Integration of Energy and Fire Prevention Systems in Green Building

Journal of Energy and Power Engineering Volume 9, Number 9, September 2015 (Serial Number 94)

David Publishing

David Publishing Company www.davidpublisher.com

Publication Information: Journal of Energy and Power Engineering is published monthly in hard copy (ISSN 1934-8975) and online (ISSN 1934-8983) by David Publishing Company located at 1840 Industrial Drive, Suite 160, Libertyville, IL 60048, USA. Aims and Scope: Journal of Energy and Power Engineering, a monthly professional academic journal, covers all sorts of researches on Thermal Science, Fluid Mechanics, Energy and Environment, Power System and Automation, Power Electronic, High Voltage and Pulse Power, Sustainable Energy as well as other energy issues. Editorial Board Members: Prof. Ramesh K. Agarwal (USA), Prof. Hussain H. Al-Kayiem (Malaysia), Prof. Zohrab Melikyan (Armenia), Prof. Pinakeswar Mahanta (India), Prof. Carlos J. Renedo Estébane (Spain), Prof. Mohamed Ahmed Hassan El-Sayed (Trinidad and Tobago), Prof. Carlos Redondo Gil (Spain), Prof. Roberto Cesar Betini (Brazil), Prof. Rosário Calado (Portugal), Prof. Dr. Ali Hamzeh (Germany), Prof. Noor-E-Alam Ahmed (Australia), Prof. E. Ubong (USA), Prof. Shripad T. Revankar (USA), Prof. Almoataz Youssef Abdelaziz (Egypt), Prof. Guihua Tang (China), Prof. Mohammad Rasul (Australia), Prof. Rene Wamkeue (Canada), Prof. Ya-Ling He (China), Prof. Filippo de Monte (Italy), Prof. Masoud Rokni (Denmark), Prof. Hosni I. Abu-Mulaweh (USA), Prof. Quan Zhang (China), Prof. Peng-Sheng Wei (Taiwan), Prof. Vinod Kumar (India), Prof. Yuan-Kang Wu (Taiwan), Dr. Kaige Wang (USA), Dr. Fude Liu (Hong Kong), Prof. Isa Salman Hasan Qamber (Bahrain), Dr. GM Shafiullah (Australia), Dr. Kelum Gamage (UK), Dr Isam Janajreh (UAE), Prof. Marco Casini (Italy), Prof. Meicheng Li (China). Manuscripts and correspondence are invited for publication. You can submit your papers via Web Submission, or E-mail to [email protected] or [email protected]. Submission guidelines and Web Submission system are available at www.davidpublisher.com. Editorial Office: 1840 Industrial Drive, Suite 160, Libertyville, IL 60048, USA E-mail: [email protected]; [email protected]; [email protected].

[email protected];

Copyright©2015 by David Publishing Company and individual contributors. All rights reserved. David Publishing Company holds the exclusive copyright of all the contents of this journal. In accordance with the international convention, no part of this journal may be reproduced or transmitted by any media or publishing organs (including various websites) without the written permission of the copyright holder. Otherwise, any conduct would be considered as the violation of the copyright. The contents of this journal are available for any citation. However, all the citations should be clearly indicated with the title of this journal, serial number and the name of the author. Abstracted / Indexed in: Database of EBSCO, Massachusetts, USA Database of Cambridge Science Abstracts (CSA), USA Chinese Database of CEPS, American Federal Computer Library Center (OCLC), USA Ulrich’s Periodicals Directory Summon Serials Solutions Chinese Scientific Journals Database, VIP Corporation, Chongqing, China Chemical Abstracts Service (CAS) ProQuest CNKI Google Scholar

Subscription Information: Price: Print $800 (per year) Online $600 (per year) Print and Online $880 (per year) David Publishing Company Tel: 1-323-984-7526, 001-323-410-1082; Fax: 1-323-984-7374; 323-908-0457 E-mail: [email protected] Digital Cooperative Company: www.bookan.com.cn

D

DAVID PUBLISHING

David Publishing Company www.davidpublisher.com

Journal of Energy and Power Engineering Volume 9, Number 9, September 2015 (Serial Number 94)

Contents Clean and Sustainable Energy 761

Integration of Energy and Fire Prevention Systems in Green Building Sergio Eduardo Palmiere, Carmen Elena Martinez Riascos and Luis Alberto Martinez Riascos

769

Extending the Use of Hothouses through Heating with Residual Agricultural Biomass Erol Murad, Edmond Maican, Cătălin Dumitrescu and Sorin-Ştefan Biriş

775

Proof of Concept of a Novel PEM Fuel Cell Stack Design with Hydraulic Compression Ulrich Rost, Cristian Mutascu, Jeffrey Roth, Christoph Sagewka and Michael Brodmann

784

Performance of R141b Ejector with Thermal Storage for Solar Air Conditioning Chakri Sripanom and Sarayooth Vaivudh

791

Stress-Strain Behavior of Nylon-Carbon Composite Subjected to High Strain Rate Impact Loading Noori Hassoon Mohammed Al-Saadi

Power and Electronic System 796

Solar-Powered Direct-Current Loads in Small Buildings Prakash Sarnobat and Simon Lannon

805

Design and Simulation of a Solar Regulator Based on DC-DC Converters Using a Robust Sliding Mode Controller Gaga Ahmed, Errahimi Fatima and ES-Sbai Najia

813

Conductive Channel for Energy Transmission Victor Victorovich Apollonov

D

Journal of Energy and Power Engineering 9 (2015) 761-768 doi: 10.17265/1934-8975/2015.09.001

DAVID

PUBLISHING

Integration of Energy and Fire Prevention Systems in Green Building Sergio Eduardo Palmiere, Carmen Elena Martinez Riascos and Luis Alberto Martinez Riascos CECS (Engineering and Social Science Center), UFABC (Federal University of ABC), Santo Andre, SP 09210-580, Brazil Received: June 18, 2015 / Accepted: July 10, 2015 / Published: September 30, 2015. Abstract: Green buildings should be sustainable, efficient, economical, and to improve the quality of life of users and nearby residents. Unfortunately, some architecture features fostered by energy and environmental certifications do not comply with fire prevention codes. Several factors can influence a building’s energy efficiency and fire safety, such as bigger or smaller glass areas, building material thermal inertia, and the use of sustainable solutions such as green roofs. In this work, a methodology for integrating energy efficiency and fire prevention in green buildings is presented. A case study comparing a full glass skin and a building with 40% window size is considered. Also, the methodology permits modifications for obtaining a safe and energy efficient building. Key words: Energy efficiency in buildings, fire prevention, EnergyPlus, Pyrosim, pathfinder.

1. Introduction Commercial and residential buildings are some of the largest energy consumers, accounting for 30%-40% of all primary energy consumed in developed countries [1]. Most of this energy is consumed by the cooling or internal heating systems, and on a smaller scale to supply lighting and security systems. For the commercial sector, the air conditioning system represents 48% of all final consumption, lighting 24%, building control equipment 15%, elevators and other loads 13%. Buildings with appropriate planning and best orientation can reduce energy consumption by up to 30%, when compared to other buildings, that offer similar comfort levels without an energy aware design. An increase as little as 2% on the initial investment of a building—construction costs—could save up to 20% of its operational costs over its entire life cycle [2]. A building design should ensure sustainability, comfort, quality of life and energy efficiency to residents. Given these needs aroused the concept of green buildings. Corresponding author: Luis Alberto Martinez Riascos, professor, research fields: fuel cell and control systems. E-mail: [email protected].

LEED (leadership in energy and environmental design) is a building certification provided by a non-governmental, non-profit organization, USGBC (U.S. Green Building Council). LEED provides volunteer and market-oriented rating systems based on energy efficiency and environmental standards [3-5]. Unfortunately, some architecture features fostered by LEED do not comply with fire prevention codes [6-8]. Amongst the most dangerous features are internal voids, which eliminate the compartmentalization protection, double skin facades, excessive natural ventilation, and the choice of low thermal insulation construction materials without fireproof treatment. In order to achieve maximum energy efficiency with proper fire safety, the buildings must be designed to integrate the existing standards and fire prevention codes [9]. Small modifications of the building envelope can save cooling system energy, reducing the possibility and severity of fire damage. The main objective of this work is to propose a methodology to integrate fire prevention with LEED standard requirements, ensuring the HVAC (heating, ventilation and air conditioning) system maximum efficiency and maintaining fire safety. A case study applying the São Paulo Fire State

762

Integration of Energy and Fire Prevention Systems in Green Building

code in Brazil is introduced [10], however, any fire code could be applied with the proposed methodology. This paper is organized as follows: Section 2 discusses some building energy efficiency and fire protection issues; Section 3 presents a methodology for integrating energy efficiency and fire safety applied to a case study; Section 4 analyzes the results. Finally, conclusions are stated.

2. Building Energy Efficiency and Fire Protection Issues Green buildings are designed to be sustainable, efficient and architecturally attractive. Many of the applied techniques to integrate sustainability and energy efficiency can increase fire risks. For example, the use of green roofs can increase the building energy efficiency but also increase the fire risk by providing more fuel. The integration of energy efficiency and fire safety maintaining the desired sustainability requires that appropriate studies be conducted. Among the main factors, that directly affect the building energy efficiency and its fire risks are included:  building architectural design;  specific surface of building materials;  characteristics of building construction materials;  amount and distribution of combustible materials;  fire starting place;  openings between interior spaces allowing fire spread;  weather conditions such as temperature and relative humidity;  architectural design of surrounding areas;  energy control and fire prevention systems.

3. Methodology for Energy Efficiency and Fire Safety For energy efficiency simulation, the following programs were used: Google Sketch UP 8 was used to model the building envelope and EnergyPlus 7-2 was used to perform energy simulations. EnergyPlus is a free open source simulation program created by DOE

(Department of Energy, USA) for modeling buildings and their heating, ventilation and air conditioning equipment. It is indicated for energy simulations by ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) Standard 90-1-2010 [11], however, since EnergyPlus is a rendering engine and does not have a friendly graphical interface, Open Studio 0.11.0, which presents a graphical interface facilitating the data integration and visualization, was also used. Fire simulation was conducted in Pyrosim, version 2012 [12]. The program provides a GUI (graphic user interface) for applying the FDS (fire dynamics simulator) + evacuation, developed by NIST (National Institute of Standard. and Technology), and simulate fire scenarios applying CFD (computational fluid dynamics). Pyrosim has a limited combustive material data library, but facilitates the visualization, simulation and use of FDS simulator, also it is applied to simulate building ventilation according to the building envelope. Evacuation simulation was performed in Pathfinder [13], which permits to analyze different scenarios. For considering the integration of energy efficiency in green buildings in view of fire safety standards, a commercial building was analyzed, comparing two envelope options: a fully glass skin and an opaque skin with 40% window size. For simplification, simulations were performed considering these options as two different buildings. The analyses of the two buildings followed the methodology presented below. 3.1 Pre-project The building contains 11 floors (ground floor plus 10 elevated floors), they are 8.5 m wide by 25.5 m long. Each floor is 2.7 m high with external walls 25 cm thick and internal walls 15 cm thick. The elevated floors consist in two rooms per floor with building access through an emergency stairwell and an elevator, except for the ground floor, which contains only one room. Fig. 1 shows the elevated floor building plan.


763

materials control, emergency exits, building emergency response team, emergency light, fire alarms, emergency signaling, fire extinguishers, and fire hydrants. 3.5 Building Materials

Fig. 1 Elevated floor building plan.

3.2 Building’s Population The population determines what types of combustive material will be incorporated to the building, and the psychology and physical aptitude of the users, if it helps or delays evacuation. This building was designed to attend an average fixed population of 120 people (medics, psychologists, lawyers, engineers, etc.), and a floating population of 150 people (clients). The physical and psychological conditions of the building population are defined as average. 3.3 Building Location The correct building location determines the weather file to be used with EnergyPlus. The geo localization considers the influences of the local climate, solar radiation and temperature. The building was geo locate with the Google Earth tool program integrated with Google Sketch UP 8, latitude: -23.671470S, longitude: -46.552697W, time zone: -3 and elevation: 803 m, São Paulo, Brazil. Sketch Up also helps to simulate shadow generated by nearby buildings. 3.4 Fire Code In this case, the São Paulo Fire Code, Decree 56819 was applied. Brazil has different and outdated fire codes in some cities and states. Decree 56819 applies tables using the building height to determine the fire loads and fire protections required by law. The building has 29.7 m high and 700 MJ/m2 fire load, demanding 90 min of passive fire resistance time for all construction materials, 25 tons. Firefighter vehicle access, structural fire safety, horizontal and vertical compartmentalization, finishing and construction

The fifth step consists of determining the building material sets and construction layer. EnergyPlus uses this information to determine the building energy consumption. Table 1 presents two common construction material sets that should be recycled, have low environmental impact, and can be extracted near the construction site. 3.6 Energy Consumption Analysis The energy consumption information gathers the quantity and the use of the HVACs. Eleven thermal zones were considered, one for each floor. Those zones represent areas where thermal difference can exist and are close enough to interact with each other. The HVAC systems have VAV (variable air volume) with PFP (parallel fan powered) boxes and reheat. The building internal loads represent the energy released as heat by people (0.3 radiant fraction), lamps (6.4 W/m2), and electrical equipment (9.7 W/m2). The infiltration represents the airflow that penetrates into unintended areas of the thermal environment, usually caused by opening and closing doors and windows. Table 1 Construction material lists applied in the simulation. Elements

List 1

Roof

Heavy concrete

Interior slab Interior wall Exterior wall

Tile Light concrete Air interlayer Plaster Mortar plastering Brick (25 cm) Granite Mortar plastering Light concrete Mortar plastering

Floor

Heavy concrete

Glass

Translucent 3 mm

List 2 Exterior tile Heavy concrete Ceramic floor Mortar plastering Light concrete Mortar plastering Ceramic brick Mortar plastering Ceramic brick Ceramic floor Mortar plastering Heavy concrete Translucent 3 mm

764


According to ASHRAE, a value of 0.75 air changes of the total volume by hour was adopted. Table 2 shows building’s occupancy schedule, for determining the higher and lower energy consumption periods. 3.7 Fire Risk Analysis The simulation should consider the worst situation. Vertical proliferation of flames and smoke is faster, affecting upper floors quickly. Then, the fire is supposed to start in lower floors. The building’s population can include people with disabilities, for example, wheelchair users.

Table 2 Occupancy (%). Schedule 8:00-9:00 9:00-10:00 10:00-12:00 12:00-14:00 14:00-16:00 16:00-17:00 17:00-18:00 18:00-19:00

Monday to Friday 25 100 100 50 100 100 100 25

Saturday 0 25 100 50 100 25 0 0

Sunday & holydays 0 0 0 0 0 0 0 0

3.8 Energy Efficiency Simulation The simulation determines the energy consumption based on the construction materials, the best building orientation for reducing HVAC energy consumption, and the required LEED baseline building, which will serve as a standard comparative for green building certification. A baseline building should be determined for each of the proposed models to serve as energy efficiency reference, considering the average consumption obtained from four different building rotations such as 0°, 90°, 180° and 270° (ASHRAE 90-1 2010, appendix G). Fig. 2 shows the influence of the two construction material lists on the energy consumption, it also shows that, the energy consumption is proportional to the window size. Figs. 3 and 4, respectively show the glass skin and 40% window size building consumption results considering different rotations and the most efficient materials (list 2 from Table 1), it also shows that, the 40% window size building is the most efficient. Fig. 5 shows the end user participation of each equipment type in 270º orientation, although 90º orientation shows a little better efficiency it is opposed to the street orientation. Fig. 6 shows that, in the glass skin building, the HVAC system consumes a bigger portion of the energy. Different types of glasses were simulated. The

Fig. 2 Construction material influence over energy consumption. G.S.—glass skin, 40% W.—40% window size.

Fig. 3 Glass skin consumption results for different rotations.

Fig. 4 40% window area consumption results for different rotations.


Heating Cooling Lighting Equipment Ventilation Pumps

Total

Fig. 5 End use participation by building type: glass skin vs. 40% window size (90º orientation).

Fig. 6

765

Glass skin chart results.

simulation shown that, Galaxy Clearvision double glazing with an insulating gas interlayer of 13 mm and 6 mm clear window has a better efficiency. Figs. 6 and 7 show the simulation energy consumption results considering the glass skin and 40% window buildings with different structural modifications. Note that, the glass skin building will not consider sun blind shutter or awnings. Applying blinds and efficient glazing in the 40% window building, the refrigeration energy consumption is reduced, but the energy for heating, ventilation and pump systems increases more, resulting in a final consumption of 4,516 kW/h, greater than the annual consumption applying only efficient glasses. 3.9 Fire Simulation Pyrosim [12] was applied for fire simulation (only paper and wood are available as aggregated materials). The desired heat release rate curve followed the São Paulo’s Decree 56819 (700 MJ/m2). For simulating ignition until the flashover, Pyrosim reference burner starts at 33,716 W/m2, representing 0.2% of the total

Fig. 7 40% window area chart results. 1—baseline, 2—270º orientation, 3—efficient glazing (double glass Galaxy Clearvision + 13 mm air interlayer (+ 6 mm clear window), 4—blinds and 270º orientation, 5—three modifications simultaneously (270º orientated, blinds and efficient glazing).

fire load (it has 0.50 m2 size and a heat release rate of 16,858 kW). The mesh encloses 10 m beyond the building [13]. The fire starts in the middle of the 3rd floor’s left room. Thermocouples are placed near the ceiling, in the middle and the corners (in both sides of each room windows). The glasses are set to disappear from simulation in each floor when a temperature of 250 °C was reached in thermocouples near the windows, causing the outside air penetration. Figs. 8a and 9a show the fire dynamics in the 40% window size and glass skin buildings; Figs. 8b and 9b present Pyrosim thermometer results, illustrating the temperature inside the 3rd floor where the fire is meant to start. Sprinklers are not mandatory in buildings lower than 30 m in Brazilian territory; Fig. 10 shows the temperature evolution considering an automatic sprinkler system. Four evacuation sceneries were simulated applying Pathfinder [14], as follows: Scenery 1: Only healthy people are in the building and only the emergency stairwell is an escape route; Scenery 2: One wheelchair person is in 3rd floor (fire starting place), which must be evacuated by the elevator; Scenery 3: Three persons with physical disabilities using the elevator, and healthy people being evacuated by the emergency stairwell;

766


420.0

1,200.0

360.0 340.0

1,000.0

300.0

800.0

260.0 600.0

220.0 180.0

400.0

140.0 200.0

100.0 60.0

0.0 100.0

20.0

200.0

300.0

400.0

500.0 600.0 700.0

800.0

900.0 1,000.0

Time (s)

(a)

(b)

Fig. 8 (a) 40% window size building, (b) fire proliferation results. 570.0 515.0 480.0

1,400.0

405.0

1,200.0

350.0

1,000.0

295.0

800.0

240.0

600.0

185.0 130.0 75.0

400.0 200.0 0.0

20.0

(a)

(b)

Fig. 9 (a) glass skin building, (b) fire proliferation results.

Table 3 shows the scenery results, also considering a 3 min pre-evacuation time [15-18]. Pathfinder and Pyrosim do not work together, although both programs belong to the same company. Thus, the evacuation simulator do not consider the effect of people been reached by fire.

5. Results Analysis

Fig. 10 Sprinkler time control results.

Scenery 4: Most people using the emergency stairwell, but also healthy and people with disabilities will use the elevator.

The energy simulation shows that, buildings with smaller window size are more efficient when applying common glasses. Applying special efficient glasses the energy savings is 27% in the glass skin building and 20% in the 40% window size. Note that, this modification greatly increases the energy efficiency in

Integration of Energy and Fire Prevention Systems in Green Building Table 3 Evacuation time results. Disability

Scenery 1 Scenery 2 Scenery 3 Scenery 4

Wheelchair

-

104.1

104.1

208.1

Lame

-

67.2

173.9

360.6

Blind

-

84.1

233.5

419.3

Total time

155.3

159.5

233.5

419.3

Total evacuation time plus 3 min of pre movement Disability

Scenery 1 Scenery 2 Scenery 3 Scenery 4

Wheelchair

-

284.1

284.1

388.1

Lame

-

247.2

353.9

540.6

Blind

-

264.1

413.5

599.3

Total time

335.3

339.5

413.5

599.3

the glass skin building, representing more credits for LEED green building certification. However, the total consumption is still higher in the glass skin building. Figs. 8b and 9b show reduced fire proliferation rate in smaller window size buildings (small openings). Glass skin buildings have ventilation excess, producing fast fire proliferation to the upper floors. Other reason for this result is the thermal inertia characteristic of the concrete, granting a higher fire protection and reducing the spread of fire. The results of both simulations (considering buildings without sprinkler protection) show that, the alarm will be activated 80 s after the fire starts, with the third floor becoming hazardous 20 s after the alarm sounds. Mass casualties will happen, since the occupants of the other floors would spend from 336 s to 600 s for a full building evacuation (considering 3 min. for pre-evacuation), because the fire spreads to the upper floors. Even ignoring the pre-evacuation time, the minimum evacuation time would be from 156 s to 420 s, therefore building modifications are necessary to guaranteed the occupant safety (for example, fireproof stairs or implementing fire shelters). The floors below the fire source or close to the roof have a better protection for its occupants. In the 40% window size building, the floors affected by the fire were from 3rd to 9th, and in the glass skin, building floors from 3rd to 11th were affected. For performing modifications, the proposed

767

methodology suggests their implementation in backward direction, until results are satisfactory. In this case, it is necessary to return to Step 5 (building materials), for example, considering isolated fireproof stairs, fire shelters, or fire resistant glasses. If results are still not satisfactory, then it is necessary to return to Step 4 (fire code), for example, considering sprinkles in both buildings. However, if results are still not satisfactory, then it is necessary to return to Step 1, for example, considering setback or balconies in the building plan (only possible in the 40% window building). In this case, modifications for the building location (Step 3) or building’s population (Step 2) are not considered.

6. Conclusions A methodology considering energy efficiency and fire safety was introduced. The methodology permits to analyze different building modifications (construction material types, different fire code requirements, etc.) for determining the better practices to be adopted. It will facilitate to observe the fire risks for each structural modification and new project. Fire simulations can be performed for determining the building’s fire risk, its potential direct and indirect losses and whether the procedures adopted are enough to ensure the survival of the occupants. Simulation programs are worth tools, helping to develop more efficient and more secure buildings, facilitating the integration of several factors, but any software can truly foresee people survive attitudes inside a building on fire, thus the simulation serves only as a small reference. Pyrosim lacks of a fully fire load library, only paper and wood were tested as fire aggregated loads. It is highly recommended to simulate the full aggregated fire loads that could be found inside a building to get a better understand of the fire risks. Different opening sizes directly impact the energy efficiency and fire spread in buildings, although special glazing grants better energy savings, it may


768

become a rescue barrier for firefighters. The 40% window building also permits more structural changes, such as applying setbacks or balconies, therefore, it can be more secure than the glass skin building. The green building fire risk analysis is a new study field. More studies should be conducted for integrating fire risk and energy efficiency.

References [1]

[2]

[3]

[4] [5]

[6]

[7]

[8]

Escrivã, E. G., Santamaría, O., and Mugarra, L. F. 2012. “Continuous Assestment of Energy Efficiency in Commercial Buildings Using Energy Ratings Factors.” Elsevier Energy and Buildings 49 (June): 78-84. Salman, A., Wade, A. C., Darren, O., and Irtishad, A. 2011. “Building Information Modeling for Sustainable Design and LEED® Rating Analysis.” Elsevier Automation in Construction 20 (2): 217-24. USGBC (U.S. Green Building Council). 2009. “LEED 2009 for New Construction and Major Renovations.” USGBC. GBCI (Green Business Certification Inc.). 2012. “LEED Certification Policy Manual.” GBCI. Dall’O, G., Speccher, A., and Bruni, E. 2012, “The Green Energy Audit, a New Procedure for the Sustainable Auditing of Existing Buildings Integrated with the LEED Protocols.” Elsevier Sustainable Cities and Society 3 (July): 54-65. Meacham, B., Poole, B., Echeverria, J., and Cheng, R. 2012. “Fire Safety Challenges of Green Buildings.” In The Fire Protection Research Foundation, Springer, MA: Worcester Polytechnic Institute. Meacham, B. 2012. “Fire Safety Challenges of Green Building Presentation.” Presented at the Fire Safety and Sustainable Building Symposium, Chicago, USA. Tidwell, J., and Murphy, J. J. 2010. Bridging the Gap: Fire Safety and Green Buildings. A Fire and Building

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Safety Guide to Green Construction. Quincy: National Association of State Fire Marshals. Patterson, M. G. 1996. What is Energy Efficiency? Concepts, Indicators and Method. Palmerston North: Elsevier Sci. Ltd. São Paulo’s Fire Department. 2011. “São Paulo State Law 56819.” São Paulo’s Fire Department. Accessed September 10, 2015. http://www.ccb.policiamilitar.sp.gov.br/credenciamento/d ownloads/Decreto_Estadual_56819.pdf. ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) 90-1. 2010. “Energy Standard for Buildings Except Low-Rise Residential Buildings.” ASHRAE. Thunderheard Engineering. 2013. “Pyrosim User Manual.” Thunderheard Engineering. Accessed September 10, 2015. http://www.thunderheadeng.com. Mazzoni, F. 2010. Fire Simulation: A Study of the Cacique Bldg. Porto Alegre: UFRGS (Universidade Federal do Rio Grande do Sul). Thunderheard Engineering. 2013. “Pathfinder User Manual.” Thunderheard Engineering. Accessed September 10, 2015. http://www.thunderheadeng.com. Tong, D., and Canter, D. 1985. “The Decision to Evacuate: A Study of the Motivations Which Contribute to Evacuation in the Event of Fire.” Fire Safety Journal 9 (3): 257-65. Kobes, M., Helsloot, I., Vries, B., and de Post, J. G. 2010. “Building Safety and Human Behavior in Fire: A Literature Review.” Elsevier Fire Safety Journal 45 (1): 1-11. Proulx, G. 1995. “Evacuation Time and Movement in Apartment Buildings.” Elsevier Fire Safety Journal 24 (3): 229-46. Chow, W. K., and Chow, C. L. 2005. “Evacuation with Smoke Control for Atria in Green and Sustainable Buildings.” Elsevier Building and Environment 40 (2): 195-200.

D


DAVID

PUBLISHING

Extending the Use of Hothouses through Heating with Residual Agricultural Biomass Erol Murad1, Edmond Maican1, Cătălin Dumitrescu2 and Sorin-Ştefan Biriş1 1. Department of Biotechnical Systems, Politehnica University of Bucharest, Bucharest 060042, Romania 2. Hydraulics and Pneumatics Research Institute (INOE-2000 IHP), Bucharest 040558, Romania Received: June 19, 2015 / Accepted: August 08, 2015 / Published: September 30, 2015. Abstract: The paper presents a study on increasing energy independence of a 200 m2 horticultural hothouse, by means of heating it with thermal energy from a TLUD (top-lit-up-draft) gasification procedure of local residual agricultural biomass, chopped at 10-50 mm and dried at 10%-15% RH (relative humidity). It produces an average of 14% higt quality biochar. Hot-air heating system and forced circulation are equipped with two GAZMER 40/150G energetic modules, which are rechargeable, simple, safe, efficient and environmentally friendly. They can gasify chopped or pelletised biomass. To study the microclimate evolution, it was used a complex numerical model for a 200 m2 hothouse for growing vegetables. Simulated experiments were carried out for frosty days and, estimated, for the whole warming period. Each year 13.44 t of biomass are consumed, resulting 1.78 t of biochar which, when introduced in soil, produces a -6.2 t/year negative balance of CO2. Key words: Hothouse, energy independence, TLUD, biochar, automat control, simulation.

1. Introduction Essential to increase food security of Romania and to increase the health of the local population, is production of current consumption of vegetables throughout the winter. Winter vegetable production comes up to about 2/3 out of unheated hothouses and the difference comes from industial type greenhouses. Due to competition in the global floricultural products, much of protected crop surfaces are allocated to flowers production, at the expense of production of vegetables. By reducing expenses necessary for heating hothouses and greenhouses, there can be produced more vegetables, with high quality and making an adequate profit both for producers and consumers. To increase the life of the vegetable hothouses in order to increase vegetable production during the cold season, heating is required with simple thermal systems, Corresponding author: Erol Murad, associate professor, research fields: automation and control systems, thermo-chemical gasification of biomass, modeling and simulation of unconventional processes. E-mail: [email protected].

with low costs both in investment and operation. To reduce dependence on fossil fuels, with high CO2 emissions, currently the share of biomass for heating increases and, more rarely, biomass is used in cogeneration plants. Currently, it is used the biomass compacted into pellets, easy to carry and use as fuel, which required an extension of the use of this form of compaction. Burning is done with high superficial speeds which produces a higher concentration of PM that clogs the heat exchanger and pollute the environment. EU environmental legislation heavily restricts the concentration of PM (particulate matter), which leads to use of smoke filters or towards reducing the intensity of the thermal process. These in turn have the effect of price increases and a decrease of overall performance. For the production and transportation of pellets, an average of 200 kWh/t is consumed, which for a thermal/electric conversion efficiency of 25% represents about 16%-20% of the calorific energy of the compacted biomass. It is issued into the atmosphere, a large amount of CO2 from the production of

770

Extending the Use of Hothouses through Heating with Residual Agricultural Biomass

electricity consumed and from the manufacturing of production equipment. The current price is 140-160 €/t, slightly higher during the cold season [1]. Alternatively, a more ecological and economic alternative for thermal energy production needed to extend the useful life of hothouse with three months, or working like a greenhouse during the entire cold season, it is proposed and analyzed the use of local agricultural biomass residues at the level of current technology, with minimal mechanical processing [2-4]. Agricultural biomass, derived from fruit growing, viticulture and vegetable growing, is chopped at 10-50 mm and naturaly or forced dried at 10%-15% RH (relative humidity). Energy consumption from harvesting to use is 100 kWh/t less than in case of pelletization, and 50 kWh/t less for transport on distances under 30 km. On average is 50 kWh/t. This represents about 25% of the power consumption pelleting, or 4%-5% of the calorific power of used biomass. It is estimated that, this type of local fuel can be used at maximum 50 €/t [2, 5, 6]. The most important reasons for using this type of biomass consists of ecological and economic advantages (lower consumption of fossil fuels), dependency reduction on unstable geopolitical entities. It also has the effect of encouraging the efficient use of local energy resources, innovation and employment growth in rural areas.

2. Production of Heat with TLUD Energetic Modules To convert the local chopped biomass at 10-50 mm, with 10%-15% moisture, in thermal energy, there can be used relatively low conversion efficiency combustion plants that produce a lot of PM (smoke), or thermo-chemical gasification can be applied, whic means lack of smoke and good efficiency. Conventional gasification plants are still relatively complex equipment, are senzitive to moisture and size of biomass, require highly trained operators, thus slowing down their use in agriculture. In order to

reduce both the initial investment and the running operation costs, it is proposed the TLUD (top-lit-up-draft) process of gasification, which is tolerant and functional auto-adaptive to the particle size and moisture of biomass. It is also characterized by PM and CO very low emissions. Energetic modules for producing thermal energy based on the TLUD process have a simple design, are durable and safe under operation, and benefit from a cheaper price [4-7]. The TLUD gasification process is a reverse downdraft process that works in batch mode with reduced thermal load. Kinematics and dynamics of the gasification process are similar. At 25%-30% excess of air for gasification, not all carbon in biomass is reduced. Thereby at the end of the charge results 10%-15% charcoal, called biochar, with incorporated ash. Biochar has proven to be a valuable agricultural amendment to increase soil fertility. Considering biochar as a high usable value product, is obtained a thermal energy conversion efficiency of the completely gasified part of biomass of 92%-96%, according to the operating conditions and characteristics of used biomass [5-8]. From the tests, it was found that, the LHV (lower heating value) of produced syngas is 3.5-4.5 MJ/Nm3 which, summed with hot gas enthalpy at 500 °C, leads to an available energy of 1.2-1.5 kWhth/Nm3 [6, 8-10]. In the reactor, the produced gas has a very low superficial velocity vs = 0.03-0.06 m/s, providing a very low concentration of solid particles (PM2.5  2.4 mg/MJ), practically smokeless operation, far below standards imposed on the field. To operate with a high energy conversion efficiency and with minimum emissions, the minimum load of energetic modules is limited to 40% [6, 8, 10].

3. Hot Air Heated Hothouse To determine the specific greenhouse heating with TLUD energetic modules, a quonset-metric type hothouse was chosen. It consists of 28 base modules having 6 m, height 3 m, pitch of 1.2 m, and has a


surface area on the ground of 200 m2 and a volume of 475 m3. It is covered with a double layer, inflatable, polyethylene, with high thermal resistance [3, 4]. The hothouse is heated with hot air, distributed through a flexible duct which has openings for the jets in the greenhouse space. Part of the inside air is recirculated and mixed with the outside air necessary for venting the greenhouse. The mixture is heated with an internal-flue HE (heat exchanger). This design eliminates the need for regular ventilation, whom is producing energy losses and decreases in temperature, which could prove harmful if not properly managed. Indoor crop consists of Vidra Romanian variety of tomato, for which experimental data for breath and perspiration are available. In order to protect the plants from contact with too warm air currents, the maximum temperature of the jets is limited to 40 °C, which required a constant mass flow (Dam = 2 kg/s) of the air heating the greenhouse. Fig. 1 shows the block diagram of a hothouse heated by two BGM (biomass-gasification-module) thermal modules. The two modules are connected to the burner mounted at the entrance of the internal-flue heat exchanger operating at a constant flow of heated air, and therefore, having a minimum efficiency of 85%. The experiment was simulated for a winter day

771

temperatures within -25-(-5) °C, and constant indoor temperature (Ti = 10 °C) and humidity (Ui = 70%). It has been determined the maximum thermal power required to heat the greenhouse: Pincmax = 58 kW. It follows that, the maximum thermal power produced by a BGM module (PBGM) must be at least: PBGMn 

Pinc max

SC N BGM



58 0.85  2

 34.12 kW

(1)

where, ηSC is minimum return of the heat exchanger, NBGM is number of modules. There have been selected two GAZMER 40/150G energetic modules with a 400 mm reactor diameter and a hight of 1,500 mm. Reactor size was determined by the possibility of using agricultural biomass, whose bed density is between 200 kg/m3 and 300 kg/m3. Rated power of an energetic module is 40 ± 5 kW, with an average specific biomass consumption of 100 kg·bm/(m2·h).

4. Simulated Experiments The block diagram in Fig. 1 shows the basic structure of the SERMGB27.DP simulation program, which is designed for hothouses microclimate. Simulated system consists of four subsystems: hothouse conditioning process, air heater, power unit with two BGM modules,

Fig. 1 Block diagram of a hothouse heated by TLUD energetic modules.

772


subsystem for automatic control. To model the process of conditioning the hothouse in the block diagram have been highlighted the following specific parameters: input flow Din having temperature Tin and relative humidity φin; evacuated Dev and recirculated Drec air flows. BGM energy subsystem has two modules loaded with biomass, out of which every hour are consumed Cbm1 and Cbm2 (kg·bm/h), depending on the air flow for gasification Dag (kg·air/s), regulated by Cag1 and Cag2 flaps. The fuel gas produced is burnt with an air flow Dar adjusted with the flap Cab. Flue gas enters the HE and transfers heat to the air that warms the hothouse [3, 4, 11]. Automatic control subsystem must be able to adjust key parameters that define the hothouse microclimate: indoor temperature (Ti) and humidity (Ui), CO2 concentration (CCO2), as well as burner flame temperature (Tfl). Setting the indoor temperature is done with a PID (proportional-integral-derivative) numerical algorithm, which develops a control parameter that changes the energy load of modules by varying gasification flow rates (Dag). Simultaneously, a dynamic optimization algorithm changes the flow of combustion air, so that the flame temperature (Tfl) to be as close as possible to the maximum value, thus maximizing the efficiency of burner [3, 4, 11-13].

5. Results and Discussion Fig. 2 presents the result of an experiment with the φ model of studied hothouse, for an average winter day, used to calculate heating requirements in Ilfov County, with a reference for the indoor temperature rTi = 10 °C by night and 12 °C during the daytime, and for an indoor relative humidity rUi = 70%. Variations of the following parameters are displayed: Tex—external temperature, Rads—solar radiation, Ti, Ui, Pth—produced thermal power, Mcomb—biomass consumption. It is found that, Ti has little variation compared to the reference, but Ui varies somewhat more than the imposed reference. This shows that, the automatic control algorithm does not achieve full

performance requirements. Table 1 summarizes the simulation results of the heating for the cold season months when using chopped agricultural biomass, compared to using diesel. The first important resulted value is the annual consumption of biomass which is of 13,440 kg·bm/year, and which occupies a volume of 54 m3 in warehouse in 108 containers of 0.5 m3. Another valuable characteristic is the average specific consumption of biomass for heating during winter cbms = 13.2 kg·bm/(K·day). With this value, can be quickly and accurately size heating systems for greenhouses and hothouses. From biomass gasification in BGM results 1,780 kg of biochar each year, that can harness 200 €/t, or that can be incorporated as amendment for the soil in solarium with a negative annual CO2 balance of -6.2 t/year. Average specific cost of biomass for hothouse heating during winter is 0.672 €/m2·month, which is about 20 times lower than with diesel and four times lower than with pellets. This clearly shows the economic efficiency of minimally processed local agricultural biomass into gas-producing TLUD.

6. Conclusions By heating hothouses with thermal generators based on TLUD gasification process, local agricultural biomass residues, minimally processed, can be used economically and environmentally efficient at a price of 50 €/t. The specific costs for heating are four times lower than for pellets and 20-fold lower than for diesel fuel. Analyzed hothouse can be used as greenhouse throughout the entire cold season with a consumption of 13.5 tons biomass, from which results 1.8 tonnes of biochar that can be sold as charcoal, or that can be incorporated as amendment for the soil in solarium with a negative annual CO2 balance of -6.2 t/year. Biomass hothouse heating in the analyzed version is done with virtually no emissions of toxic effluents, pollutants, and with negative balance of atmospheric carbon.


773

Time (h) Fig. 2 The evolution of a hothouse microclimate in a winter average day. Table 1 Simulated experimental results. Parameter Heating period Average outside temp. Indoor average temp. Average required heating Biomass specific consumption Biomass consumption Biochar production Price—chopped agric. biomass Price—biochar as amendment Cost biomass for heating Revenue capitalization biochar Net cost heating fuel Specific cost—biomass heating

Month

day/month grade C grade C K·day/month kg.bm/(K·day) t·bm/month t·bc/month €/t €/t €/month €/month €/month €/(m2·month) Net cost—biomass heating €/(m2·month) Diesel consumption for heating L·mt/(kg·bm) Diesel—monthly cost €/month Specific cost—diesel heating €/(m2·month) CO2 balance with carbon sequestration t (CO2)/month

Oct.

Nov.

Dec.

Jan.

Febr.

Mar.

15 7.00 10.50 52.50 13.20 0.693 0.092 50.00 200.00 34.65 18.36 16.29 0.173 0.081 0.35 327.03

30 5.20 10.50 159.00 13.20 2.099 0.278 50.00 200.00 104.94 55.62 49.32 0.525 0.247 0.35 990.44

31 -0.10 10.50 328.60 13.20 4.338 0.575 50.00 200.00 216.88 114.94 101.93 1.084 0.510 0.35 2,046.92

31 1.20 10.50 288.30 13.20 3.806 0.504 50.00 200.00 190.28 100.85 89.43 0.951 0.447 0.35 1,795.88

28 5.60 10.50 137.20 13.20 1.811 0.240 50.00 200.00 90.55 47.99 42.56 0.453 0.213 0.35 854.65

15 7.00 10.50 52.50 13.20 0.693 0.092 50.00 200.00 34.65 18.36 16.29 0.173 0.081 0.35 327.03

Cold season 150 3.71 10.50 1,018.10 13.20 13.439 1.781 50.00 200.00 671.95 356.13 315.81 0.672 0.316 0.35 6,341.96

1.64 -0.32

4.95 -0.97

10.23 -2.00

8.98 -1.76

4.27 -0.84

1.64 -0.32

6.34 -6.20


774

A comprehensive program was developed to simulate the hothouse microclimate, SERMGB27.DP, the model also including the TLUD process thermal generators model. Automatic control algorithm for hothouse climate has been verified, which ensures both maintaining optimal microclimate parameters and automatic control of heat generator. The model developed for simulation is a strong basis for research and development in the field, being open source. Results require an experimental validation on a real hothouse, both for confirming the results of simulated experiments and to determine the real economic aspects derived from extending the useful life during cold season. Intensive use of local agricultural biomass resources for the production of heat and biochar leads to an increase in the use of labor in rural areas, to a better use of local resources, and to a higher living standards and sustainable development in the area.

[5]

[6]

[7]

[8]

[9]

[10]

References [1]

[2]

[3]

[4]

Risovic, S., Dukic, I., and Vukovic, K. 2008. “Energy Analysis of Pellets Made of Wood Residues.” Croat Journal for Engineering 29 (1): 95-108. Daugherty, E. C. 2001. “Biomass Energy Systems Efficiency Analized trough a Life Cycle Assesement.” Master’s of Science thesis, LUND University. Murad, E., Maican, E., Haraga, G., and Biriş, Ş. S. 2011. “Greenhouse Module Heating with Biomass.” Presented at the International Symposium, Horticulture—Science, Quality, Diversity, Harmony, Iasi, Romania. Murad, E., Maican, E., Biriş, S. S., and Vlăduţ, V. 2011. “Heating Greenhouses with TLUD Biomass Energy Modules.” Presented at the 3rd International Conference

[11]

[12]

[13]

on Research People and Actual Tasks on Multidisciplinary Sciences, Lozenec, Bulgaria. Murad, E., Achim, G., and Rusănescu, C. 2012. “Energy Recovery from Biomass and Organic Orchards Cuts, Scientific Communications Session.” Research and Development Institute for Industrialization and Marketing of Horticultural Products—ICEDIMPH-HORTING. Murad, E., and Dragomir, F. 2012. “Heat Generators with TLUD Gasifier for Generating Energy from Biomass a Negative Balance of CO2.” In Proceedings of the International Conference—HERVEX-2012, 440-7. Murad, E., Culamet, A., and Zamfiroiu, G. 2011. “Biochar- Economically and Ecologically Efficient Technology for Carbon Fixing.” Presented at the 2011 International Salon of Hydraulics and Pneumatics—HERVEX, Calimanesti-Caciulata, Romania. Mukunda, H. S. 2010. “Gasifier Stoves—Science, Technology and Field Outreach.” Curent Science 98 (5): 627-38. Reed, T. B., and Das, A. 1988. “Handbook of Biomass Downdraft Gasifier Engine Systems.” U.S Department of Energy. Varunkunar, S. 2012. “Packed Bed Gasification-Combustion in Biomass Domestic Stove and Combustion Systems.” Ph.D. thesis, Department of Aerospace Engineering Indian Institute of Science. Ramírez-Arias, A., Rodríguez, F., Guzmán, J. L., Arahal, M. R., Berenguel, M., and López, J. C. 2005. “Improving Efficiency of Greenhouse Heating Systems Using Model Predictive Control.” Copyright © 2005 IFAC (International Federation of Automatic Control). Murad, E. 2005. “Optimisation of Biomass Gasification Load Regime.” Presented at the CIEM (International Conferece Energie-Mediu) 2005, Bucureşti, Romania. Murad, E., Dumitrescu, C., Haraga, G., and Dumitrescu, L. 2012. “Pneumatic Measurement of the Biomass Consumption for TLUD Generator.” Presented at the International Conference HERVEX 2012, Călimăneşti, Romania.

D


DAVID

PUBLISHING

Proof of Concept of a Novel PEM Fuel Cell Stack Design with Hydraulic Compression Ulrich Rost1, 2, Cristian Mutascu1, Jeffrey Roth1, Christoph Sagewka1 and Michael Brodmann1 1. Westphalian Energy Institute, Westphalian University of Applied Sciences, Gelsenkirchen D-45145, Germany 2. Faculty of Industrial Chemistry and Environmental Engineering, Politehnica University Timisoara, Timisoara 300222, Romania Received: July 03, 2015 / Accepted: August 03, 2015 / Published: September 30, 2015. Abstract: In this study, a novel design concept for PEMFC (polymer electrolyte membrane fuel cell) stacks is presented with single cells inserted in pockets surrounded by a hydraulic medium. The hydraulic pressure introduces necessary compression forces to the membrane electrode assembly of each cell within a stack. Moreover, homogeneous cell cooling is achieved by this medium. First, prototypes presented in this work indicate that, upscaling of cells for the novel stack design is possible without significant performance losses. Due to its modularity and scalability, this stack design meets the requirements for large PEMFC units. Key words: PEMFC, modular stack design, hydraulic compression, homogeneous cooling.

1. Introduction In the year 2050, Germany intends to generate over 80% of its electrical energy by renewable sources [1]. To succeed in transforming, the German energy sector towards this high share of renewable energy sources huge effort has to be made in improving the electrical grid infrastructure as well as in extending energy storage. To fulfil carbon dioxide emission targets, it is common sense to result in carbon free strategies. Producing hydrogen (H2) and oxygen (O2) in water electrolyzers is a potential solution for using excess energy in periods of low demand [2]. Storing large amounts of H2 in caverns under ground and transportation of it in pipelines is shown at few locations worldwide [3]. In order to generate electrical energy in times of high demand, fuel cell systems can be powered by H2 and O2 or air. Due to high power density and dynamic part load operation, PEM (polymer electrolyte membrane) based fuel cell or electrolyzer systems are appropriate for this purpose [4, 5]. Corresponding author: Ulrich Rost, M.Eng. (Ph.D. candidate), research fields: hydrogen energy systems and energy storage. E-mail: [email protected].

The NOW (national organization for hydrogen in Germany) calculated the demand for hydrogen energy systems as positive and negative supply for varying power demand for the German energy economy in 2025 to be in the range of a up to 5 GW [6]. German demonstration projects combining renewable energy sources with hydrogen systems, like “H2 Herten” or “Falkenhagen”, started operation within the year 2014 [7, 8]. The main component in PEMFCs (polymer electrolyte membrane fuel cells) is the MEA (membrane electrode assembly). MEAs for PEMFCs consist of an ions conducting but electrical insulating membrane, which is coated from both sides with catalytic material (CCM (catalyst coated membrane)) and, furthermore, an electrical conducting GDL (gas diffusion layer) on the anode as well as on the cathode side. Such a MEA is placed between two pole plates with gas channels (flow field). Together these components form a single fuel cell (Fig. 1). H2 and O2 or air (about 21% O2) are supplied over the flow field and finely distributed to the catalyst due to the high porosity of a GDL. H2 is delivered to the anode and O2 is delivered to the cathode, where product water evolves (H2 + ½O2

776

Proof of Concept of a Novel PEM Fuel Cell Stack Design with Hydraulic Compression

2. A Modular PEMFC Stack Design with Hydraulic Compression

Fig. 1 Schematic drawing of a PEMFC in cross section.

→H2O + ΔHL with ΔHL equals 3 kWh·m-3 according to the LHV (lower heating value) of H2) [9-12]. Typical fuel cell stacks consist of several single fuel cells connected electrically in series which increases stack voltage at constant current. To reduce electrical resistance, pole plates are constructed in bipolar design, so that a flow field is arranged on each side of the plates. Mono-polar plates are placed on the stacks extremities with a flow field only on one side. Furthermore, special pole plates which have channels for cooling water can be placed within the stack to protect it from overheating. Over two strong endplates which are pulled together by tie bolts, necessary compression forces are induced to improve contact resistance between the electrode components as well as catalyst contact to the membrane [12]. To reduce reactance within PEMFC stacks without setting down porosity of the electrode’s layers, an optimal compression over the whole cell surface is needed. Moreover, well distributed current density is a central requirement for PEMFC systems to avoid hot spots which may accelerate fuel cell degradation and malfunction [13-21]. In this work, a novel PEMFC stack design is presented which offers homogeneous operation condition for each cell of a stack by hydraulic compression. Inherent stack cooling by a hydraulic medium is a further advantage of the described stack design. Two laboratory scale prototypes indicate modularity and scalability which are required in order to develop large scale PEMFC systems.

A new stack design for electrochemical cells based on hydraulic compression of single cells has been suggested [22-24]. With this stack design, PEMFC systems can be constructed which contain a variable defined number of single cells electrically connected in series over flexible copper ribbon cables. These cells are arranged in flexible pockets within a pressure tank filled up with fluid. At operation cells are fixed within the tank by pressurizing this fluid. However, e.g., for maintenance reasons, pressure can be released so that changing cells is possible. During operation a hydraulic medium surrounds, each cell of the presented stack design in a way that surface pressure on each cell is identical. The hydraulic pressure within the stack is controlled by a pneumatic actuator. The system’s pressure control keeps a constant pressure level. This is necessary at e.g., start-up procedure, when stack temperature rises to about 70 °C to 80 °C (operating temperature). Furthermore, temperature of the whole stack can be controlled by cooling the hydraulic medium. This guaranties homogeneous temperature distribution for each cell as well as protection against overheating. Waste heat recuperation for a CHP (combined heat and power) system can easily be realized. Hence, operation conditions with this type of stack are nearly ideal, if the parameters of the delivered process gases are constant. In order to deliver gases with identical physical characteristics to each fuel cell, these are connected in parallel to temperature controlled manifolds (for O2/air as well as for H2). Cells have clips which connect them to the gas manifolds (Fig. 2). However, distribution of necessary gases in series is also possible with another kind of manifold which joins a cell’s outlet with the next cell’s intake. Due to unfavorable pressure drop after each cell, delivering the gases in series is limited to only few cells. Two fuel cell stack prototypes with hydraulic compression have been realized so far. Water is used

Proof of Co oncept of a Novel N PEM Fu uel Cell Stack k Design with Hydraulic Co ompression

7777

matic drawing of a modular PEMFC P stack with Fig. 2 Schem four cells.

Fig.. 3 Photograaph of a PEM MFC stack with w hydraulicc com mpression.

for cell com mpression as teemperature liimit for PEM MFCs is about 80 °C. Howeveer, high tempperature PEM MFC systems (up to 200 °C) caan be realizedd by using therrmal stable oils innstead. The fiirst prototypee A is construucted with three single fuel cells with an active cell areea of 20.25 cm² (445 mm × 45 mm) m and a foour channels flow f field. Photoggraph of this small scale module m is shhown in Fig. 3. The seconnd prototype B is a stackk with four cells. c Each cell haas an active ceell area of 1822.35 cm² (93 mm × 197 mm). Flow field is different duee to rectangle pole plate design. Furthermoree, flow field is i designed by 13 parallel chaannels. Electrrical contactss are realizeed at back side off the stack.

the small protottype (four chhannels flow w field, GDL L poccket geometryy, sealing). Uniform U currrent distributtion is expeected due too hom mogeneous compression c and heat traansfer to thee hyd draulic mediuum. Regardiing the inveestigation off inveestigation onn current denssity distributiion, a speciall polee plate has been develooped for the small scalee PEM MFC prototyppe which offfers eight active segments.. Flow w field desiign (four chhannels flow field, GDL L poccket geometryy, sealing) is identical to standard s polee plattes for proototype A. Insulation between b thee segm ments is acchieved by eepoxy resin. The activee geo ometrical cell area is decreeased due to epoxy e framess arou und each seggment. One seegment has a geometricall areaa of 2 cm² ressulting in an active cell arrea of 16 cm²² (Fig g. 4). Each segment iss connected to a shuntt resiistance with 0.1 0 ohm. Voltage drop att operation iss meaasured. A typpe K thermoocouple is in ntegrated intoo each h segment. Teemperature is monitored sim multaneously.. In n order to prroof the noveel modular deesign conceptt for PEMFC staccks based onn hydraulic compression, c , opeeration charaacteristics hhave been deetermined att labo oratory condditions. Stackk housing, cell slots andd polee plates for thhe presented prototypes with w hydraulicc com mpression arre manufactuured at the laboratory’ss worrkshop. For both b stacks poole plates aree milled from m

3. Experim ments One mainn advantage of the presented novel stack s s cell comppression duee to design is homogeneou h hydraulic staack pressure. In order to deetermine presssure distribution over the acttive cell area, investigatioon is carried outt with pressure sensittive film from f ® FUJIFILM . For qualitattive determinnation a film type Prescale Supper Low Presssure LLW (66-25 bar) is used. u Pressure diistribution iss compared to a typiccally compressed PEMFC staack (four tiee bolts, 20 mm pressure platte) with identtical flow fieldd geometry as for

778


ogged into a gases are detected. All meaasures are lo ntral data baase. Air and hydrogen are a deliveredd cen with hout additionnal humidificaation. Air stoichiometry iss set to λ = 4 which means foour times mo ore oxygen iss deliivered to thee fuel cell as would be needed for ann entiire reaction. The T λ-control is implementted in the testt ben nch’s PLC (prrogrammable logic controlller).

4. Results R and d Discussion 4.1 Pressure Disstribution

matic drawing of a segmenteed pole plate (eeight Fig. 4 Schem segments).

graphite (Scchunk Kohlennstofftechnikk type FU44113/7) due to its higgh electrical and a thermal conductivity. c MEAs foor the fuel cell prototyypes have been b prepared by commerciallly available components c f from Gore®. CC CM type Prim mea® FCM and GDL type Carbel® GDM G CNW220B have been b used. Cell assembly is handmade. Both staccks are charracterized acccording to their t polarization behavior in a full autom mated test beench with proceess media treatment and a monitorring. Polarization curves for thhe presented fuel fu cell stackss are determined by an electronically conttrolled load type Höcherl & Hackl H ZS18066NV at constaant current mode. m Load currennt is increaseed step wise (1 A for 2 s per step). Currennt and correspponding voltaage are measuured simultaneouusly. Stacks aree investigatedd in H2/air operation o at rooom temperature. Stacks havee start-up conddition. MEAss are conditioned by fast cyclinng proceduree according too the Gore® data sheet. H2 witth high purityy 6.0 is delivvered with slightlyy overpressuure of 0.2 bar. b Hydrogeen is circulated byy a diaphragm m pump from the anodic ouutlet to the anodic intake in order to savve fuel. Onlyy the amount of consumed H2, is fed to thee anodic gas path. p Ambient air is delivered to t the cathode by a diaphrragm pump. Masss flow, temperrature and preessure of the feed

n order to visualize ceell compresssion due too In hyd draulic mediuum, pressure distributio on has beenn anaalyzed by the aid of pressuure sensitive film. f Pressuree disttribution for each fuel ceell within pro ototype A iss com mpared with a mechanicaally compressed fuel celll stacck (Fig. 5 [255]). Itt becomes obvious tthat, the mechanically m y com mpressed cells (Figs. 55a-5c) show w a typicall inho omogeneous pressure disttribution acco ording to thee fourr tie bolts of o the investtigated stack k, via whichh com mpression forrces are inducced. This resu ult underliness thatt, homogeneoous compression is a seriou us issue evenn for small PEM MFC stacks w with only few fe cells. Inn con ntrast to the mechanically m ccompression,, the pressuree disttribution withhin the hydraaulically com mpressed testt cellls is homogenneous (Figs. 5d-5f). Thiss nearly ideall com mpression cann be explainedd by the pressure medium m whiich surroundds each cell of the noveel test stackk com mpletely. Thherefore, at any point within thee presssurized stack, the same ppressure forcces affect thee sing gle fuel cells. 4.2 Determinatioon of Currentt Density and d Temperaturee Disstribution Current C dennsity distrib ution and temperaturee disttribution meaasurements are carried out for 15 min att con nstant current (4, 8 and 16 A A). Every seccond, voltagee drop p over the shhunt is measuured automattically by thee aid of an analogg input moduule type Beck khoff EL30688 (0-1 10 V, 12 bit resolution). Current density for eachh segm ment is deterrmined from Ohm’s law. Temperaturee


(a)

( (b)

(c)

(d)

( (e)

(f)

7799

(g)

Fig. 5 Analyysis of pressuree distribution over the activee cell area by pressure p sensittive film ((a)-(cc): mechanicall compression;; (d)-(f): hydraaulic compressiion; (g): measu uring scale).

is measuredd by a typpe K thermocouple module Beckhoff EL L3318 (eight channels, 0.1 °C resolutiion). Results of cuurrent densityy and temperature distribuution measuremennts are presennted in Figs. 6 and 7. Relaative deviation frrom arithmetiic average current densitty is given in per cent. Segmentss No. 4 and Noo. 8, which arre situated near to the gas outlet, show neggative deviatiion from aveerage current denssity for all thrree measurem ments. Especially, current density of segm ment No. 4 is i lowest forr all measuremennts carried out. This may be b an evidencee for gas flow differences d w within the floow field at this specific segm ment, as parttial pressure is i decreased with w the length off the flow fielld (due to gass consumptioon). However,, it becomes obvious thatt, current dennsity over the flow w field is hoomogeneouslyy distributed as a result of homogeneous h s operation conditions. For constant cuurrent operatiion at 4 A, current dennsity determined for each seggment differss from -4.8% % to +2% from the average value of 0.25 0 A·cm-2. For operation att 8 A (averaage of 0.5 A·cm A -2) and 16 A (average of 1 A·cm-2), thee measured deviation d is -77.7% to +4.3% annd -7.1% to +44.6%, respecttively. As a resuult of coolinng by the hyydraulic meddium temperature for each segm ment is nearlyy identical. Mean M

mperature of each e segment measured sim multaneouslyy tem to current c densityy distributionn is in the rang ge of 33.0 °C C ± 0.8 0 °C at connstant currentt operation with w 4 A. Forr opeeration at 8 A as well as 16 A, the mean m segmentt tem mperature is 33.8 °C ± 0.6 °C and 40.2 °C ± 1.2 °C,, resp pectively. Ass a result of inncreasing, waaste heat celll tem mperature rises with higherr load currentt. 4.3 Polarizatioon Behavior of Presentted PEMFC C Pro ototypes Fig. F 8 shows polarization p ccurves of prottotypes A andd B. Current C density is presentted as a funcction of stackk volttage as well as a cell as of vvoltage of the first cell off each h stack (U1).. Open circuiit voltage of both b fuel celll prottotypes is in the range off 0.9 V for each e cell. Forr prottotype A 2.7 V and, respectively, for prototype B,, 3.7 V has been measured in open circuit. At fuel celll opeeration, voltaage drops nnearly linearrly with thee app plied load up to t the point of maximum power p output.. Thiis result from m the internal rresistance as well as from m ioniic membrane losses. For F nearly am mbient conditiions (stack teemperature att 30 °C, ° operationn with un-hum midified ambiient air) MPP P (maaximum pow wer point) off fuel cell pro ototype A iss foun nd for an outpuut voltage of 11.15 V (27 W//0.43 W·cm-2)..


2

3

4 5 Segment

6

7

8

4.6 0.4

1.6

-1.2 -4.4

-2.8

-2.8 -7.7 -7.1

-10 1

-1.1 -0.6 -4.8

-8 0.0

0.1 1.0

2.5 -1.9

-2

-6

0.2

3.4

4.1 0.1

0

-4

0.243

0.1

2

Deviation (%)

0.494

2

1.1

6

4.3

1.046 0.517 0.254

0.500

0.496

0.243

0.238

0.251

0.462

0.491 0.250

0.250

0.510

0.522 0.255

A·cm-2

0.4

1.00 A·cm-2

8

4

0.8 0.6

0.50 A·cm-2

10

2.2

1.0

0.25 A·cm-2

0.956

0.929

1.025

1.041

1.022

1.2

1.00 A·cm-2

1.010

0.50 A·cm-2

0.25 A·cm-2

0.994

780

1

2

3

4 5 Segment

6

7

8

(a) (b) Fig. 6 (a) current density distribution and (b) deviation from arithmetic average current density for 4, 8 and 16 A load current. T @ 0.25 A·cm-2

44

-2

38.8

40.8

38.8

40.1

40.5 39.0

40

41.8

T @ 1.00 A·cm-2

42

34.0 34.4

32.5 33.4

32.1

33.3 34.0

33.1

34.1 33.0

34

33.0 33.6

36

34.0 34.7

38

32.2 33.1

T (°C)

41.5

T @ 0.50 A·cm

32 30 1

2

3

4 5 6 7 8 Segment Fig. 7 Temperature distribution for 4, 8 and 16 A load current.

Maximum power output of fuel cell prototype B is 290 W (0.38 W·cm-2) at an output voltage of 1.92 V. Although active cell area is different by a factor of nine specific maximum power output is in the same range. As linear upscaling of a pole plate design has not been possible

for prototype B, slightly design differences have been necessary (e.g., channel length, channel diameter or channel number). Therefore, differences in power output of prototypes A and B occur mainly due to different flow field designs. Furthermore, increase of resistance losses is likely due to higher current of prototype B (150 A @ MPP). However, increasing diameter of the connecting copper cables can be an appropriate solution. Obtained power density for the presented prototypes is similar to specific power output reported for commercially available MEA based PEMFC test cells with mechanical cell compression [26-28]. Further improvement for PEMFC stacks with hydraulic compression is expected by the optimization of pole plates, especially, the flow field. However, the novel stack design offers specific advantages. Very large cells with any planar shape can be realized. Waste heat recuperation can easily be maintained (for higher operation temperature). Furthermore, each cell can be exchanged due to the modular stack construction with separate slots.


4.0

U-I stack A

U-I stack B

U1-I stack A

U1-I stack B

P-I stack A

P-I stack B

781

0.5

3.5

U (V)

2.5

0.3

2.0 0.2

1.5 1.0

P (W·cm-2)

0.4

3.0

0.1

0.5 0.0

0 0.0

0.2

0.4

0.6 I (A·cm-2)

0.8

1.0

1.2

Fig. 8 Polarization curves on cell level (U1) as well as on stack level for PEMFC prototypes A and B.

The influence of hydraulic compression on PEMFC operation characteristic in long term tests is matter of a prospective study.

5. Conclusions Results indicate that, upscaling the presented small cell geometry for PEMFC prototype A to a factor of nine decreases specific power output of just 12%. Furthermore, current density distribution is homogeneous (about ±7% over the active cell area) due to hydraulic compression. It is expected that, current density for the larger prototype is also well distributed. It can be concluded that, upscaling effects known for state of the art PEMFC stacks are of minor importance due to even cell compression and cooling by the surrounding hydraulic medium. Very large planar cell geometries can be realized. Special cooling plates which are necessary in large state of the art PEMFC stacks are not needed for the novel stack design. Increasing number of cells from three to four does not affect specific power output significantly. This effect occurs due to nearly identical operation condition for each cell of a stack. Therefore, it can be

concluded that, designing large scale stacks with any number of cells is possible. This may attract PEMFC based stationary applications in the near future. Few design changes are needed to convert the presented stacks into energy systems with waste heat recuperation, e.g., a CHP system for household heating can operate by heat exchanging a PEMFC stack’s inherent cooling water which is also used for cell compression. The possibility to exchange single cells of the stack may be useful for maintenance. Cells which are degraded after some thousand operating hours can be replaced without affecting the housing. For prospective stationary applications with a lifetime of several years, the possibility to maintain the stack could be an advantage for the presented novel stack design with hydraulic compression.

Acknowledgments The presented study has been carried out in the scope of research projects funded by the public, financially supported by the ministry of innovation, science and research of North Rhine Westphalia, Germany, within


782

the program “Ziel2.NRW EFRE (Regionale Wettbewerbsfähigkeit und Beschäftigung)” (funding number: 005-1111-0010) as well as by the German Federal Ministry of Economic Affairs and Energy within the program “Zentrale Innovation Mittelstand” (funding number: KF2687002ZG3).

References [1]

[2] [3] [4]

[5] [6]

[7]

[8]

[9]

German Federal Ministry of Economic Affairs and Energy, German Federal Ministry for the Environment, Nature Conservation, Building and Nuclear Safety. 2010. “Energiekonzept.” German Federal Ministry of Economic Affairs and Energy, German Federal Ministry for the Environment, Nature Conservation, Building and Nuclear Safety. Accessed May 17, 2015. http://www. bundesregierung.de/ContentArchiv/DE/Archiv17/_Anlag en/2012/02/energiekonzept-final.pdf?blob=publicationFil e&v=5. Stolten, D. 2010. Hydrogen and Fuel Cells. Weinheim: Wiley- VCH, 207-26. Stolten, D. 2010. Hydrogen and Fuel Cells. Weinheim: Wiley-VCH, 121-48. Ballard. 2010. “Ballard Tests 1 MW PEM Power Plant.” Ballard. Accessed May 17, 2015. http://www. h2euro.org/2010/ballard-tests-1-mw-pem-power-plant. Stolten, D. 2010. Hydrogen and Fuel Cells. Weinheim: Wiley-VCH, 755-86. Smolinka, T., Günther, M., and Garche, J. 2011. “Stand und Entwicklungspotenzial der Wasserelektrolyse zur Herstellung von Wasserstoff aus Regenerativen Energien.” NOW GmbH. Accessed May 17, 2015. http://www.now-gmbh.de/fileadmin/user_upload/RE_Pub likationen_NEU_2013/Publikationen_NIP/NOW-StudieWasserelektrolyse-2011.pdf. Klug, K. 2013. “Wasserstoff als Energiespeicher.” Presented at the 13th Annual Meeting of the Hydrogen and Fuel Cells Network North Rhine-Westfalia, Düsseldorf, Germany. Accessed May 17, 2015. http://www.brennstoffzelle-nrw.de/fileadmin/daten/jahres treffen/2013/11_Klug_WF-HS-GE_Windstrom-Herten_ NBW20131114.pdf. E.ON. 2013. “E.ON Inaugurates Power-to-Gas Unit in Falkenhagen in Eastern Germany.” E.ON. Accessed May 17, 2015. http://www.eon.com/de/presse/pressemitteilungen/presse mitteilungen/2013/8/28/eon-nimmt-power-to-gas-pilotanl age-im-brandenburgischen-falkenhagen-in-betrieb.html. Zhang, J. 2008. PEM Fuel Cell Electrocatalysts and Catalyst Layers Fundamentals and Applications. London: Springer Verlag, 1-88.

[10] Heinzel, A., Mahlendorf, F., and Roes J. 2006. Brennstoffzellen Entwicklung, Technologie, Anwendung. Heidelberg: C.F. Müller Verlag. [11] Hamann, C., and Vielstich, W. 1998. Elektrochemie. Weinheim: Wiley-VCH. [12] Barbir, F. 2013. PEM Fuel Cells Theory and Practice. Waltham: Elsevier. [13] Zhou, P., Lin, P., Wu, C. W., and Li, Z. 2011. “Effect of Nonuniformity of the Contact Pressure Distribution on the Electrical Contact Resistance in Proton Exchange Membrane Fuel Cells.” International Journal of Hydrogen Energy 36 (10) 6039-44. [14] Asghari, S., Shahsamandi, M. H., and Ashraf Khorasani, M. R. 2010. “Design and Manufacturing of End Plates of a 5 kW PEM Fuel Cell.” International Journal of Hydrogen Energy 35 (17): 9291-7. [15] Asghari, S., Mokmeli, A., and Samavati, M. 2010. “Study of PEM Fuel Cell Performance by Electrochemical Impedance Spectroscopy.” International Journal of Hydrogen Energy 35 (17): 9283-90. [16] Garcia-Salaberri, P. A., Vera, M., and Zaera, R. 2011. “Nonlinear Orthotropic Model of the Inhomogeneous Assembly Compression of PEM Fuel Cell Gas Diffusion Layers.” International Journal of Hydrogen Energy 36 (18): 11856-70. [17] Park, S. 2013. “Simulation and Experimental Analysis of the Clamping Pressure Distribution in a PEM Fuel Cell Stack.” International Journal of Hydrogen Energy 38 (15): 6481-93. [18] Qiu, D., Yi, P., Peng, L., and Lai, X. 2013. “Study on Shape Error Effect of Metallic Bipolar Plate on the GDL Contact Pressure Distribution in Proton Exchange Membrane Fuel Cell.” International Journal of Hydrogen Energy 38 (16): 6762-72. [19] Wang, X., Song, Y., and Zhang, B. 2008. “Experimental Study on Clamping Pressure Distribution in PEM Fuel Cells.” Journal of Power Sources 179 (1): 305-9. [20] Xing, X. Q., Lum, K. W., Poh, H. J., and Wu, Y. L. 2010. “Optimization of Assembly Clamping Pressure on Performance of Proton-Exchange Membrane Fuel Cells.” Journal of Power Sources 195 (1): 62-8. [21] Pharoah, J. G., and Burheim, O. S. 2010. “On the Temperature Distribution in Polymer Electrolyte Fuel Cells.” Journal of Power Sources 195 (16): 5235-45. [22] Brodmann, M., Greda, M., Roth, J., and Mutascu, C. 2009. Vorrichtung Zur energieumwandlung, insbesondere brennstoffzellenstack Oder elektrolyseurstack. German Patent DE102,009,057,494 A1, filed December 10, 2009, and issued June 16, 2011. [23] Brodmann, M., and Greda, M. 2012. “Taschenkonzept Schafft Neue Freiheitsgerade Modulare Stacks Mit Austauschbaren Verschleißteilen.” Hzwei 12 (4): 14-5.

Proof of Concept of a Novel PEM Fuel Cell Stack Design with Hydraulic Compression [24] Rost, U., Brodmann, M., Şerban, V. A., Mutascu, C., Roth, J., and Zekorn, B. 2012. “State of the Art PEM Fuel Cells with a Focus on a Modular Fuel Cell Stack with Haydraulic Compression.” Scientific Bulletin of the “Politehninica” University of Timisoara Romania Transactions on Mechanics 57 (71): 1-6. [25] Fujifilm. “Instruction Manual Fuji Prescale Film.” Fujifilm Europe GmbH. Accessed September 1, 2015. http://www.fujifilm.eu/fileadmin/products/prescale/media /Two_Sheet_Type_for_Super_Low_Pressure_klein.pdf. [26] Yan, W. M., Wang, X. D., Lee, D. J., Zhang, X. X., Guo, Y.

783

F., and Sud, A. 2011. “Experimental Study of Commercial Size Proton Exchange Membrane Fuel Cell Performance.” Applied Energy 88 (1): 392-6. [27] Wang, X. D. 2008. “Effects of Operating Temperatures on Performance and Pressure Drops for a 256 cm2 Proton Exchange Membrane Fuel Cell: An Experimental Study.” Journal of Power Sources 185 (2): 1040-8. [28] Roth, J. 2011. “Optimization of PEMFCs by Applying Hot Pressed MEAs Using Different GDL and Favourable Driving Scenarios for Operation without Humidification.” Master thesis, Fachhochschule Gelsenkirchen.

D


DAVID

PUBLISHING

Performance of R141b Ejector with Thermal Storage for Solar Air Conditioning Chakri Sripanom and Sarayooth Vaivudh SERT (School of Renewable Energy Technology), Naresuan University, Phitsanulok 65000, Thailand Received: July 07, 2015 / Accepted: August 05, 2015 / Published: September 30, 2015. Abstract: This paper was designed to determine the performance of the R141b ejector includes analysis in economics. The first step is to determine the operating condition and ejector geometry through computer calculation program. That found at the generator temperature 84 °C and evaporator temperature 8 °C, diameter of nozzle throat is 2 mm, diameter of nozzle exit is 8 mm, diameter of mixing chamber inlet is 25 mm, diameter of constant area section is 8 mm. Area of evacuated solar collector is 10 m2, thermal storage tank size is 0.33 m3, cold thermal storage size is 2.3 m3. The entrainment ratio and COP (coefficient of performance) of computer calculation program are 0.295 and 0.235, respectively. The second step ejector is fabricated and equipped to solar ejector refrigeration system, it is found that, average COP is 0.265. The economics analysis of solar ejector cooling system are invested in the investment cost was 158,158 baht. When calculating payback period was 7.73 years, the return value on a NPV (net present value) was 60,872.63 baht of lifetime of the system throughout a period of 15 years, and IRR (internal rate of return) is 13.57%. Key words: Solar energy, ejector, energy storage system.

1. Introduction Currently, global warming has been causing an increase in global temperature. As a result, there has been an increase in air conditioning demand as well, which can be observed from the worldwide growth of air conditioning market. The growth rate assessment is at about 17% [1]. Approximately 15% of electronic products worldwide are mostly air conditioning and refrigeration systems [2]. In Thailand, electricity consumption can be categorized into three main parts: 40% consumed by industry sector, 35% consumed by business sector, and 25% consumed by residential sector. Over 50% (approximately 50 × 103 GWh/year) of electricity cost is consumed by air conditioning systems. Electronic productions from power plant thusly have to support the air conditioning system at about 600 MW each year [3]. The common cooling technologies being used are vapor compressed refrigeration systems that are made to be compatible Corresponding author: Chakri Sripanom, research field: renewable energy. E-mail: [email protected].

with solar energy in order to extract electricity from solar cells, which are desiccant system, passive cooling system, absorption refrigeration system, and ejector refrigeration system. According to literature review conducted by many researchers, an ejector refrigeration system has been studied and developed due to many of its advantages, for example, the simplicity of the installation, design and operation, the relatively low cost in heat operation system, the possible usage in residential sector, and it can be used as an environment-friendly refrigerator as well. However, the ejector refrigeration system is prone to be lacking in energy storage system. This present study will design ejector refrigeration system using R141b as refrigerant. Designed parameters which are obtained from program for fabricating the system are such as condenser pressure and temperature, optimal mass flow rate of entrained fluid, load required at condenser, generator and evaporator, ejector performance and COP (coefficient of performance). This work is to develop a residential

Performance of R141b Ejector with Thermal Storage for Solar Air Conditioning

785

solar ejector refrigeration system with an energy storage system for air conditioning—the performance analysis and economic feasibility.

2. System Design 2.1 Ejector Design Fig. 1 [4] shows geometry and sections of ejector, they are consist of three main parts: a suction chamber, a constant area and mixing chamber and a diffuser. When the primary flow goes through a converging-diverging nozzle in the ejector, vapor is drawn from the evaporator. The secondary flow is accelerated to a high velocity vapor stream and reaches subsonic velocity. Mixing starts at the onset of the constant-area section (section y-y, hypothetical throat (Fig. 1)). In section y-y, both streams develop uniform pressure; choking of the secondary flow occurs. A combined stream develops into a transient supersonic stream and shocks at section s-s. The velocity of the mixing fluid must be high enough to increase the pressure after deceleration in the diffuser to a suitable condensing pressure. Computer calculation program of ejector refrigeration system has been developed to obtain parameters which are used to design and fabricate the refrigeration system. The flow chart of simulation can be expressed as Fig. 2, and assumption for computer simulation program as efficiency of nozzle ( ) is 0.85, efficiency of suction (  ) is 0.85, efficiency of diffuser ( ) is 0.85 [5], generator temperature ( ) is 84 C, condenser temperature ( ) is 28 C, evaporator temperature ( ) is 8 °C, specific heat of R141b at 25 C ( ) is 1.16 kJ/(kg·K). 2.2 Evacuated Tube Solar Collector A solar system unit is used to supply the heat to the generator as a major source of energy for the ejector cooling system. Hot water is used in this research generate from evacuated tube solar collector. Evacuated tube collectors use liquid-vapor phase change materials

Fig. 1 Ejector geometry and sections.

to transfer heat at high efficiency. The useful energy collected from a collector can be obtained from Eq. (1): (1) It is defined that, the solar collector efficiency is the ratio of the useful heat gained over any time period to the incident solar radiation over the same period. The instantaneous energy efficiency of the solar collector can also be expressed in the form of the average Bliss coefficient (FRτα = 0.80) and the heat loss coefficient (FRUL = 1.5) [4] in Eq. (2): (2) 2.3 Energy Storage System The most widely used method of sensible heat storage is a well-mixed liquid storage. The transient energy balance equation for a hot liquid storage tank can be considerably simplified if the temperature in the storage tank is assumed to be uniform, which is called well-mixed storage. Assuming that, the storage tank has a uniform temperature , energy balance on a storage tank can then be expressed as: (3) where, M is the mass of media in storage tank, is the rate of heat addition by the collector, is the rate of heat removal by the load, U is the overall heat transfer collection between media in the tank, Ta is the


786

,

1

1 2 ∆

,

,

,

NO 2

,

.

,

, ⁄

2 ,

1

1

YES ,

1

1

1 2

2 1

∆

1

1

,

2

1

1

2

,

,

1 2

,

NO

2

1

2

1 1

, 2 2

1

,

1 1

1

2

1

2

1 1

YES .

1

1

,

2

·

·

·

·

2

· 1

2 1

2

2

1

1

2

1 2

1 2

1

or NO

if if

, ,

∆ ∆ YES

Fig. 2 Flow chart of the ejector simulation.

1

. .


ambient at temperature, and is the surface area of the storage tank. For a reasonable time period ∆ , the rates of heat additional and removal are assumed to be constant, can be written for each time interval as: ,

,

∆

,

(4)

2.4 Performance of Ejector Refrigeration System The performance of ejector refrigeration system can be expressed in term of coefficient of performance is defined as the heat flow rate of evaporator in Eq. (5): and generator (5) Terms of

and

can be expressed in form of

enthalpy difference and system performance is also written as:



(6)

Ejector performance is measured by the entrainment ratio, which is defined as mass flow ratio of secondary flow to primary flow: 

Solar System Unit

Fig. 3 Schematic diagram of system.

(7)

787

3. Experimental Investigation of Solar Ejector Refrigeration System The experimental practice of the R141b ejector refrigeration system uses the operating condition and ejector geometry via the result of the simulation program. The experiment was tested from 9:00 a.m. to 16:00 p.m. at the School of Renewable Energy Technology, Naresuan University, Thailand. Fig. 3 shows a schematic diagram of system. The evacuated tube solar collector provided hot water in to storage tank before start the test, water temperature increase effect from solar radiation which shown in Fig. 4. After that, turn on the hot water pump for it to circulate to the generator to exchange with R141b. Then, turn on the water pump by opening cooling water valve to let water flow through the condenser. After the R141b is increased to the set point, the vapor enters to the primary nozzle of the ejector by opening the balancing valve manually when the R141b vapor reaches the temperature of 84 C, pressure at 4-5.5 bars. After that, the temperature of generator is steady, by opening balancing valve, the secondary flow is entrained to ejector suction and mixed with the primary flow at mixing chamber. In order to continuously make the system operate, the

Ejector Cooling Unit

Cooling Load Unit

Temperature (°C)


Radiation (W/m2)

788

Radiation Water temperature (°C)

Time Fig. 4 Temperature of water at various times with radiation.

refrigeration pump should be turned on to feed the refrigerant from the receiver tank to the generator. Storage tank, the temperature output from collector must be higher than 80 C, given collector efficiency is 0.5 approximately, solar radiation is 715 W/m2.

4. Results and Discussion The result of the ejector geometry is calculated as follows. Nozzle throat diameter is 2 mm, nozzle inlet diameter is 7.75 mm, nozzle exit diameter is 8 mm, mixing chamber inlet diameter is 25 mm, mixing chamber outlet diameter is 8 mm, diffuser outlet diameter is 17.8 mm, mixing chamber length is 25 mm, diffuser chamber length is 56 mm. Calculated the mass flow rate is 0.0152 kg/s, solar collector area of 10 m2 for heated water 328.13 kg in the 0.33 m3, for supply the hot water from storage tank to ejector cooling system, to produce the chilled water 8-10 C, stored to cold thermal storage, for provide chilled water to air condition, in this case, flow rate is 0.104 kg/s, remove heat from water of 2,257 kg in the 2.3 m3 for 6 h at 25 C. For cooling system capacity is 3.5 kW. In case constant evaporator temperature and varied generator temperature range from 80 C to 90 C, the mass flow rate of primary fluid is increase 0.0051 kg/s to 0.0064 kg/s, while mass flow rate of secondary fluid is 0.0016 kg/s, since it is lead cause to reduce the

entrainment ratio and COP. In addition to high generator, temperature leads to higher angle of expansion of primary fluid from primary nozzle. Found increase generator temperature, the saturated vapor primary will spread and obstruct the secondary fluid is entrained to mixing chamber section. This reason to entrainment ratio and COP is decrease. That is shown in Fig. 5. The experimental study varied generator temperature range from 80 C to 90 C, constant evaporator temperature. Entrainment ratio and COP was decreased. As shown in Fig. 5, the mass flow rate of primary fluid is 0.0233 kg/s, mass flow rate of secondary fluid is decrease from 0.0043 kg/s to 0.0029 kg/s. The comparisons COP of design result and experimental result, as shown in Fig. 6, trend graph is similar. Found increase generator temperature, entrainment ratio and COP is decrease. Because of the saturated vapor primary will spread and obstruct, the secondary fluid is entrained to mixing chamber section. The system performance was shown and compared the results with experimental result. The deviation of COP has average 0.182%. The relationship among temperature of condenser evaporator and chilled water, with time, from Fig. 7, it can be seen that, evaporator and chilled water temperature curve give the same trend. Evaporator and


789

Fig. 5 Relationships of the COP, entrainment ratio and mass flow rate with the generator temperature from design result.

Fig. 6 Comparison of the performance results of design and experimental with the generator temperature.

0.277 at 16:00 p.m.. That is shown in Fig. 8, the change of COP can be explained as the change in evaporator temperature. COP is higher in the first period as higher evaporator temperature and COP is decrease as lower evaporator temperature at longer time. This result is given the same as Pollerberg, et al. [7] and Yen, et al. [8] Furthermore, the results show condenser temperature significantly affects COP as well. That is, as condenser temperature decreases, COP increases which this result is similar to Petrenko, et al. [9] and Pollerberg, et al. [7]. 4.1 Economic Analysis

Fig. 7 Temperature of condenser, evaporator and chilled water in storage tank.

chilled water temperatures are lowest at 9 C and 12 C, respectively. Allouche, et al. [6] studied the change of evaporator temperature by predicting from CFD (computational fluid dynamics) program which the results are similar to this research. The average COP obtained is 0.265 and the refrigeration system performance yields the highest COP of 0.301 at 11:15 a.m., and COP reduces to be

The economics of solar energy systems are particularly complex with many inevitable uncertainties due to several factors. The principal reason to use the solar energy for heating or cooling is the cost reduction. Therefore, an economic analysis must be carried

Fig. 8 The COP of system, ambient temperature and condenser temperature at the various times.


790

out to determine whether a particular solar system is economically advantageous for a particular project. The solar ejector cooling system is invested in the investment cost 158,158 baht. Loaning interest rate 8.7% [10], when calculating payback period is 7.73 years. The return value on a NPV (net present value) is 60,872.63 baht of lifetime of the system throughout a period of 15 years, and IRR (internal rate of return) is 13.57%.

5. Conclusions Solar ejector refrigeration system has been designed and experimentally fabricated. Computer program of R141b developed to study effect of operating condition and ejector geometry on ejector performance and COP. From testing results of solar ejector refrigeration system, that found, entrainments ratio, average COP and chilled water temperature are 0.295, 0.265 and 10 °C, respectively. Thermal storage tank size is 0.33 m3, cold thermal storage tank size is 2.5 m3, solar collector area is 10 m2, and economics analysis of invested in the investment cost is 158,158 baht, payback period is 7.73 years. The return value on a NPV is 60,872.63 baht of lifetime of the system throughout a period of 15 years, and IRR is 13.57%.

References [1]

Henning, H. M. 2007. “Solar Air-Conditioning and Refrigeration.” Task 38 of the IEA (International Energy Agency) Solar Heating and Cooling Programme. Accessed May 10, 2011. http://lmora.free.fr/task38/pdf/matin/Henning.pdf.

[2]

Chunnanond, K., and Aphornratana, S. 2004. “Ejectors: Applications in Refrigeration Technology.” Renewable and Sustainable Energy Reviews 8 (2): 129-55. [3] Sun, D. W., and Eames, I. W. 1996. “Performance Characteristics of HCFC-123 Ejector Refrigeration Cycles.” International Journal of Energy Research 20 (10): 871-85. [4] Pridasawas, W. 2006. “Solar-Driven Refrigeration System with Focus on the Ejector Cycle.” Ph.D. thesis, KTH (Kungliga Tekniska Högskolan). [5] Huang, B. J., Chang, J. M., Wang, C. P., and Petrenko, V. A. 1999. “A 1-D Analysis of Ejector Performance.” International Journal of Refrigeration 22 (5): 354-64. [6] Allouche, Y., Bouden, C., and Varga, S. 2014. “A CFD Analysis of the Flow Structure inside a Steam Ejector to Identify the Suitable Experimental Operating Conditions for a Solar-Driven Refrigeration System.” International Journal of Refrigeration 39 (March): 186-95. [7] Pollerberg, C., Heinzel, A., and Weidner, E. 2009. “Model of Solar Driven Steam Jet Ejector Chiller and Investigation of Its Dynamic Operational Behavior.” Solar Energy 83 (5): 732-42. [8] Yen, R. H., Huang, B. J., Chen, C. Y., Shiu, T. Y., Cheng, C. W., Chen, S. S., and Shestopalov, K. 2013. “Performance Optimization for a Variable Throat Ejector in a Solar Refrigeration System.” International Journal of Refrigeration 36 (5): 1512-20. [9] Petrenko, V. O., Huang, B. J., and Shestopalov, K. O. 2011. “Innovative Solar and Waste Heat Driven Ejector Air Conditioner and Chiller.” In Proceedings of the 2011 2nd International Conference on Environmental Science and Technology IPCBEE (International Proceedings of Chemical, Biological and Environmental Engineering), (V1-338)-(V1-343). [10] Bank of Thailand. “Loan Rates of Commercial Banks as of 19 May.” Bank of Thailand. Accessed May 19, 2015. https://www.bot.or.th/thai/statistics/_layouts/application/i nterest_rate/in_rate.aspx.

D


DAVID

PUBLISHING

Stress-Strain Behavior of Nylon-Carbon Composite Subjected to High Strain Rate Impact Loading Noori Hassoon Mohammed Al-Saadi Department of Civil Engineering, Dijlah University College, Baghdad, Iraq Received: June 24, 2015 / Accepted: August 03, 2015 / Published: September 30, 2015. Abstract: The aim of this study is to investigate the dynamic stress-strain relation for the hybrid composite (nylon +carbon). Three groups of specimens are used with different number of carbon layers. The specimens were subjected to high velocity impact with different strain rates. SHPB (split Hopkinson pressure bar) is used in this investigation. The results show that, the stress-strain relation various with the strain rate. The maximum stress and strain are proportion directly with the strain rate. Also, the results revealed that, as the number of carbon layer increased, the maximum strain decreased. Key words: High strain rate, split Hopkinson pressure bar, composite, high strain rate impact.

1. Introduction The split-Hopkinson pressure bar is among the most commonly used experimental techniques for the determination of the dynamic behavior of materials strained at rates of up to around 103 s-1. Polymer matrix composite materials and structures are finding increasing uses in high performance applications. High strain rate loading is one of the possibilities in many of such applications. The mechanical properties of composite materials may be different at high strain rate loading compared with those at quasi-static loading. This necessitated carrying out investigations on how mechanical properties of composites would change with strain rate. The widely used technique for the determination of behavior of composites under high strain rate loading is SHPB (split Hopkinson pressure bar) apparatus. The working of this apparatus is based on one-dimensional wave propagation theory in elastic bars [1]. Various studies were carried out on high strain rate behavior of unidirectional and cross-ply glass Corresponding author: Noori Hassoon Mohammed Al-Saadi, Dr., assistant professor, research field: applied mechanics. E-mail: [email protected].

composites [2, 3] and woven fabric glass composites [4] under compressive loading. Generally, it was observed that, the compressive strength and modulus increase compared with those at quasi-static loading. Various studies were also carried out on high strain rate behavior of unidirectional and cross-ply carbon composites [5-7] and woven fabric carbon composites [8] under compressive loading. It was observed that, the compressive strength and modulus increase compared with those at quasi-static loading. The objective of the present study is to determine the behavior of typical nylon-carbon/epoxy composite under high strain rate compressive loading. Force versus time plots based on strain gage signals obtained from incident bar and transmitter bar are derived and compared.

2. Theoretical Considerations The design of SHPB is based on one-dimensional wave propagation in elastic bars which deals with the motion of particles in longitudinal direction. The one-dimensional system can ideally be considered to be of infinite length and negligible diameter. The analytical relations to calculate strain rate, strain, and stress as a function of time in the specimen in SHPB

792

Stress-Strain Behavior of Nylon-Carbon Composite Subjected to High Strain Rate Impact Loading

testing are [9]:

transmitter bars, part of the pulse is reflected and the (1) d

(2)

remaining part is transmitted through the specimen to the transmitter bar. The strain gages are placed at the centers of incident and transmitter bars, respectively for providing the time-resolved measure of the signals.

(3)

Strain gage on incident bar measures both the incident

where, c is the elastic wave velocity in the bars

bar measures the transmitted pulse. The strain gages

), L is the specimen gage length, Ab is the

are installed midway of the incident and transmitter

(

cross-sectional area of the bars, AS is the cross-sectional area of the specimen, E and ρ are the Young’s modulus and density of the bars, εR is the reflected strain pulse, εT is the transmitted strain pulse and t is the time duration.

and reflected pulses, whereas strain gage on transmitter

bars to avoid overlapping of the reflected signal with the incident signal and to keep the time taken the same for both the signals to avoid time calculations during measurements of the strain. The specimen then undergoes dynamic elastic-plastic deformation. From the reflected pulse, the strain rate applied and the strain in the specimen are estimated, and the transmitted

3. Experimental Considerations

pulse provides a measure of the stress. The entire

3.1 SHPB Apparatus

strain/deformation history within the specimen can be

The compressive SHPB apparatus shown in Fig. 1 was used for the present study [8]. The main parts of the compressive SHPB apparatus are: 1—the incident bar, 2—the transmitted bar, 3—the striker bar, 4—the specimen, 5—strain gages, 6—supports, 7—momentum mass, 8—shooting gun, 9—the screw driver, 10—striker’s clamper, 11—the rigid beam. The diameter of the incident and transmitter bars is 12 mm .When the striker bar impacts the incident bar,

obtained by taking measurements along the incident and transmitter bars from strain gages with the assistance of amplifier and oscilloscope. From these signals and using one-dimensional wave propagation theory, strain rate versus time, strain versus time, stress versus time and stress versus strain history in the specimen can be determined. 3.2 Specimens

an elastic stress pulse is generated and travels along the

Specimen dimension design is one of the most

incident bar. When the pulse reaches the specimen,

important considerations in SHPB testing. In order to

which is sandwiched between the incident and

minimize the effects of the longitudinal and lateral

9

8 10

1 3

Fig. 1 Schematic of the split Hopkinson pressure bar.

2

6 5A A

4

11

5

7


793

Table 1 Specimen specifications. Group number

Lay up sequence

A

313

B

323

C

333

Number of Lamination layup method layers 3nylon + 1carbon fiber + 7 3nylon 3nylon + 2carbon fiber + 8 3nylon 3nylon + 3carbon fiber + 9 3nylon

= 588.0/s = 567.1/s = 497.0/s = 485.8/s = 468.0/s

inertia and wave dispersion within the specimen, the overall specimen dimensions are required to be small. Previous

work

in

SHPB

suggested

that,

0

2,000

4,000

6,000

8,000

the

thickness/diameter ratios of 0.5-2.0 are suitable for

Fig. 2 Dynamic stress-strain curves for (313) lay-up.

cylindrical specimen tests under compression loading [8]. Also the specimens’ size must be smaller than the incident and transmitted rods diameters. A cylindrical specimen of 10 mm diameter was chosen in this work

= 526.1/s

for all types of tested materials. Table 1 shows the

= 492.7/s

groups of the samples.

= 460.7/s = 415.0/s

4. Results and Discussion The dynamic experimental results for three tested groups are shown in Figs. 2-4. It can be seen that, the maximum stress and strain are increased with increasing the strain rate. Also, it can be indicated that, the increasing the number of layers of carbon fibers yielded higher mechanical properties. The results clearly show that, the material response is rate dependent. In general, both the slope of the stress-strain curves and the maximum stress and strain are affected. Figs. 2 and 3 show that, the groups A and B show less rate sensitivity than group C. The strain for the group C is less than for the other configurations due to increasing the volume fraction of carbon fiber. Tables 2-4 show that, the strain rate decreasing as the number of carbon layers increasing i.e., the strain increasing with increasing the carbon layers due to increasing the specimen stiffness. Also, the results indicate that, the strain rate is increasing with the impact velocity. Figs. 5-9 show the variations of stress-strain behavior with the impact velocity. It can be concluded that, the specimen with one layer of carbon has

= 383.4/s

0

2,000

4,000

6,000

8,000

Fig. 3 Dynamic stress-strain curves for (323) lay-up.

= 422.5/s = 393.0/s = 380.0/s = 367.3/s = 358.2/s 0 1,000 2,000 3,000 4,000 5,000 6,000

Fig. 4 Dynamic stress-strain curves for (333) lay-up. Table 2 Maximum stress and strain for (313) lay-up. Group number

A

588

7,500

Maximum stress (MPa) 342

567.1

7,470

321

Strain rate (1/s) Strain (micro)

497

7,000

300

485.8

6,680

282

468

6,200

255

794


Table 3 Maximum stress and strain for (323) lay-up. Group number

Maximum Strain rate (1/s) Strain (micro) stress (MPa) 526.1 6,000 388.5

B

492.7

5,750

361.8

460.7

5,690

338

415

5,678

291

383.4

5,600

267.9

10%, 13%, 7%, 14% and 18% for group B and by 28%, 30%, 23%, 24% and 23% for group C. This is means that, the rate of change in (deformation) of a material with respect to time for group C is higher than groups A and B.

Table 4 Maximum stress and strain for (333) lay-up. Group number

422.5

5,089

Maximum stress (MPa) 450

393

4,870

378

380

4,621

353

367.3

4,002

300

358.2

3,810

288

Strain rate (1/s) Strain (micro)

C

0

2,000

4,000

6,000

8,000

Fig. 7 Effect of number of carbon layer on stress-strain relation (v3 = 9.62 m/s).

0

2,000

4,000

6,000

8,000


0 1,000 2,000 3,000 4,000 5,000 6,000 7,000


0

2,000

4,000

6,000

8,000


maximum strain rate for all values of impact velocity. Compared with group A, the strain rate is decreased by

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000



5. Conclusions Based on the results were obtained, the compressive strength increases with an increase in strain rates. The maximum stress and minimum strain are observed in the group C (333). On the other hand, the strain rate is increased with the impact velocity while it found the strain rate for the first group A (313) is higher than the other groups at all impact velocities.

References [1]

[2]

[3]

[4]

Chen, W. W., and Song, B. 2011. “Split Hopkinson (Kolsky) Bar Design, Testing and Applications.” In Mechanical Engineering Series. Berlin: Springer Science + Business Media LLC. Tarfaoui, M. 2011. Experimantal Investigation of Dynamic Compression and Damage Kinetics of Glass/Epoxy Laminated Composites under High Strain Rate Compression. Brest: ENSIETA/MSN. Accessed June 24, 2015. http://www.intechopen.com. Tsai, J., and Sun, C. T. 2004. “Dynamic Compressive Strengths of Polymeric Composites.” International Journal of Solids and Structures 41 (11-12): 3211-21. Accessed June 24, 2015. http://www.elsevier. com/locate/ijsolstr. Ravikumar, G., Jayram, R. P., Makarand, J., Kiran, A.,

[5]

[6]

[7]

[8]

[9]

795

Santosh, K., and Naik, N. K. 2013. “Analytical and Experimental Studies on Mechanical Behavior of Composites under High Strain Rate Compressive Loading.” Materials and Design 44 (February): 246-55. Accessed June 24, 2015. http://www.elsevier.com/locate/matdes. Daniel, I. M., and Hasiao, H. M. 1999. “Effects of Strain Rate and Fiber Waviness on Compressive Behavior of Composite Laminates.” In Eng. and Applied Sc. Evanston: Northwestern University, 412-22. Accessed June 24, 2015. http://www.iccm-central.org. Koerber, H., and Camanho, P. P. 2011. “High Strain Rate Characterization of Unidirectional Carbon-Epoxy IM7-8552 in Longitudinal Compression.” Composites Part A 42 (5): 462-70. Accessed June 24, 2015. http://www.elsevier.com/locate/compositesa. Koerber, H., and Camanho, P. P. 2009. “Characterization of Unidirectional Carbon-Epoxy IM7-8552 in Longitudinal Compression under High Strain Rates.” Presented at the 3rd International Conference on Integrity, Reliability and Failure, Porto, Portugal. Ruaa, A. Y., and Dr. Khaldoon, A. J. 2013. “Investigation of the Effects of Fast Impact on Composite Plates Using Hopkinson Split Bar Method.” Master thesis, University of Al-Mustansiriya. Jadhav, A. 2003. “High Strain Rate Properties of Polymer Matrix Composites.” B.S. Master thesis, University of Pune.

D


DAVID

PUBLISHING

Solar-Powered Direct-Current Loads in Small Buildings Prakash Sarnobat1, 2 and Simon Lannon1 1. Welsh School of Architecture, Bute Building, Cardiff CF10 3NB, UK 2. A Flows, 53 Llwynygrant Road, Penylan, Cardiff CF23 9HL, UK Received: January 22, 2014 / Accepted: August 07, 2015 / Published: September 30, 2015. Abstract: Traditional light bulbs (e.g., incandescent, fluorescent) use too much electricity, convert very little energy into light of sufficient quality and in their production use toxic contaminants. During the last few years, a new type of light source, LED (light emitting diode) bulb, has gained increasing popularity and its costs are set to plunge even further. LED bulbs offer many advantages over traditional sources, and they can be used as a direct replacement to existing lighting. This paper will use a spreadsheet-based analysis with hourly solar data supplied by Ecotect to show that, the efficiency of LED installations can be increased when used in conjunction with photovoltaic modules, as the two generate (and use) DC (direct-current) electricity, thereby eliminating intermediate-level losses in the electronic circuitry. If a storage battery is included, the solar panels generate electricity during the times when the occupants are not necessarily using the lighting, but the stored electricity can be used to power the lighting when the energy is required. The latest results demonstrate that, a slight reduction in the required floor area to be lit allows the solar-battery-LED system to be implemented in small buildings using a storage battery size that is within the range of present commercial devices. Key words: Energy efficiency, photovoltaics, LED lighting, simulation.

1. Introduction and Background Energy efficiency has been taken on an increasingly important role, given the threat of dwindling resources and climate change. Buildings are one of the largest users of energy in the western world, and there is plenty of scope for reducing its use. Lighting contributes around 20% of the energy use, and the potential exists to half this figure. A technology that has matured during the last few years is semiconductor LED (light emitting diodes), which use DC (direct current) electricity for their operation. Another technology that has also gained popularity in the built environment in recent years is semiconductor PV (photovoltaics), which generates DC electricity directly from sunlight. The two technologies can be used in combination to potentially produce an efficient and sustainable interior lighting system. Traditional lighting systems used incandescent (i.e., Corresponding author: Prakash Sarnobat, consulting engineer, research field: building physics. E-mail: [email protected].

filament) light bulbs. Although cheap to purchase, and possessing a good light quality which mimics natural daylight, less than 5% of the electrical power is converted into light with the remainder being emitted as heat into the room. This means energy is wasted both in powering the bulb, and in the building cooling systems to extract the emitted heat from the room. A few decades ago, fluorescent lighting came on the market: It is more efficient than incandescent lighting (around 10%), but produces a rather artificial light output profile which can cause psychological discomfort to occupants, and also produces toxic substances (e.g., mercury) making their safe disposal quite problematic. LEDs were first invented in the 1960s and were mainly used in lighting applications for electronic equipment, due to them emitting a particular colour, and due to emitting light in a quite narrow beam (less than 30 degrees). Compared to old light sources, they have a much longer lifetime, typically 25 years as opposed to 5 years CFL (compact fluorescent lights)

Solar-Powered Direct-Current Loads in Small Buildings

and 2 years (incandescent). The last decade has seen unprecedented improvements, LEDs can generate warm-white light with a spectrum that is almost as good as daylight, sophisticated optics that allow light emission at both wide and narrow beam angles have been developed, and efficiencies are now at least as high as the corresponding CFL. The big breakthrough came in 2008 [1, 2] when it was shown that, it is possible to take advantage of the processes in the manufacturing of computer chips to reduce the cost of producing LEDs by up to tenfold, the typical LED now costs around three times as much as its CFL counterpart but its physical performance is at least as good as; if not better. According to Shailesh [3], the operating costs over 25 years can be reduced by 80% if the constant use of fluorescent lamps is replaced by LED lighting combined with sensors for daylight and occupancy levels. The one outstanding issue that remains is thermal management: Although less heat is produced than in other light sources, if it is not extracted away from the device then the light output will degrade [4], or even worse damage will occur. Both Parry [5] and Narendran [6] discuss how to address the issue using computational fluid dynamics. Even a single luminaire can have its spectral output programmed to be time-varying such that it can mimic the behaviour of daylight over a full day [7]. Work is currently underway to develop “Wi-fi” LED lighting, where it is hoped that, information can be transmitted using optical photons as opposed to the current method using wireless (non-visible) photons. No doubt, there will be other innovative uses for LED lighting. PV technology has traditionally been the domain of remote, off-grid systems, due to its efficiency losses when implemented in the form of a centralized, large-scale power generating plant. Improvements in performance and cost have made PV panels increasingly popular in being integrated into the building architecture, especially if they are

797

roof-mounted. Most PV are made out of single-junction crystalline silicon, and as of 2012, the average efficiency of PV modules is around 15% with a typical cost of around £2,000 per installed kW, and this is set to improve even further. One must be careful not to confuse (industrially-manufactured) module efficiency with lab efficiency, where the latter can be a lot higher. Single-juction cells have an upper limit on the efficiency of 30% due to the Shockley-Quessier limit which has its origins in statistical thermodynamics. The development of multi-junction cells is underway, and lab efficiencies of 40% have been achieved, but module efficiencies are less than 10% and costs are much higher than silicon. Solar electricity is DC, yet many of the appliances in a building are AC (alternating current), and an inverter is needed to make the required conversion, this will result in significant efficiency losses. Nevertheless, Liu [8, 9] has performed a system optimization for using PV and battery to power residential buildings in Queensland, and finds that, 6 kW roof-mounted panels with an angle of 20-25 degrees can provide nearly two-thirds of the electricity requirements. Given that LED lighting is also DC, this makes it ideal to use PV panels to power LED luminaires for interior room lighting, there are efficiency savings on not involving the use of an inverter. However, sunlight is not constant, and the lighting energy is sometimes needed when the sun does not shine, for a residential building, energy is generated during the day when the occupants are out, and it is needed during the evening when the occupants have returned. Clearly, some sort of storage is required in the form of a suitable battery. During the winter months sunlight is minimal and electricity must be drawn from the grid, and correspondingly during the summer months more electricity will be produced than is needed for the building; the excess is sold to the grid at an externally determined rate. According to tests by Sastry [10], the

798


combined PV and battery energy sizing can be reduced by up to 50% if PV modules are used to power LED lamps rather than CFL lamps. Note that, although an inverter is not needed when PV is used to directly power LEDs, it is needed when use is made of the grid. Having said that, Boeke, et. al. [11] show that, using PV to power LED lights still results in electricity savings of 15% compared to using AC mains alone. Moreover, Boeke, et. al. [12] then state that, by having an appropriate local DC electricity grid for rural businesses, the PV and battery costs (and the overall economic costs of PV-powered systems) can be significantly reduced. Patel [13] discusses a systematic procedure for calculating the overall efficiencies when various types of components are involved, and pode discusses strategies for encouraging uptake of PV-LED systems [14]. The lifetime of the battery is determined by the number of charging-discharging cycles it can undergo when discharging to 80% of the full capacity—this is said to be a “deep cycle”. The lifetime itself will depend on some key factors, including the discharge depth, rate of discharge and temperature. Shallow cycling of batteries is used for large discharge rates in order to minimize the heat generated within the battery and prolong its lifetime. In addition, the “cycle efficiency” is the percentage energy hysteresis between charging and discharging. Leadbetter and Swan [15] give a broad and up-to-date overview of the comparison of different battery types that are suitable for renewable energy systems: lead-acid (Pb-A), sodium-sulphur (Na-S), vanadium-redox (VRB), and lithium-ion (Li-Ion). Nickel-based batteries are being

gradually phased out due to cost and environmental concerns. At the present time, the most common is lead-acid, as this is a mature technology which has low upfront costs, however, it has a limited lifetime. sodium-sulphur has both low cost and long lifetime, however, heat must be provided to keep the sodium in its liquid state (costing energy) during times when the battery is not operating, else a self-discharge of 20% per day occurs. This makes it unsuitable for PV systems, as the battery requirement tends to be seasonal. Vanadium redox has a very long lifetime, but needs to operate within a narrow temperature range of 10-35 °C, requiring local climate control (and the resulting additional energy), and a large storage space is required. Lithium-ion batteries had their origins in microelectronic devices, but recently their capabilities have been scaled up to larger systems. Of the four battery types, they are the only ones that possess a cycle efficiency of close to 100%. And unlike their lead-acid counterparts, they last twice as long, and do not require regular maintenance. At the present time the initial costs are high, around 40 p/kWh this is expected to drop below 30 p/kWh in a few years, making lithium-ion batteries competitive [16]. AllCell Technologies [17] give a worked example comparing lead-acid and lithium-ion technologies, demonstrating that the latter are already better suited to hot climates. Following [15], the relevant properties of the four battery types are summarized in where for lead-acid and lithium-ion batteries the “energy cell” figures have been chosen as opposed to the “power cell” figures, as we want to avoid using large currents for LED lighting.

Table 1 Comparison of the various battery types. Cycle life Energy density (Wh/L) Daily self-discharge (%) Cycle efficiency (%) Capital cost (US$ per kWh)

Pb-A 200-1,800 50-80 < 0.5 63-90 200-600

Li-Ion 3,000+ 200-500 0.1-0.3 80-98 600-1,200

Na-S 4,500 150-250 20 75-90 350

VRB 10,000+ 16-33 Negligible 75-80 150-1,000


As the lighting load will vary during the year due to daylight, it is assumed that, a suitable MPPT (maximum power point tracking) algorithm is in operation (Ref. [18] and references contained therein). An innovative method of increasing battery charging capacity by nearly 80% has been suggested by Huang [19], which states that instead of using MPPT for the PV in relation to the load, operate at near maximum power point while using pulse width modulation to control discharging of the battery. This also has the advantage of reducing the MPPT conversion loss when an undersized load is used. It has been suggested that, using solar-angle tracking for PV systems can increase power output by up to 50% [20]. Although this is not considered in the present work, it is nevertheless being considered by the authors in ongoing work. An earlier investigation by the authors [21] of the effectiveness of using wall-mounted PV panels to power LED lighting systems for large, multi-level office buildings showed that the most important system parameters were the PV cost and efficiency, number of desk lights, feed-in tariffs and cost of the old lighting system being replaced. This article focuses on small buildings, especially residential, with PV panels mounted on a south-facing roof as opposed to the walls. Sometimes the roof has a complex shape, constraining the layout of the PV panels, and one cannot really associate any individual panel(s) as belonging to any particular room. In this case the building must be analysed as a whole. For simple roofs possessing a high degree of symmetry, then one can imagine dividing the roof space (and its PV panels) as “belonging to” a particular floor of a two-level building. Indeed, this decomposition can be applied to some terraced houses that have been converted into separate apartments, each occupying one floor. Indeed, the simulation data was obtained for a detached house, but one could also imagine that it equivalently forms part of a row of terraced houses. There are several tools on the market for both PV

799

and lighting analysis, the most popular being Homer, PV-Sys and TRNSYS (TRaNsient SYStems) for PV (which can also include the battery analysis), and Dialux, Ecotect, and IES-VE (integrated environmental services-virtual environment) for lighting (to a lesser extent IES-VE can also do PV analysis, but not batteries). At present, there is no single tool that can perform PV, battery, and lighting analysis all in one place. RetScreen is an all-round tool in this regard, but it only does monthly analysis. The spreadsheet developed for this work is intended to perform all three types of analyses based on hourly data. Work is underway to generalize the tool to also deal with hourly data obtained from thermal modelling. The aim of this work is to give a first indication of the energy requirements, such that key decisions on several “what-if” scenarios can be made during early-stage design. Once the optimum building configuration has been chosen, then detailed analysis can proceed as usual, indeed, the resulting value of the lighting energy from this can be input into the spreadsheet used in this analysis to more accurately determine the PV area and battery capacity. It is hoped that, this shall become a useful tool for architects, engineers and building managers alike. Further, the analysis is extrapolated from the present day to consider long-term horizons up to 2050, where we show that, the payback period will be a third of what it is today. Although this work focuses on the United Kingdom, the fundamental methods can be applied anywhere in the world.

2. Methodology The system under consideration is a Detached house that consists of PV panels on its south-facing roof (Fig. 1), and its interior consisting of a battery and the DC loads, in this case the LED lighting. Both commercial software and excel spreadsheet-based analyses were used. Annual hourly solar radiation data for Bristol was obtained from EcoTect, and converted into watt-hours (Wh) for PV

800


panels. Based on peak values of this data, the PV system was sized at 1.5 kW. The battery capacity was sized in relation to the daily excess of the load requirements versus the PV input, averaged over the year, and this was termed the daily deficit. In sizing the battery, a margin of 50% spare capacity was allowed to account for various losses. Its state of charge was determined by the difference between energy input from the PV, and energy extracted by the loads, including a battery self-discharge of approximately 2% per month. As the hourly self-discharge is relatively small, of the order of a fraction of a Wh, this behaviour can be assumed as linear. The state of charge on the battery is determined by the difference between the PV input and its use by the loads. If this difference is greater than the maximum capacity of the battery, then any excess is fed to the grid. Conversely, if the PV input is insufficient to power the loads, then the (hourly) deficit will be taken from the grid. An hourly profile for the room lighting was specified to be on between 7 a.m. and 9 a.m., and from 6 p.m. to 11 p.m., with the lights off outside these hours, daylight was not considered. Based on this lighting schedule, and on the PV input, annual hourly values for the state of charge, deficit and excess were calculated. From these quantities, the monthly values over the year of highest and lowest excess/deficit are obtained, and whether or not there is an annual net use of the grid. The price for using the grid is 12 p/kWh. Excess energy fed to the grid will result in a price being paid by the government to the building owner, called a FIT (feed-in tariff). The FIT gradually decreases every year, and as of early 2013, it is 16 pence per kWh for generation (irrespective of whether or not it is used locally) plus an additional export tariff of 4.5 p/kWh. This payment, in addition to the savings on the electricity bill prior to installation, can be used to offset the initial cost. The time it takes for this to happen is called the payback period, and it can

Fig. 1 Cross section of the house, with PV panels mounted on the south-facing roof.

depend on a number of factors. When describing the efficiency of a luminaire, one must only consider the wavelengths (and corresponding light energies) that are sensitive to human vision, and not anything outside this range. One therefore talks of lighting power in lumens, which, approximately speaking could be regarded as “optical watts”, and the number of lumens reaching a square metre of the working plane (which is an imaginary surface one metre above the floor) is termed lux. The ratio of lighting power to electrical power is termed luminous efficacy, or just “efficacy”. The lumen method regards the light from a luminaire as corresponding to a mathematically equivalent source that is uniformly distributed over a certain area of the ceiling, and being emitted vertically downwards over that same area of the working plane. If the luminaire has power P and luminous efficacy η, then, the total number of luminaires N that are required to produce a given lux level E at the working plane of area A is: EA N PMU where, M is the maintenance factor, and U is the utilization factor. M accounts for the degradation of the luminaire over time (e.g., due to dirt), and U describes the fraction of light from the luminaire that actually reaches the working plane.


The key determinant of the economic viability of the whole system is its lifetime costs, and in particular, the payback period taking into account inflation interest rates. The sum of the annual payments forms a geometric series, which can be evaluated in closed form. For a given capital cost of the PV-battery-LED system, annual FIT (G), market discount rate for investment (d%), annual rate of electricity price inflation (i%), and annual operating costs of the old lighting system (L), the payback period (T) in years is given by:  C i  d  ln 1  G  L   T   1 i  ln   1  d 

As the battery lifetime (12.5 years) is shorter than the PV and LED lifetime (25 years), more than one battery will be used. The cost of subsequent future batteries must be discounted to the present value using the following NPV (net present value) factor, where τ is the number of years before the replacement battery is installed: 1 NPV  1  d  Given the current economic climate, it seems reasonable to assume that, a typical value of d is 3%, and a typical value of i is 5%. For the old lighting system, the capital cost of replacement luminaires must also be adjusted by using an appropriate NPV factor. This cost can be accounted for in the payback period by subtracting it from the upfront capital cost (C), thus creating a “modified upfront capital cost”. The manufacturing and disposal cost are not considered here, as this is not the responsibility of the building owner, else it is implicit in the capital costs. It is more instructive to analyse the building as consisting of two separate floors, each using half the space available for PV generation on the south-facing roof. The question that we ask is “to what extent can half the roof power one floor?”.

801

Although this particular house is a detached family home, it can equivalently be regarded as consisting of a terraced house being sub-divided into two separate apartments. Moreover, if each apartment consists of one or two people, then not all of the rooms on each floor will be occupied. It makes sense to only keep the lights on in the occupied room(s), where sensors are able to detect occupancy, with the corridor light always switched on for safety reasons. It is assumed that, this configuration amounts to 50% of the apartment’s floor area being lit (in reality, suitable lighting controls will be required with carefully timed dimming, so that, the on-off switching does not cause visual discomfort to the occupants), and a comparative analysis was done for 100% of the floor area being lit (no sensors) against 50% of the floor area being lit (with sensors). The loads will depend on the type of luminaire we use, but a typical luminous efficacy for warm white LEDs (as of 2012) is 60 lm/W, and this is set to improve even further. For a house with 7.2 m × 8.2 m floor area and 2.4 m room height (implying K-factor and utilization factor of 1.17 and 0.9, respectively), one can use the lumen method to show that, if the working plane requirement of 150 lux is to be satisfied using 7 W ceiling-mounted luminaires, 14 of these luminaires are required resulting in a total power requirement of 102.5 W. The old lighting system that is being replaced is a mains-powered CFL based system of efficacy 60 lm/W, lifetime of 5 years, and capital cost of £2 per 1,000 lumens. The area of the south-facing part of the roof is 34 m2, and the whole of this area is to be fitted with PV panels, giving 17 m2 of PV panels to provide power to each floor. It was assumed that, there were no exterior obstructions to create shadowing. The PV efficiency is assumed to be 15% with a cost of £2,000 per installed kWh, and the battery cost is 40 p/kWh. In predicting the future long term horizons, the following inputs described in Table 2 were used.


802

Table 2 The cost and efficiency figures used to predict the long term payback trend. Year 2005 2010 2015 2020 2025 2030 2040 2050

PV efficiency 12 14 16 18 20 21 21.5 22

PV cost ($/kWh) 3,830 2,250 1,660 1,330 1,080 1,000 830 750

LED efficacy 35 56 150 175 200 200 200 200

LED capital cost per 1,000 lm 40 18 7.5 3 2 1 1 1

Battery cost ($/Wh) 0.63 0.45 0.35 0.3 0.27 0.25 0.24 0.24

3. Results and Discussion If there are no occupancy sensors, then there is an average daily PV excess for six months of the year, as can be seen by the red line in Fig. 2 (when the red line is above the horizontal axis there is a net PV excess, conversely below horizontal axis denotes grid use), and the battery size is determined by the maximum daily deficit PV energy in comparison to the load (green line), averaged over the whole year. It was found to be -1,650 Wh, indeed shows that, there was little variation in maximum daily deficit each month. Allowing for a factor of 1.5 to consider the non-ideal battery behaviour, this also resulted in a battery size exceeding 400 Ah, which (as of 2013) is beyond the range of typical devices on the market. In the presence of occupancy sensors to light only 50% of the floor space, a more interesting picture emerges (Fig.3) There is an average daily PV excess for eight months of the year, the annual average maximum daily deficit, is around 780 Wh, and the battery size is just under 200 Ah, which (as of 2013) is now well within the range of commercial devices on the market. It seems that by reducing the required lighting power, in this case, by reducing the required floor area to be lit, we correspondingly reduce the required battery size. Fig. 4 explores how the future projections of PV and LED performance determine the payback period: As the LED luminous efficacy improves from 60 lm/W to 120 lm/W, the payback is expected to reduce by around three years. In comparison, if the PV cost per

Fig. 2 The monthly difference between the generated PV energy and load use if there are no occupancy sensors.

Fig. 3 The monthly difference between the generated PV energy and load use when there are room occupancy sensors.

installed kW were to half, then so would the payback period. Indeed, once the PV cost drops to £500 per kW, then the payback period is of the order of a few years—similar to that contributed by the LED luminaires themselves. The effect of solar tracking has not been considered


Fig. 4 The effect of the PV costs and LED efficacy on the payback period.

803

buildings have been investigated. It has been shown that, for the typical system sizing and costs (as of 2013), the energy storage requirements are beyond the range of current commercial batteries. However, if it is desired that 50% less floor area needs to be lit for the given PV roof area, then commercial batteries of around 200 Ah can accommodate this storage requirement. Also, at present, the payback period is dominated by the high PV cost, and this needs to reduce by at least half before its contribution is comparable to that of the battery and LED luminaires. Nevertheless, during the next few years, costs and efficiencies are set to improve significantly, and we should begin to see large scale uptake of this technology. Further, this solar-battery system can be used to power other DC applications, e.g., solar panels on garages powering electric cars.

Acknowledgments

Fig. 5 Long-term prediction of the payback time.

here, but work by the authors is currently ongoing, and it is expected that, this would increase the annual PV output by an extra 30%-40%. The long-term behavior up to 2050 (Fig. 5) showed that, there is a rapid reduction in the payback from 30 years to 10 years up to 2020, and then there is only a slight reduction in the payback of 7 years up to 2050. As can be inferred from Table 2, the PV and LED costs drop quite rapidly, but the battery cost reduction is more gradual. Thus it is expected that, by 2020, the payback has shortened to encourage a large-scale uptake of the PV-Battery-LED systems.

The authors would like to acknowledge the discussions with numerous other colleagues at the Welsh School of Architecture, and for the availability of a research grant from the Low Carbon Built Environment Project, supported by the European Regional Development Fund through the Welsh Government.

References [1]

[2] [3]

[4]

4. Summary

[5]

The economic viability of PV-LED systems for interior lighting of residential and other small

[6]

Zhu, D. 2012. “High-Efficiency InGaN/GaN Quantum Well Structures on Large Area Silicon Substrates.” Physica Status Solidi (A) Applications and Materials Science 209 (1): 13-6. Humphreys, C. J. 2008. “Solid-State Lighting.” MRS Bull 33 (04): 459-70. Shailesh, K. R., and Raikar, T. S. 2010. “Application of RELUX Software in Simulation and Analysis of Energy Efficient Lighting Scheme.” International Journal of Computer Applications 9 (7): 24-35. Biber, C. 2008. LED Light Emission as a Function of Thermal Conditions. Portland: Biber Thermal Design Ltd. Parry, J. 2011. “Thermal Simulation Simplifies LED Luminaire Development.” Mentor Graphics—White Paper. Dong, T., and Narendran, N. 2009. “Understanding Heat Transfer Mechanisms in Recessed LED Luminaires.”

804

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

Solar-Powered Direct-Current Loads in Small Buildings Presented at the Ninth International Conference on Solid State Lighting, San Diego, USA. Jou, J. H. 2009. “Sunlight-Style Color-Temperature Tunable Organic Light-Emitting Diode.” Applied Physics Letters 95 (1): 013307-013307-3. Liu, G. 2012. “Techno-Economic Simulation and Optimization of Residential Grid-Connected PV System for the Queensland Climate.” Renewable Energy 45 (September): 146-55. Liu, G. 2012. “Simulation and Optimization of Residential Grid-Connected PV System in Queensland, Australia.” Advanced Materials Research 347-353 (October): 715-24. Sastry, O. S. 2010. “Development of White LED Based PV Lighting Systems.” Solar Energy Materials and Solar Cells 94 (9): 1430-3. Boeke, U., Wendt, M., and Yseboodt. L. 2011. “Combined Solar and AC Mains Powered LED Lighting System.” In Proceedings of the 2011 14th European Conference on Power Electronics and Applications, 1-8. Panguloori, R., Mishra, P., and Boeke, U. 2011. “Economic Viability Improvement of Solar Powered Indian Rural Banks through DC Grids.” In Proceedings of the 2011 Annual IEEE India Conference, 1-4. Patel, A. R. 2011. “Modeling and Simulation of Photovoltaic Based LED Lighting System.” World Academy of Science, Engineering and Technology 73 (March): 647-51. Pode, R. 2010. “Solution to Enhance the Acceptability of

[15]

[16]

[17]

[18]

[19]

[20]

[21]

Solar-Powered LED Lighting Technology.” Renewable and Sustainable Energy Reviews 14 (3): 1096-103. Leadbetter, J., and Swan, L. G. 2012. “Selection of Battery Technology to Support Grid-Integrated Renewable Electricity.” Journal of Power Sources 216 (October): 376-86. Braun, M. 2009. “Photovoltaic Self-Consumption in Germany Using Lithium-Ion Storage to Increase Self-Consumed Photovoltaic Energy.” Presented at the 24th European Photovoltaic Solar Energy Conference and Exhibition, Hamburg, Germany. Albright, G., Edie, J., and Al-Hallaj, S. 2012. “A Comparison of Lead Acid to Lithium-Ion in Stationary Storage Applications.” Alternative Energy Emagazine Industry. Esram, T., and Chapman, P. L. 2007. “Comparison of Photovoltaic Array Maximum Power Point Tracking Techniques.” IEEE Transactions on Energy Conversion 22 (2): 439-49. Huang, B. J. 2010. “Development of High-Performance Solar LED Lighting System.” Energy Conversion and Management 51 (8): 1669-75. Kelly, N. A., and Gibson, T. L. 2009. “Improved Photovoltaic Energy Output for Cloudy Conditions with a Solar Tracking System.” Solar Energy 83 (11): 2092-102. Sarnobat, P., and Lannon, S. 2013. “Simulation of PV-Battery-LED Systems in Office Buildings.” In Proceedings of the 13th Conference of International Building Performance Simulation Association (at press).

D


DAVID

PUBLISHING

Design and Simulation of a Solar Regulator Based on DC-DC Converters Using a Robust Sliding Mode Controller Gaga Ahmed, Errahimi Fatima and ES-Sbai Najia Laboratory Renewables Energy and Intelligent Systems, Faculty of Science and Technology, University Sidi Mohamed Ben Abdellah, Fez BP 2202, Morocco Received: June 06, 2015 / Accepted: August 03, 2015 / Published: September 30, 2015. Abstract: The MPPT (maximum power point tracking) is one of the most important features of a regulator system that processes the energy produced by a photovoltaic generator. It is necessary, in fact, to design a controller that is able to set the output value of the voltage and ensure the working within the maximum power point. In this paper, we propose the application of the robust sliding mode control technique to a DC-DC buck converter which is combined with a classical P & O (perturbation and observation) algorithm to enhance the solar system efficiency. Dynamic equations describing the boost converter are derived and a sliding mode controller for a buck converter is designed. It is shown that, this control approach gives good results in terms of robustness toward load and input voltage variations. The effectiveness of the proposed work is verified by the simulation results under PowerSim environment. Key words: Photovoltaic generator, MPPT, DC-DC converter, Buck converter, sliding mode controller, P & O, PowerSim.

1. Introduction The MPPT (maximum power point tracking) is one of the most important features of a system that processes the energy produced by a photovoltaic generator. It is necessary, in fact, to design a controller that is able to set the value of voltage or current of the generator and always ensure the working within its maximum power point. This point can substantially change its position during the day, essentially due to weather variations, then sunshine and temperature. So, the tracking control of the maximum power point is a complicated problem, to overcome these problems and ensure the high efficiency of the PV system, different solar regulators based on many MPPT strategies have been developed, such as FOCV (fractional open-circuit voltage), FSCC (fractional short-circuit current), fuzzy Corresponding author: Gaga Ahmed, PhD student, research fields: energy efficiency, renewable energy and intelligent systems design. E-mail: [email protected].

logic, neural network, P & O (perturbation and observation) and incremental conductance algorithms. These strategies have some disadvantages such as high cost, difficulty, complexity and instability. To resolve the problems affecting the solar MPPT regulator, especially the stability drawback, we propose a new structure of regulator which is based on double DC-DC converter. The boost converter that allows the nonlinear solar generator to provide maximum power to the photovoltaic terminal, and the buck converter that ensures the voltage regulation at a desired reference point. The used switched mode DC-DC converters are nonlinear and time variant systems, and do not rule by a linear control theory. The state space averaging method is usually applied for the DC-DC converter characterization. With this approach, an equivalent average model is developed by circuit averaging in a switching case.

806

Design and Simulation of a Solar Regulator Based on DC-DC Converters Using a Robust Sliding Mode Controller

Sliding mode approach for VSS (variable structure systems) offers an alternative way to implement a control action which exploits the inherent variable structure nature of DC-DC converters [1, 2]. The converters model are presented in state space and controlled in such way to force the system trajectory to remain on a suitable selected surface on the phase space called the sliding surface. The most remarkable feature of sliding mode control is its ability to result in very robust control systems.

2. DC/DC Converter Modeling The switching DC/DC converters are most widely applied in photovoltaic systems as an intermediary between the PV and the load to follow up the MPP (maximum power point). The DC-DC converters are hybrid dynamical systems characterized by both continuous and discrete dynamic behavior. In this section, we present a general modeling approach of DC-DC converters by application of the state space averaging technique of the buck and boost converter for the case of a continuous conduction mode. Different topologies and different design approaches could be used for DC/DC converters. In this part, two different models of converters are introduced, buck and boost converters. The diagrams in Figs. 1 and 2 show the structure of these converters with the switching period T and duty cycle d. During a period T, a switching converter has two working topologies. First, when the switches are closed, the converter model is linear. The state-space equations of the circuit can be written as follows:

When the switches are opened, the converters can be modeled by another linear state-space representation written as follows:

From the above equation, we can determine the averaged model given by the following state space

IL Vi

Fig. 1

V0

VS

IS

Equivalent circuit of a boost converter. VL

IL Vi

VD

V0

Fig. 2 Equivalent circuit of a buck converter.

equation for an entire switching cycle T:

where, the matrices follows:

,

,

,

are defined as

1 1 1 1 And , and are, respectively the average of x, y and u during the switching period T. The state space representation can be expressed for these converters as follows:

where: and 2.1 Buck Converter The buck converter is known as the voltage step down and current step up converter [3]. This gives a hint of its typical application of converting its input voltage into a lower output voltage, where the ⁄ conversion ratio M varies with the duty ratio D of the switch [3, 4]. The state space equation of buck converter is as follows: 1 0 ; 1 1 0


807

so: d d

W

1

2.2 Boost ConverterThe boost converter is also known as the step-up converter. The name implies it is typically application of converting a low input-voltage to a high output voltage, essentially functioning like a reversed buck converter [3, 4]. The state space equation of boost converter is as follows: 1

;

0

So: d d

1

1

3. MPPT Technique MPPT algorithms are necessary in PV applications because the MPP of a solar panel varies with the irradiation and temperature, so, the use of MPPT algorithm is required in order to obtain the maximum power from a solar array. The P & O method is one of the most commonly used methods in practice, it is operate by periodically perturbing, i.e., incrementing or decrementing, the array terminal voltage and comparing the PV output power with that of the previous perturbation cycle. If the PV array operating voltage changes and power increases, the control system moves the PV array operating point in that direction. Otherwise, the operating point is moved in the opposite direction, Fig. 3 demonstrates the principle of the P & O algorithm. The P & O algorithm is based on the “hill-climbing” principle, which consists of moving the operation point of the PV array in the direction in which the power increases [5, 6]. Hill-climbing techniques are the most popular MPPT methods thanks to their ease of implementation and good performance when the irradiation is constant [6]. The advantages of P & O

PV array voltage (V) Fig. 3 Principle of P & O algorithm.

method are the simplicity and low computational power they need. The drawbacks of this technique are mainly two, the main one is that they can easily lose track of the MPP if the irradiation changes rapidly [6-8]. The other handicaps of P & O method are the oscillations of the voltage and current around the MPP in the steady state [7, 9-11]. The Fig. 4 shows the flowchart of the P & O technique.

4. Sliding Mode Controller The SMC (sliding mode control) is naturally well suited for the control of variable structure system, it is a nonlinear control solution for VSC (variable structure

Fig. 4 Flow chart of P & O MPPT algorithm.


808

control) derived from the variable structure system theory. DC-DC power converters are variable structure systems. So, it is appropriate to apply SM control on these power converters [12]. Moreover, SM control offers excellent large-signal handling capability, which is important for variable structure system, it is known to be robust against modeling inaccuracies and system parameters fluctuations, it was successfully applied to electric motors, robot manipulators, power systems and power converters. Since the design of conventional PWM (pulse-width modulation) controllers is small-signal based, the converters being controlled operate optimally only for a specific condition [13] and often fail to perform satisfactorily under large parameter or load variations (i.e., large-signal operating condition) [14-16]. By replacing linear PWM controllers with a nonlinear sliding mode controller, power converters can reach better regulation and high dynamical performance for a varied operating range. In this part, we will present the general principle of the SMC and the controller design principle. First, let us consider the nonlinear system represented by the following state equation: , , where, x is n-dimensional column state vector, f and g are dimensional continuous functions in x, u and t vector fields, u is the control input. For the considered system the control input is composed by two components a discontinuous component and a . continuous one The continuous component insures the motion of the system on the sliding surface whenever the system is on the surface. The equivalent control that maintains the sliding mode satisfies the following condition: 0 The equivalent control ∂ ∂

,

maybe calculated as: ∂ ∂

∂ ∂

The nonlinear control component brings the system states on to the sliding surface. The nonlinear control component is discontinuous. It would be of the following general form: , ,

0 0

In this work, the SMC will be applied to a buck to ensure the voltage stabilization in the solar regulator terminal. We consider the following sliding surface where k is the sliding coefficient and e is the output voltage error defined as follows: where: By considering the mathematical model of the DC-DC Buck converter, the sliding surface can be expressed by the following expression: 1

1

and its derivative is given by: 1

Now, we consider the positive definite Lyapunov function V defined as follows: 1 2 The time derivative of V must be negative definite 0 to insure the stability of the system and to make sure that the surface S is attractive. Such condition leads to the following inequality: 0 To satisfy the above condition, the nonlinear control component can be defined as follows: sign

5. Simulations Results ,

Fig. 5 shows the block diagram of the complete solar


809

PV panel

Ipv & Vpv sensors

Reference voltage Fig. 5

Block diagram of complete solar regulator.

system regulator based on double DC-DC converter and sliding mode controller. The circuit of the complete photovoltaic system is simulated under PowerSim environment, the PV system is modeled using solar panel bloc, the MPPT algorithm is modeled using microcontroller bloc and a SMC is modeled using embedded function bloc. Fig. 6 shows that, the PV is connected to a resistive load without MPPT system. The load imposes its own characteristic which is generally different from maximum power point.

The simulation result of the circuit in Fig. 6 shows that, the load cannot extract the available power of the PV terminal (Fig. 7). To overcome this problem, we insert a DC-DC boost converter with variable duty cycle, the interest of this addition is to change the load characteristic which becomes a function of the duty cycle D and allows to move on the solar panel characteristic regardless of the load value. The boost converter is used to raise the input DC voltage to a higher output DC voltage.

VS

CS

Fig. 6

PV panel without MPPT system. 60 ∆P = Pmax – Pout 40

20

Pmax Pout

0

0.05 Fig. 7 PV-output power without MPPT system.

0.10

0.15

0.20

0.25


810

Now, when we can control the output characteristic by varying the duty cycle of the DC-DC boost converter which inserted between the PV panel and load, we need to make this duty cycle automatic, to do this, we must add a microcontroller as a digital processing bloc in which we must implement our MPPT algorithm. Fig. 8 shows the complete circuit diagram based on a PV panel, DC-DC boost converter, load, voltage and current sensors and MPPT controller. The algorithm shown in Fig. 4 is implemented in the MPPT controller bloc shown in Fig. 8. The following result demonstrates the performance of the MPPT algorithm implementation under irradiance step of 500, 1,000 and 1,500 W⁄m . The MPP values are shown in Pmax

VS

Ve IL

Current sensor

Insolation

blue, whereas the real values are shown in red. Fig. 9 illustrates the results of the simulation. For the SMC voltage regulator which based on a DC-DC converter, we consider the analysis given in the preview part, the proposed SMC was applied to a buck converter characterized by the parameters given in Table 1. Fig. 10 presents the simulated output voltage by application of a PWM control signal of 50% duty cycle. The voltage response corresponds to a second order damped system response with an overshoot. We can notice also that, the load variation affect directly the voltage in the load terminal using open loop control for our buck converter as shown in Fig. 11.

D1 PWN

Ce

Temperature Solar panel

MOSFET

CS

Voltage sensor

Pe

MPPT_controler

Fig. 8 PV panel with MPPT system.

Pmax at: 500, 1,000 and 1,500 W/m2

0

0.01

Fig. 9 PV-output power with MPPT system.

0.02

0.03 Time (s)

0.04

0.05

0.06

Load


To resolve this problem, we propose a complete SMC presented in Fig. 12. Fig. 13 presents the obtained result by application of

Table 1 Studied buck converter parameters. 17 V 22 µF 3 mH 50 ohm

Vin C L R

the proposed SMC to the studied controller for a 7.5 V for voltage reference. We can see clearly that, the observed voltage overshoot obtained on the open-loop response disappeared by application of the SMC. The SMC is tested also for the case of the load and input voltage variation. Fig. 14 presents the obtained

Ve VS

0 Fig. 10

0.005

0.010 0.015 0.020 Time (s) Open loop responses of the buck converter.

20

results for the case of the variation of the load resistance from 100 Ω to 500 Ω at 0.1 s and also three steps of input voltage variation. It is clear that, this perturbation is quickly rejected because the output voltage attends rapidly to the reference voltage. Fig. 14 shows the robustness, accuracy and stability of the applied SMC.

Ve

15

811

VS

6. Conclusions

10 5 0 0 Fig. 11

0.05

0.10 0.15 Time (s) Open loop responses with load variation.

Ve

0.20

IL

This paper presents the MPPT voltage regulator for solar systems, the simulation of both, MPPT P & O algorithm and sliding mode controlled DC-DC buck converter are presented in this study. The proposed control schemes are robust with respect to load and

VS

IS

PWM

Fig. 12

SMC for buck converter.


812

[4]

20

Ve

15 10

VS

[5]

5 PWM

0 0

1

Fig. 13 converter.

2

3 4 5 Time (ms) Application of the SMC to the studied buck

Ve

Ve VS

[7] VS

PWM 2.0 0 0.5 1.0 1.5 2.0 Time (s) Fig. 14 Output voltage evolution by application of the SMC for the variation of the input voltage (Vref = 2.5 V, Vref = 8.5 V). 0

0.5

PWM 1.0 1.5

input voltage step variations. The simulation results show the validity of the sliding mode controller for buck converter model and the robustness of this control technique against change in load and supply parameter. The proposed system offers a powerful ability which has several advantages such as stability even supply and load variation, robustness and good dynamic behavior. With additional development, the proposed integrated system will be widely applied to intelligent solar system control. Experimental control projects using FPGA (field programmable gate array) for hardware implementation is our next research effort.

References [1] [2]

[3]

[6]

Utkin, V. I. 1978. Sliding Modes and Their Application in Variable Structure Systems. Moscow: MIR Publishers. Mattavelli, P., Rossetto, L., Spiazzi, G., and Tenti, P. 1993. “Sliding Mode Control of SEPIC Converters.” In Proceedings of the ESPC (European Space Power Conferences), 173-8. Knopf, H. 1999. “Analysis, Simulation, and Evaluation of Maximum Power Point Tracking (MPPT) Methods for a Solar Power Vehicle.” M.S. thesis, Portland State University.

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

Ustun, T. S., and Mekhilef, S. 2010. “Effects of a Static Synchronous Series Compensator (SSSC) Based on Soft Switching 48 Pulse PWM Inverter on the Power Demand from the Grid.” Journal of Power Electronics 10 (1): 85-90. Nguyen, T. L., and Low, K. S. 2010. “A Global Maximum Power Point Tracking Scheme Employing DIRECT Search Algorithm for Photovoltaic Systems.” IEEE Transactions on Industrial Electronics 57 (10): 3456-67. Sera, D., Kerekes, T., Teodorescu, R., and Blaabjerg, F. 2006. “Improved MPPT Algorithms for Rapidly Changing Environmental Conditions.” In Proceedings of the 12th International Conference on Power Electronics and Motion Control, 1614-9. Femia, N., Petrone, G., Spagnuolo, G., and Vitelli, M. 2004. “Optimizing Sampling Rate of P & O MPPT Technique.” In Proceeding of the IEEE PESC (Power Electronics Specialists Conference), 1945-9. Hussein, K. H., Muta, I., Hoshino, T., and Osakada, M. 1995. “Maximum Photovoltaic Power Tracking: an Algorithm for Rapidly Changing Atmospheric Conditions.” IEEE Proceedings on Generation, Transmission and Distribution 142 (1): 59-64. Femia, N., Petrone, G., Spagnuolo, G., and Vitelli, M. 2005. “Optimization of Perturb and Observe Maximum Power Point Tracking Method.” IEEE Transactions on Power Electronics 20 (4): 963-73. Zhang, C., Zhao, D., Wang, J., and Chen, G. 2009. “A Modified MPPT Method with Variable Perturbation Step for Photovoltaic System.” In Proceedings of the Power Electronics and Motion Control Conference, 2096-9. Xiao, W., and Dunford, W. G. 2004. “A Modified Adaptive Hill Climbing MPPT Method for Photovoltaic Power Systems.” In Proceedings of the IEEE PESC, 1957-63. Venkataramanan, R. 1986. “Sliding Mode Control of Power Converters.” Ph.D. thesis, California Institute of Technology. Middlebrook, R. D., and Cuk, S. 1976. “A General Unified Approach to Modeling Switching Power Converter Stages.” In Proceedings of the IEEE Power Electronics Specialists Conference Record, 18-34. Forsyth, A. J., and Mollow, S. V. 1998. “Modelling and Control of DC-DC Converters.” IEEE Power Engineering Journal 12 (5): 229-36. Kassakian, J. G., Schlecht, M. F., and Verghese, G. C. 1992. Principles of Power Electronics. Reading, MA: Addison-Wesley. Mitchell, D. M. 1998. DC-DC Switching Regulator Analysis. New York: Mc-Graw Hill.

D


DAVID

PUBLISHING

Conductive Channel for Energy Transmission Victor Victorovich Apollonov A. M. Prokhorov General Physics Institute, Russian Academy of Sciences, Moscow 119991, Russia Received: June 11, 2014 / Accepted: July 08, 2014 / Published: September 30, 2015. Abstract: Laser spark obtained by using a conical optics is much more appropriate to form conducting channels in atmosphere. Only two types of lasers are actively considered to be used in forming high-conductivity channels in atmosphere, controlled by laser spark: pulsed sub-microsecond gas and chemical lasers (CO2, DF (deuterium fluoride)), short pulse solid-state and UV (ultraviolet) lasers. Main advantage of short pulse lasers is their ability in forming of super long ionized channels with a characteristic diameter of ~100 mm in atmosphere along the beam propagation direction. At estimated electron densities below 1,016 cm-3 in these filaments and laser wavelengths in the range of 0.5-1.0 mm, the plasma barely absorbs laser radiation. In this case, the length of the track composed of many filaments is determined by the laser intensity and may reach many kilometers at a femtosecond pulse energy of ~100 mJ. However, these lasers could not be used to form high-conductivity long channels in atmosphere. The ohmic resistance of this type a conducting channels turned out to be very high, and the gas in the channels could not be strongly heated (< 1 J). An electric breakdown controlled by radiation of femtosecond solid-state laser was implemented in only at a length of 3 m with a voltage of 2 MV across the discharge gap (670 kV/m). Not so long ago scientific group from P.N. Lebedev physical institute has improved that result, the discharge gap (-1 m) had been broken under KrF laser irradiation when switching high-voltage (up to 390 kV/m) electric discharge by 100-ns UV pulses. Our previous result -16 m long conducting channel controlled by a laser spark at the voltage -3 MV was obtained more than 20 years ago in Russia and Japan by using pulsed CO2 laser with energy -0.5 kJ. An average electric field strength was < 190 kV/m. It is still too much for efficient applications. Key words: Laser, exploding wire, conductive channel, laser spark, electric discharge, pulse-periodic laser, energy transmission.

1. Introduction For many years, the attempts to create super long conductive channels were taken in order to study the upper atmosphere and to settle special tasks, related to the energy transmission. There upon the program of creation of “Impulsar” represents a great interest, as this program in a combination with high-voltage high repetition rate electrical source can be useful to solve the above mentioned problems. It looks like as a kind of “renaissance of Nikola Tesla ideas” for the days of high power lasers. The principle of conductive channel production can be shortly described as follows. The “Impulsar”-laser jet engine vehicle-propulsion take place under the influence of powerful high repetition rate pulse-periodic laser radiation. In the experiments, Corresponding author: Victor Victorovich Apollonov, professor, research fields: high power lasers and interaction of laser radiation with matter. E-mails: [email protected], [email protected].

the CO2-laser and solid state Nd YAG (neodymium-doped yttrium aluminium garnet laser) laser systems had been used. Active impulse appears thanks to air breakdown (< 30 km) or to the breakdown of ablated material on the board (> 30 km), placed in the vicinity of the focusing mirror-acceptor of the breakdown waves. With each pulse of powerful laser, the device rises up, leaving a bright and dense trace of products with high degree of ionization and metallization by conductive nano-particles due to the ablation process. Conductive dust plasma properties investigation in our experiments had been produced on the basis of two very effective approaches: high power laser controlled ablation of different materials and by electrical explosion of wire. Experimental and theoretical results of conductive canal modeling will be presented. The estimations show that, with already experimentally demonstrated figures of specific thrust impulse, the lower layers of the ionosphere can be

814

Conductive Channel for Energy Transmission

reached in several hundred seconds that is enough to keep the high level of channel conductivity and stability with the help of high repetition rate high voltage generator. At present, many laboratories still continue the search for an efficient laser-based lightning protection system [1-3] and for producing a controlled high-conductivity channel in the atmosphere [4-8], capable of conducting a short-circuit current [9] in a natural or artificial electrical circuit. In Ref. [7], demonstrated a 1.5-fold increases in the length of the discharge gap broken under laser irradiation when switching extended (~1 m) high-voltage (up to 390 kV) electric discharges by 100-ns UV pulses of a KrF laser. Laser-based lightning protection systems, as is known, rely on the so-called long laser spark, which provides the conditions for connecting a thunderstorm cloud with a grounded metal rod, i.e., a classical lightning rod. Maximum lengths (~16 m) of the laser-spark-controlled electric discharge at a voltage of 3,000 kV were obtained in Russia and Japan [3] using a 0.5 kJ pulsed CO2 laser with spherical optics. Such control conductivity channels can be used in energy transmission, overvoltage protection systems, transport of charged particle beams, plasma antennas, etc. [8].

2. Lasers for Producing Sparks It was shown in Refs. [10, 11] that, a laser spark produced using conical optics demonstrates much better characteristics from the point of view of formation of conductivity channels in the atmosphere. Currently, among the huge variety of different lasers, only two types are being actively studied to be used in the formation of laser-spark-controlled high conductivity channels in the atmosphere: submicrosecond pulsed gas and chemical lasers (CO2, DF (deuterium fluoride)) and femtosecond solid-state lasers [4, 6, 12]. The main advantage of femtosecond lasers is the ability of producing superlong ionized channels (so-called filaments) with a characteristic diameter of

about 100 µm in atmosphere along the laser beam propagation direction. With an estimated electron density of 1016 cm-3 in these filaments and the laser wavelength in the range of 0.5-1.0 µm, the plasma hardly absorbs the laser light. In this case, the length of the track consisting of many filaments is determined by the laser intensity and can reach many miles at a femtosecond pulse energy of ~100 mJ. However, these lasers could not be used to form long high-conductivity channels in atmosphere. The ohmic resistance of the thus formed conducting channels turned out to be very high, and the gas in the channels could not be strongly heated (the femtosecond laser energy is less than  1 J). In particular, an EB (electric breakdown) controlled by radiation of a femtosecond solid-state laser was implemented in Refs. [5, 6] only at a length of 3 m (with a voltage of 2 MV across the discharge gap). As shown in Refs. [13, 14], to form a high-conductivity channel in the atmosphere, which could commute an artificial or natural discharge (lightning), the gas must be heated to a temperature from 6,000 K to 8,000 K, when thermal ionization begins. This condition is crucial, because the high initial electron density in the atmosphere is a necessary but not sufficient condition for realizing high conductivity in a long atmospheric channel. As was mentioned in Ref. [13], it is important not only to produce but also to maintain the electron density in an atmospheric channel that is sufficient for the breakdown. During gas heating, thermal ionization becomes the main mechanism of electron “production” in the channel. Specifically, this is the main reason why the streamer mechanism cannot lead directly to an electric breakdown in the presence of high electron density (filaments) in a cold channel [13]. In the case of submicrosecond CO2 or DF lasers with a high energy per pulse (several hundreds of joules), long conducting laser-plasma channels can be formed with conical optics. In contrast to spherical optics, one can use methods of lens dynamic correction to increase aberrations and elongate the focal segment by passing


to a conical phase front [11]. For example, an axicon transforms the phase front of a plane wave so that it becomes conical rather than spherical and converges to the z axis at an angle γ. For small angles (γ < 1°), it is simply expressed in terms of the angle γ at the axicon base and the refractive index N of the axicon material: γ = (N – 1)α. Each ring element of a conical wave front with a radius R and width δR is focused into an axial element of length δz = δR/γ and the entire focal segment has the length L = R/tgγ = R/γ [11]. This length can be made significant. For example, at R = 10 cm and γ = 0.5°, it amounts to L ≈ 11 m. The cylindrical symmetry of focusing suggests that, the type of transverse intensity distribution is independent of the coordinate z, a situation corresponding to a “diffractionless” wave beam. In reality, such beams also undergo diffraction, however, the diffraction energy loss from the central part of the beam is compensated for by the distributed lateral supply of radiation. Note that, the theoretical length of the focal segment formed by the axicon is independent of the laser beam energy and power and is determined by only the beam diameter. At R = 100 cm and γ = 0.5°, it amounts to L ≈ 110 m. According to our estimates, when using CO2 lasers with an energy of ~5 kJ per pulse, the length of these channels for wide laser beams and conical optics can be fairly large (much more than 100 m). When laser radiation is focused by an axicon in a real experiment, the longitudinal intensity distribution I(z) along the focal length L depends on z, since the transverse intensity distribution I(R) of the input beam is transformed by the axicon into the longitudinal distribution I(z) of the Bessel beam. When the radiation intensity in the beam reaches the threshold, there occurs a breakdown in the medium and an extended plasma channel is formed along the focal length of the axicon [11]. Initially, the channel has a diameter of 20-100 µm and a length that is on order of the focal length L of the axicon. If the breakdown is stable, the plasma channel, in contrast to the case of

815

spherical optics, becomes continuous for few nanoseconds. The density of the plasma, depending on the parameters of the medium, the wavelength λ, and other experimental conditions, varies from 1017 cm-3 to 3 cm-3, the temperature reaches 30-40 eV, and specific conductivity is 10 (ohm·m)-1. A few microseconds later the channel expands to 10 mm. Accordingly, the temperature drops to 1 eV and retains at this level for about 100 µs, followed by a relaxation of the plasma [11]. Fig. 1a shows a photograph of a plasma channel in scattered heating radiation with a wavelength λ = 1.06 µm at the breakdown onset (τi ≈ 50 ns, exposure time 5 ns, delay time td = 0 ns) [11]. The channel structure is affected by the combination of such parameters as the pulse width and the angle γ. For example, for a long pulse, laser-supported detonation may arise at inhomogeneities of the channel structure [15], as a result of which the structure becomes herringbone [11]. The photograph of a CLS (continuous laser spark) in Fig. 1b was obtained by focusing CO2-laser radiation in the atmosphere by a conical mirror with an opening angle γ = 177.8° [8]. One can see that, a laser spark looks like a continuous cord, composed of bright tapered elements. There are no discontinuities that are characteristic of sparks produced by focusing radiation by spherical optics. A similar pattern was observed at laser energies below 300 J [8]. In the photograph of a CLS fragment in Fig. 1c, the longitudinal structure of the spark looks like a set of bright tapered elements, which are in contact and equally spaced. Fig. 1d shows a photograph of the electric breakdown in the atmosphere for a channel formed by a CLS. When studying the spark between two rods immersed in its plasma and spaced by d = 80 cm [8], an electric breakdown occurred at a laser energy of 140 J and minimum average electric field strength Emin ≈ 77 V/cm. The up-to-date experiments aimed at implementing laser-spark-guided electric discharges shows another fundamental difference in the case of femtosecond solid-state and long-wavelength lasers. In the presence

816


(a)

(b)

(c) (d) Fig. 1 (a) Schematics of focusing laser radiation by a CL (conical lens) and a photograph of the channel in scattered heating LR (laser radiation) with a wavelength 1.06 µm at the breakdown onset (50 ns, exposure time 5 ns, delay time = 0); (b), (c) photographs of a continuous laser spark for CO2-laser radiation focused by a conical mirror with an opening angle = 177.8°; and (d) a photograph of electric breakdown in atmosphere in a CLS-induced channel.

of a laser-plasma channel composed of filaments, the breakdown voltage decreased by 30% [5]. The use of CO2 and DF lasers made it possible to reduce it by a factor of more than 10. As was pointed out in Ref. [16], the guiding properties of CLS (with a high voltage delayed by 10-20 µs) are similar to the action of a chain of metal balls. After an optimal time of 10-20 µs, the expansion of individual breakdown “zones” leads to the formation of a continuous quasi-cylindrical conducting channel with a low gas density (~0.1 of the normal density), high equilibrium temperature (2,700-2,900 K), and electron density in the range from 107 cm-3 to 108 cm-3.

3. Pulse-Periodic Laser Another approach to the formation of a conducting channel relies on the application of a pulse-periodic laser. It was shown in Ref. [17] that, electric-discharge and gas-dynamic lasers with a high peak power can operate in a high-frequency pulse-periodic regime.

This lasing regime allows one to produce a long conducting channel in the atmosphere, which exists for a long time. To this end, it is necessary to align a laser beam with an inverted Cassegrain telescope system, which consists of two spherical mirrors. The first (main) mirror of large diameter is immobile, while the second mirror can move according to a certain law. When moving the second mirror, the common real focus of the system changes its position in time from infinity to some minimum value, which is determined by the optical scheme of the telescope. Calculations show that, at a certain repetition frequency of laser pulses and a certain motion law for the second mirror, continuous extended plasma channels can be formed. However, practical implementation of this scheme with a “running” focus meets a number of fundamental difficulties. The light distribution in the focal plane, yielded by a well correlated lens, is in essence due to the Fraunhofer diffraction. The specific features of the out-of-focus 3D monochromatic images of a point


source, obtained with a round hole, were considered for the first time by Lommel. Along the optical z axis, the length of the focal segment that determines the longitudinal size of the energy-accumulation region depends quadratically on the focal length of the system. The transverse size of this region is proportional to the focal length and the angular divergence of laser radiation, therefore, the volume within which energy is concentrated is proportional to the focal length in the fourth power. For this reason, to form a homogeneous plasma channel several kilometers long, one must use a pulse-periodic laser with a shaped pulse energy and the second telescope mirror moving according to a complex program. A practical implementation of this method with a pulse-periodic CO2 laser with an average power of 1 kW yielded a continuous plasma channel ~1 m long [18].

4. Formation of a Current-Conducting Channel A completely different approach to the formation of a current-conducting channel can be realized using laser firing according to the “Impulsar” program [19]. Motion of an “Impulsar” laser engine under irradiation by a high-frequency pulse-periodic laser with a pulse energy sufficient for producing a breakdown in the focus of the optical system of a LRE (laser reactive engine) gives rise to a continuous conducting channel in air as a result of the formation of a mixture of aerosol [20] with atmospheric air and its subsequent ionization [19]. In this case, according to the calculations, the aforementioned range of laser pulse repetition rates provides continuity of the conducting channel at optimal velocities of the focusing system. Indeed, each laser pulse focused by the optical system forms some extended plasma region with a high conductivity, which occupies a relatively small segment of the focusing system trajectory. If the laser pulses have a low repetition rate [21], these regions will form somewhat like a dashed line at some velocities of the focusing system. If the pulse repetition rate exceeds

817

10 kHz and the motion velocities are optimal for solving the problem stated, these high-conductivity regions will have no discontinuities, and a continuous conducting channel will be formed. However, when the moving optical system enters rarefied atmospheric layers, a problem of medium deficit arises. To solve it, the displaced optical system must have an efficient source of an appropriate medium (a material containing microscopic metal particles, which can easily be sublimated under laser irradiation) in the vicinity of the focus. In particular, one can use copper oxide (СuO) nano-powder [22], which provides synthesis of conducting aerosol copper particles upon laser heating. The presence of aerosol particles of heavy metals in the sublimated material makes it possible to increase the specific impulse of the LRE thrust, simultaneously with increasing the channel conductivity. To implement a long conducting channel, one must solve several problems:  development of a high-frequency pulse-periodic CO2 laser with an average power P ≈ 100 kW and a peak power sufficient for initiating a breakdown in the focus of the LRE optical system;  development of a system of dynamic correction for the wave front of a wide-aperture laser beam with a diameter D # 10 cm in the presence of small-scale turbulence in atmosphere;  consideration of the laser radiation absorption and scattering by aerosol particles in the LRE exhaust;  alignment of the starting position of LRE with the laser and the pulsed high-voltage (up to several MV) source;  obtainment of maximally possible specific impulse of the LRE thrust to reduce the time of aerosol channel formation;  choice of the material of aerosol heavy-metal particles and their dispersion composition that could provide high channel conductivity upon heating to the sublimation temperature in the LRE;  determination of the lower limit for the size of dispersion aerosol particles to ignite electric

818


breakdown of a channel capable of transmitting a current, corresponding to a short-circuit current in natural or artificial electric circuits at a minimally possible average strength of electric field. To address the latter two problems, we conducted a series of experiments. The results obtained are reported below. We consider urgent the problems of determining the dynamics of change in the channel conductivity at long delay times, and its maintenance at the optimal level by introducing additional high-frequency generators into a chain, they will be analyzed in the next stage of the “Impulsar” program.

5. Channel Formed by an Exploding Thin Wire The formation of a conducting channel based on the Impulsar technology was simulated using the results of the laboratory experiments on a controlled electric breakdown in an atmospheric channel containing aerosols [22, 23] of hot copper and copper oxide (CuО, Cu2О) particles, formed as a result of wire explosion. Note that, the proposed method for generating particles of variable size with a necessary concentration appears to be fairly efficient in terms of the price-quality criterion when one must carry out many experiments aimed at modeling the optimal conditions for high conductivity. There are a number of factors facilitating the electric breakdown (including the streamer stage) of long gaps in an atmospheric channel formed as a result of wire explosion at average external electric fields. As noted in Ref. [14], the main problem in the analysis of the formation of a spark channel in a discharge gap in atmosphere using a field of arbitrary configuration is to determine the mechanism of gas heating, which provides thermal ionization of the gas in the spark channel. The heating in the discharge channel was generally believed to be caused by the increase in the concentration of the charged particles supplied by shock gas ionization. However, the concentration of charged particles in the range of 1011-1013 cm-3 is

limiting for shock ionization. The development rate of thermal processes in a gas is limited by the velocity of the electrons that transform the electric field energy to the thermal energy of gas molecules. The thermal inertia of the gas excludes the possibility of step changes in its temperature and therefore, its conductivity, both during heating and cooling the channel. It was noted in Ref. [13] that, ionization in a highly ionized plasma differs significantly from the similar process in a weakly ionized plasma, where molecules are ionized by the electrons gaining energy directly from the electric field. A field in a highly ionized plasma supplies energy to all gas electrons. The latter are thermalized as a result of collisions and obey Maxwell’s distribution. The gas is ionized by the electrons that acquire sufficient energy during the energy exchange with other particles rather than directly from the field. It is much easier to maintain a highly ionized equilibrium plasma than a weakly ionized nonequilibrium plasma, because the former situation requires a much weaker field. For example [13], in an equilibrium plasma column of nitrogen arc at atmospheric pressure, which burns in a cooled tube of radius r = 1.5 cm at a current i = 10 A, the field supporting the arc column has a strength E = 10 V·cm-1. Under these conditions, T = 8,000 K, the gas density Na = 1018 cm-3, ne = 2 × 1015 cm-3, the degree of ionization x1 = ne/N = 2 × 10-3, and E/Na = 10-17 V·cm2. In a nonequilibrium glow discharge column in nitrogen, the strength of the discharge-supporting field is larger by an order of magnitude: E = 200 V·cm-1. Such a sharp difference is caused by the difference in the nature of ionization processes. In the case of weakly ionized plasma, the field must be sufficiently strong to accelerate electrons to an energy of 13-15 eV (ionization threshold). At the same time, for highly ionized equilibrium plasma, the field can be much weaker, because in this case, it is sufficient to accelerate electrons to only kΤ ≈ 1 eV. As indicated in Ref. [13], electrons supply atoms with energy, and,


819

position of the end face of polished quartz fiber, which transfers radiation from the discharge channel to a high-speed pin-photodiode. The end face of the polished quartz fiber was located at a distance of 2-10 cm from the wire axis near the discharge cathode or anode. The shape of the current pulse was measured using a differential screened Rogowski loop with a time resolution of ~30 ns [24]. The shape of the voltage pulse across the discharge gap was determined with the aid of a screened voltage divider of mixed type (connected to the capacitor battery anode), which

during the energy exchange, all particles jointly concentrate energy in separate electrons, which serve as ionization sources. Our experiments on the formation of a high-conductivity channel in atmosphere was performed with a PEV (personal electric vehicles)-2 copper wire in an enamel shell, the wire was 90-155 cm long and had a diameter of 90 mm. Fig. 2a shows a schematic of the experimental setup, the arrangement of the sensors for measuring current by a Rogowski loop and voltage on the capacitor battery using a voltage divider, and the

(a)

(b)

(c) Fig. 2 (a) Schematics of the experiment on igniting an electric breakdown in a wire-explosion channel for a horizontally oriented wire: (1, 2) brass electrodes, (3) ebonite insulators, IRT-3 (ignitron switcher), RL (Rogowski loop), VD (voltage divider), TP (triggering pulse), EFW (exploding fine wire), PD (photodiode), QF (silica fiber), A (anode) and C (cathode); (b) a photograph of controlled electric breakdown; and (c) a photograph of wire explosion in the absence of breakdown.

820


yields a measurement error of ~1% at a width of the current-pulse leading edge > 1 µs [25]. Recording was performed using a Tektronix TDS 220 digital oscilloscope. The capacitor with a capacitance C = 2 µF was charged from a voltage source to U0 = 7 – 10 kV. When switched on by an external triggering pulse from a controlled discharger (IRT-3 ignitron), the capacitor was discharged through a copper wire 90 µm in diameter (with allowance for the enamel shell, the wire diameter was ~110 µm). As the experiments showed, the presence of an enamel shell facilitates the formation of a thin straight-line electric breakdown channel with a conducting region 4-8 mm in diameter in atmosphere and promotes a unique mechanism of aerosol formation in the form of balls 330-20 µm in diameter with a black surface (oxidized during wire explosion). Large pieces of the enamel shell (up to 10 cm long) remain integer, with burn-through areas on the surface. During the wire explosion and electric breakdown, the enamel shell stabilizes the spatial arrangement of the balls in the channel, which are centers of ionization plasma formations. Fig. 2b shows a photograph of electric breakdown in a channel formed during explosion of a horizontally oriented wire (D 90 µm, length 105 cm). This photograph was recorded through an FS-1 blue filter and a neutral NS-9 filter by an OLYMPUS C-5050 ZOOM digital camera. At an exposure of 1.6 s the photographs have an integral character, i.e., the exposure during photographing the discharge was determined by the discharge duration. The digital camera matrix operated with a maximum resolution. When taking pictures of the entire discharge channel, the distance l from the wire centre to the digital camera objective was ~2 m. To reveal the electric breakdown details, sizes of individual hot particles formed during the wire explosion and the diameter of the channel and the inhomogeneities in the structure of channel luminescence, we took pictures in the macroscopic regime, in which the digital camera objective was located rather close (at 25-50 cm) to the wire. When

taking pictures of the wire explosion products, l was chosen to be 3 cm (super macroscopic regime). The digital camera was used in the ZOOM regime. The spherical shape of the wire explosion products (D 30-20 µm) is primarily explained by the effect of surface tension forces on the molten copper drops during their formation. The large spread of particle diameters is explained by the fragmentation and aggregation of metal drops during wire explosion [26]. The black color of the particles is due to the formation of copper oxide on their surface as a result of the interaction of copper drops with air oxygen. The oxidation is reversible, i.e., upon heating to ~1,200 K, copper oxide is reduced to copper [22]. During the wire explosion, the oxygen concentration in the channel formed should sharply decrease as a result of the oxidation. It was shown in this experiment that, by the instant of electric breakdown, the voltage across the discharge gap Uа = 7 kV provides the electric field strength in the channel Eс = Uа/d ≈ 67 V·cm-1. At the storage capacitance C = 2 µF, depending on the charging voltage and wire length, wire explosion may occur without electric breakdown. Fig. 2c shows a photograph of a channel (D 90 µm, length 115 cm) formed as a result of explosion in the absence of electric breakdown. It can be seen that the channel structure is intermittent: it consists of individual isolated plasma zones of superheated drops of wire material [27-29]. At a charging voltage U0 = 10 kV and storage capacitance C = 2 µF, the voltage across the capacitor battery decreases to 7.6 kV by the end of the wire explosion. At a constant charging voltage U0, the maximum pulse current during wire explosion is mainly determined by the initial wire resistance R0 and amounts to 1.25 kA. When changing the length of 90 µm copper wire within 90-155 cm, the measured resistance ranged from 2.7 W to 4.2 W. At U0 = 10 kV and C = 2-4 µF, the maximum amplitude of the pulse current was in the range of 1-1.4 kA. After the wire explosion, a current i ≈ 10 A flows for some time through the hot channel and then decays to zero. As a result, the


residual voltage across the capacitor battery decreases to 7 kV. Despite the presence of residual current after the wire explosion, which heats additionally the channel, the electric breakdown does not occur at these parameters of the discharge circuit. The amount of energy W spent on wire explosion, calculated from the current and voltage characteristics, was found to be ~35 J. In the series of experiments performed on a horizontally oriented wire, the capacitance of the capacitor battery was taken to be constant: C = 2 µF, the only variable parameters were the wire length (90-155 cm) and charging voltage (7-10 kV). It was found experimentally that, at a wire length of 100 cm, the minimum charging voltage at which there is no electric breakdown in the gap does not exceed 7 kV, a value corresponding to the minimum accumulated energy: W0 = 49 J. In the next series of experiments on the electric breakdown in a channel, we considered a vertically oriented wire. This version is preferred, because it takes into account the possible influence of the vertical component of electric field at the Earth’s surface on the electric breakdown in the channel formed by wire explosion. Fig. 3а shows a photograph of electric breakdown in a channel formed by a vertically oriented wire (D 90 µm, length 143 cm) at the initial stored energy W0 = 190 J, which corresponds to the charging voltage U0 = 10 kV and the storage capacitance C = 3.8 µF. The electric breakdown can be seen in the oscillograms of discharge channel current, voltage and luminescence, which were recorded by a high-speed pin-photodiode with a time resolution of ~1 ns. The light arrived at the photosensitive area of pin photodiode through a silica fiber ~2 m long, whose polished receiving end face was located near the discharge anode at a distance of 1.5-4 cm from the discharge channel. The radiation pulse width at half maximum was ~30 µs and hardly differed from the corresponding current pulse width during electric breakdown (the light-pulse oscillogram in Fig. 3e

821

shows only the arc phase of the breakdown). The oscillogram of the discharge current (Figs. 3b and 3d) indicates that, it is close to zero during the current pause [30, 31], whose duration is Δ ≈ 140 µs. The results of our studies showed that the current during the pause amounted to 10-50 A (depending on the discharge-circuit parameters). There are many theories explaining this phenomenon, they will be considered elsewhere. As shown in Ref. [31], the electric breakdown delay for a copper wire depends on the average electric field strength Ec in the channel at the instant when the arc phase of the breakdown begins. The voltage oscillogram demonstrates that, the breakdown occurred at a voltage of ~8.4 kV. Using the current and voltage oscillograms, one can also determine the channel resistance during the electric breakdown at the instant corresponding to Imax = 360 A. In this case, Rmin = 17 W. Here, in contrast to the data of Refs. [32, 33], the discharge-channel diameter is of little importance, because the discharge is slow, and the inductive component can be neglected when estimating the voltage drop across the discharge channel. In Fig. 3b, which illustrates the case of the electric breakdown in a channel formed by wire explosion at C = 3.8 µF and inductance L = 2 µH, the wave resistance ρ = (L/C)0.5 ≈ 0.7 ohm. Judging by the volt-ampere characteristics of the discharge gap, L ≈ 143 cm in the case of the electric breakdown at Rp ≈ 18 ohm, Rp > 2ρ (18 > 1.4 ohm), which corresponds to the aperiodic shape of the discharge. It was noted in Ref. [9] that, the channel resistance is affected by the amount of energy released in the channel rather than the current through the latter. At a wire length of 143-152 cm and the maximum capacitance Cmax = 4.5 µF, a breakdown in a channel formed by wire explosion occurred with a probability close to 100%. Fig. 3f shows a photograph of a controlled electric breakdown at C = 4.5 µF, U0 = 10 kV, and an interelectrode distance of 155 cm. In this experiment, an electric breakdown was ignited at a

822


(b)

(c)

(d)

(f) (a) (e) Fig. 3 (a) Photograph of controlled electric breakdown for explosion of a vertically oriented wire (the interelectrode length is 143 cm); (b)-(e) oscillograms of the discharge channel ((b, d) current, (c) voltage, and (e) luminescence at a capacitance C = 3.8 µF and charging voltage U0 = 10 kV); and (f) a photograph of controlled electric breakdown at C = 4.5 µF, charging voltage U0 = 10 kV, and an interelectrode distance of 155 cm.

voltage of ~8.3 kV. The duration of the current pause was Δ ≈ 200 µs. At a wire length of 152-159 cm, an electric breakdown occurred with a probability of ~30%. For wires longer than 159 cm, the breakdown probability (at the same parameters of the scheme) was close to zero. Thus, the average minimum electric field strength in the channel formed by wire explosion, at which breakdown can occur, is Emin ≈ 52 V/cm. The suggestion about the possibility of long-term support of current breakdown with an amount of accumulated energy increased by an order of magnitude and the same charging voltage was partially confirmed. The channel for recording radiation in the visible

wavelength range has a low sensitivity, because it detects only the radiation entering the angular aperture of the polished receiving end face of the quartz fiber. If the distance between the end face and the discharge channel is reduced to minimum (~1.5 cm), one can detect the following phases in the oscillogram of a visible-light pulse. During the first 16 µs after the wire explosion, the visible light intensity is low (the wire explosion gives rise to radiation with wavelengths in the range of λ ≈ 2.5-0.5 µm) [29]. Then the oscillogram of visible light pulse exhibits the radiation corresponding to the streamer (weak), leader (weak), and arc phases of the breakdown.


An analysis of many photographs indicates that, the electric breakdown trajectory (broken line, circle, etc.), which is set by spanned copper wire, is highly controllable and contains no branchings. The experiments revealed that the vertical orientation of exploded wire is indeed preferred in the case of streamer breakdown of maximally long gaps (Lmax ≈ 159 cm) at U0 = 10 kV and Cmax = 4.5 µF. Let us analyze the requirements to an ionized formation that can initiate electric breakdown. The minimal conditions that are necessary for initiating a streamer were reported in Ref. [13]. We will consider a uniform conducting sphere with a radius Rs, placed in a uniform field of strength E0. To form a self-propagating plasma channel, the external field must be displaced from the plasma formation and amplified beyond it. To displace the external field from the entire volume of the sphere, charges must be nonuniformly distributed over the surface, and there must be no less than N = 3 ε0 πR2 E0 /e electrons in the field. Taking into account that, the radius of this formation should be no less than -1 ( is the effective ionization coefficient), we can estimate the lower limit of the number of electrons at which a streamer may arise [13] as follows: Nemin ≈ 3 ε0πE0/eα2. As an example, we will consider air at atmospheric pressure. At the breakdown field strength E0 = 31.4 kV/cm and effective ionization coefficient  = 12.4 cm-1, we have Nemin = 2 × 109. The smaller the plasma formation radius, the more severe the limitation on the plasma density is. At the minimally allowable value R ≈ -1, the electron concentration in air at atmospheric pressure should be Nemin ≈ 4 × 1011 cm-3. The plasma formation size R also must be no smaller than the ionization length -1, in a field with a strength equal to the external field strength or even somewhat higher. In air at atmospheric pressure, the ionization length is -1 ≈ 0.8 mm. If the initial size of plasma formation in a homogeneous field satisfies the aforementioned requirements, this formation should generate ionization waves to both sides along the field, thus forming a

823

plasma channel. The plasma channel development is accompanied by amplification of the field at both leader heads [13]. Our experiments showed that, all above requirements to an ionized formation [13], which must be fulfilled to ignite an electric breakdown, are implemented in the channel of the discharge gap formed by wire explosion with the aforementioned parameters. The characteristics of separate plasma formations (Rs ≈ 0.5-4 mm, average distances ~3-5 mm, and temperature T ≈ 3,000-5,000 °C) with clusters of superheated drops and their aggregates at the centre, as well as low oxygen content in the wire-explosion channel, reduce significantly the average field strength that is necessary for electric breakdown. Note that, the size of individual plasma formations exceeds that of superheated drops of wire material by one to two orders of magnitude. The average density of the wire-explosion products in the channel is ρ ≈ 0.1 kg/m3. The electric field in separate spherical plasma fractions can be amplified by a factor of no more than 3. Further studies should reveal new factors leading to electric breakdown in the channel produced by wire explosion at Ec = 52 V/cm [34, 35]. We should note the original work by Komel’kov [36], who analyzed the leader-discharge channel. We showed that, the specific resistance per unit length in the channel of leader positive retarded discharge is 11.8 W/cm, while the longitudinal gradient along the channel in the end of the leader phase is 55.5 V/cm. The main parameters of the leader channel turned out to be the same as in the case of arc discharge. The result obtained in Ref. [37] was interpreted in Ref. [13] as not quite correct because of incorrect analysis of the discharge-gap voltage oscillograms. Our results disprove the conclusion of Ref. [13]. To check the influence of the size of individual plasma formations in the channel on the mechanism of electric breakdown, we performed experiments with explosion of a copper wire L ≈ 30-36 cm long and Ø 90 µm in diameter at U0 = 10 kV and Cmax = 4.5 µF, i.e., at a

824


much higher specific energy contribution. The channel formed by wire explosion with these parameters contains a fairly large number of wire-material nanoparticles [22], whereas large particles (individual plasma formations) are completely absent. In this case, the second requirement to the development of streamer mechanism of breakdown in the channel is violated. As a consequence, a fivefold decrease in the gap length did not lead to breakdown at the corresponding parameters of the discharge circuit. It occurred only for wires not shorter than L ≥ 60 cm, a minimum value at which the size of plasma formations in the wire explosion channel became sufficiently large to implement a streamer breakdown.

6. Channel Produced by a Solid-State Laser We also investigated the regime of electric breakdown in a plasma channel containing evaporated aerosol particles of metals and their compounds, which is formed by focusing radiation of a solid-state laser, operating in the modified spike regime, onto targets made of different materials [38]. To form a plasma plume on a target, we used a phosphate glass laser with a wavelength λ = 1.054 µm, operating in a modified lasing regime with a peak power of 100-500 kW. Photographs of lasers operating in this regime are shown in Figs. 4a and 4b. Solid-state lasers of this type are unique. Actually, they are laser hyperboloids (concerning the way of extracting radiation from a totally reflecting flat cavity through a round hole of large diameter, which can be seen in the upper photograph). Lasers of this type can produce large plasma plumes in both pulsed and pulse-periodic regimes. A laser pulse with a width up to 700 µs consisted of many (approximately 150) high-contrast pulses 150-300 ns wide with an average repetition rate of 200 kHz (Fig.4c). The radiation divergence at the laser output α ≈ 4 × 10-4 rad, was close to the diffraction one. The total energy per laser burst reached ~30 J (see the laser in Fig. 4a). Graphite, vinyl plastic, textolite, ferrite, permalloy and carbonyl core were

used as targets. Their exposure to focused laser radiation led to the formation of a cloud of rapidly expanding plasma, which contained aerosol particles of evaporated target material ~50-1,000 nm in size. No large particles were found in the plasma cloud. Note that, a copper target cannot be used to form an electric breakdown in this plume, because evaporated large copper drops have velocities of several hundred meters per second and cannot form a quasi-stationary channel for electric breakdown. The conductivity of this plasma has been studied little, while the dynamics of changes in its properties is of great interest to search for the ways of forming (according to the Impulsar program) superlong high-conductivity channels in order to transfer energy at large distances. The experiments were performed with a system of electrodes, between which an electric discharge was ignited (Fig. 5). An anode in the form of a copper cylinder was connected to the high-voltage output of the capacitor battery with a capacitance C = 2 µF, which was charged to a voltage of U0 = 10 kV. A duraluminum cathode (K) had a hole Ø 6 mm in diameter, through which laser radiation was guided to the target using a lens with a focal length of F = 300 mm. The distance between the electrodes was L ≈ 80 mm. The photo of the plasma plume on the target made of graphite is shown in panel (Fig. 5b). The results of image processing on the PC with the use of the HSB (hue saturation brightness) color model are presented in panel (Fig. 5c). At a laser pulse energy below 30 J, the gap was switched using the plasma formed on the target. Typical photographs of the electric breakdown in atmosphere on a graphite target (at a laser burst energy of 12 J) for a gap ~70 mm wide, obtained using an FS-1 filter, a neutral NS-9 filter, and an OLYMPUS C-5050 ZOOM digital camera, are shown in Fig. 6. In highly inhomogeneous fields, streamers are formed primarily in the regions of maximum field strength near the electrodes and, depending on the voltage polarity, are directed either to the anode or to the cathode [13].


825

(a)

(b)

(d)

(e)

(c) Fig. 4 (a), (b) General view of solid-state lasers; and (c)-(e) oscillograms of (c) the laser radiation; and (d), (e) separate laser pulses.

In the photographs in Fig. 6, the streamer directed to the anode develops more rapidly than that directed to the cathode. Streamers overlap at a distance of ~25 mm from the anode, and an electric breakdown of the gap occurs. An analysis of this process shows that, the breakdown has a random trajectory, as well as a natural lightning. In addition, the results of experiments with different targets show that electric breakdown occurs at a minimum average electric field strength of ~300 V/cm.

The experiments revealed also that, the electric breakdown delay (which is shown in the oscillograms of discharge-circuit current and luminescence) for the graphite target, could exceed 300 µs after the laser irradiation. In Fig. 6c, the capacitance C = 2 µF. The current oscillogram indicates that, the damping oscillation period T = 8 µs, therefore, the discharge inductance is 0.8 µH. The contour wave resistance ρ ≈ 0.6 ohm.

826


plasma plume. The targets made of other materials were characterized by a shorter possible delay. Changing the delay for the voltage pulse applied to the capacitor, one can measure the plasma characteristics at different instants and thus estimate the conductivity of the channel at different stages of its cooling and expansion. In our opinion, the strong electric field in these experiments is due to the absence of large particles in

According to the current-voltage characteristics of the L ≈ 70 mm discharge gap for the graphite target, the discharge gap resistance is Rp ≈ 0.5 < 2ρ ohm, a value corresponding to an oscillatory discharge with a high damping decrement. The long breakdown delay with respect to the laser irradiation onset is apparently caused by the closing of the discharge gap by the expanding vapor of the target material and discharge-gap preionization by UV radiation of the

F = 300 mm (a)

(b)

(c) Fig. 5 (a) Schematics of the system for studying the conductivity of plasma plume products (T indicates a target); (b) a photograph of a plasma plume on a graphite target; and (c) the results of image processing in the system of HSB color coordinates.


827

U = 10 kV

70 mm

(a) U = 10 kV

70 mm

(b)

6 kA

(c)

320 mks

1 500 mV CH2 500 mV BW M 50 μs

CH1

(d) Fig. 6 (a), (b) Photographs of electric breakdown of a 70 mm gap in atmosphere in plasma-plume products (photographing through an FS-1 filter and a neutral NS-9 light filter) and the results of image processing in the system of HSB color coordinates; (c) a current oscillogram at electric breakdown of the gap (voltage U0 = 10 kV, storage capacitance C = 2 µF) in plasma plume; and (d) a luminescence oscillogram for the discharge channel in aerosol plasma.

828


the erosion-plasma cloud. To carry out successfully the Impulsar program, we believe it necessary to study the dynamics of plasma conductivity and the residual aerosol products after cooling the plasma.

7. Experimental Results Below we report the estimated values of the channel conductivity, which were obtained both in our previous experiments on the breakdown of atmospheric air using long- and short-wavelength radiation and the experiments with aerosol plasma (this study). For the plasma channels obtained on graphite, ferrite, permalloy and carbonyl core targets using a phosphate glass laser, the gap specific resistance was 10 ohm/m and the specific breakdown voltage was ~300 V/cm. For the plasma containing a set of dispersion and condensation aerosols, which was obtained by wire explosion, the gap specific resistance was ~5 ohm/m and the specific breakdown voltage was ~50 V/cm. For the plasma channel formed by a pulsed CO2 laser with a conical focusing optics, the gap specific resistance was 100-400 ohm/cm and the specific breakdown voltage was ~80 V/cm. For the plasma channel obtained with the aid of a femtosecond pulsed laser, the gap specific resistance was ~0.1 MW/m and the specific breakdown voltage was ~7.5 kV/cm [4]. These estimates, based on the results of the electric-breakdown experiments, show that a voltage of 7 МB, at a total introduced energy of ~300 kJ, is sufficient to form a plasma channel ~1 km long and 8 mm in diameter [39, 40]. Note that, the minimum specific resistance of the channel depends strongly on the specific introduced energy, the transverse size of plasma formations, the concentration and size of aerosol particles, the repetition rate of laser pulses and their energy, and the humidity, temperature, and many other parameters of the medium where a conducting channel is formed. The results of the studies show that, the discharge plasma has a high conductivity and can efficiently be

used to form a conducting channel in order to transfer energy over long distances.

8. Conclusions We determined the range of sizes of copper and copper oxide particles (from 50 nm to 300 µm) and their minimum volume density (~0.1 kg/m3) that are necessary to provide an electric breakdown in order to form a plasma channel according to the “Impulsar” program. The results of the model experiments showed that, nanoparticles in the plasma channel can merge into larger aggregates during the LRE operation. The experiments on electric breakdown showed also that a high conductivity of a plasma channel with a diameter of ~8 mm and a minimum specific resistance of ~5 ohm/m is implemented at the maximum introduced specific energy: ~200 J/m. At a larger channel diameter, the specific energy is proportional to the squared ratio of diameters. The minimum average electric field strength (~52 V/cm) in the channel, at which a streamer-leader electric breakdown of the discharge gap begins, is due to the following factors:  the presence of dispersion and condensation aerosols of hot particles of copper oxide (CuO, Cu2O) and copper with oxidized surface (smaller than 300 µm in diameter) in the channel for a long time (1-10 µs) due to their low mobility;  the presence of superheated metal drops at the centre of plasma formations, which are in the metastable (unstable) state at temperatures T ≈ 3,000-5,000 °C and play a role of hot dots [31] during thermionic emission;  formation of large linear aggregates of aerosol particles [20], which are formed during wire explosion;  low oxygen content in the channel as a result of fast oxidation of aerosol particles formed during wire explosion.

References [1]

Apollonov, V. V. 1991. “High-Power CO2 and N2O Lasers with Volume Self-Sustained Discharge Pumping.”


[2] [3]

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11] [12]

[13] [14] [15] [16]

[17]

[18]

[19]

J. Opt. Soc. Am. B 8 (2): 220. Alexandrov, G. N. 1980. Elektrichestvo 12 (2): 47. Kinoshita, F., Morooka, Y., and Uchiumi, M. 1997. Proceedings of the XII Intern. Conf. on Gas Discharge and Their Applications, 475. Diels, J. C., and Rudolph, W. 2006. Ultrashort Laser Pulse Phenomena: Fundamentals, Techniques, and Applications on Femtosecond Time Scale. Burlington: Acad. Press. Kasparian, J., Rodriquez, M., and M’ejean, G. 2003. “White-Light Filaments for Atmospheric Analysis.” Science 301 (5629): 61-4. Berge, L., Skupin, S., Nuter, R., and Wolf, J. P. 2007. “Ultrashort Filaments of Light in Weakly Ionized, Optically Transparent Media.” Rep. Prog. Phys. 70 (10): 1633. Zvorykin, V. D., Levchenko, A. O., and Ustinovskii, N. N. 2011. “Control of Extended High-Voltage Electric Discharges in Atmospheric Air by UV KrF-Laser Radiation.” Quantum Electron 41 (3): 227. Apollonov, V. V., Vasilyak, L. M., and Kazantsev, S. Y. 2002. “Electric-Discharge Guiding by a Continuous Spark by Focusing CO2-Laser Radiation with a Conic Mirror.” Quantum Electron 32 (2): 115. Bazelyan, E. M., and Raizer, Y. P. 2001. Physics of Lightning and Lightning Protection. Moscow: Fizmatlit. Apollonov, V. V. 2005. “Feasibility Study of a CO2 Laser Based Lightning-Protection System Realization.” Opt. Eng. 44 (1): 014302. Pyatnitskii, L. N., and Korobkin, V. V. 2000. Trudy IOFAN 57: 59. Teramobile. 2011. “Publications of the Teramobile Project.” Teramobile. Accessed June 11, 2014. http://www.teramobile.org/publis.html. Bazelyan, E. M., and Raizer, Y. P. 1997. Spark Discharge. Moscow: Izd-vo µFTI. Aleksandrov, G. N. 1967. Zh. Tekh. Fiz. 37: 288. Raizer, Y. P. 1974. Laser Spark and Propagation of Discharges. Moscow: Nauka. Asinovskii, E. I., and Vasilyak, L. M. 2000. Encyclopedia of Low-Temperature Plasma: Introduction. Book II. Moscow: Nauka, 234. Tret’yakov, P. K., Grachev, G. N., Ivanchenko, A. I., Krainev, V. L., Ponomarenko, A. G., and Tishchenko, V. N. 1994. Dokl. Akad. Nauk. 336 (4): 466. Apollonov, V. V., Kiiko, V. V., Kislov, V. I., Suzdal’tsev, A. G., and Egorov, A. B. 2003. “High-Frequency Repetitively Pulsed Operating Regime in High-Power Wide-Aperture Lasers.” Quantum Electron 33 (9): 753. Grachev, G. N., Ponomarenko, A. G., Smirnov, A. L., Statsenko, P. A., and Tishchenko, V. N. 2005. “A Pulsating Optical Discharge Moving in a Gas.”Kvantovaya Elektron 35 (11): 973.

829

[20] Apollonov, V. V. 2009. Proceedings of the X Intern. Conf. on Photonics and Optoelectronics, 13. [21] Apollonov, V. V. 2009. Method for the formation of conducting channels in a nonconducting medium. RF Patent 2,400,005, filed May 20, 2009. [22] Fuks, N. A. 1955. Mechanics of Aerosols. Moscow: Izd-vo AN SSSR. [23] Ageev, V. P., Barchukov, A. I., and Bunkin, F. V. 1977. “Laser Air-Breathing Jet Engine.” Sov. J. Quantum Electron 7 (12): 1430. [24] http://www.nanosized-powders.com. [25] Rukhadze, A. A., and Shpigel’, N. S. 1965. Electric Explosion of Conductors. Moscow: Mir, 341. [26] Gerasimenko, N. I., Grashina, N. A., Medvedkov, A. G., Meshcheryakov, A. B., and Pletnev, N. V. 1988. Prib. Tekh. Eksp. 1: 212. [27] Goncharenko, G. M. 1963. “Electric Power Engineering.” In MEI. Moscow: Izd-vo MEI, 146. [28] Gavrilov, V. N., and Litvinov, E. A. 1993. Zh. Prikl. Mekh. Teor. Fiz. 34: 28. [29] Kvartskhava, I. F., Plyutto, A. A., Chernov, A. A., and Bondarenko, V. V. 1956. Zh. Eksp. Teor. Fiz. 30: 42. [30] Martynyuk, M. M. 1974. “Transport of Metal in Cathode Plasma Cluster at Explosion Emission of Electrons from Metallic Edges. II. Multiple Switching of Current.” Zh. Tekh. Fiz. 44 (6): 1262. [31] Peregud, B. P., and Abramova, K. B. 1964. Dokl. Akad. Nauk SSSR 157: 837. [32] Abramova, K. B., Valitskii, V. P., Vandakurov, Y. V., Zlatin, N. A., and Peregud, B. P. 1966. Dokl. Akad. Nauk SSSR 167 (4): 778. [33] Protopopov, N. A., and Kul’gavchuk, V. M. 1961. Zh. Tekh. Fiz. 31: 557. [34] Aleksandrov, A. F., Zosimov, V. V., Kurdyumov, S. P., Popov, Y. P., Rukhadze, A. A., and Timofeev, I. B. 1971. “Dynamics and Radiation of Direct High-Current Discharges in the Atmosphere.” Zh. Eksp. Teor. Fiz. 61 (November): 1841. [35] Aleksandrov, A. F., and Rukhadze, A. A. 1976. Physics of High-Current Electric-Discharge Light Sources. Moscow: Atomizdat. [36] Komel’kov, V. S. 1947. Dokl. Akad. Nauk SSSR 58: 57. [37] Vlasto’s, A. E. 1967. “Current Pause in Exploding-Wire Discharges.” J. Appl. Phys. 38 (13): 4993. [38] Vlasto’s A. E. 1968. “Restrike Mechanisms of Exploding Wire Discharges.” J. Appl. Phys. 39 (7): 3081. [39] Apollonov, V.V., and Pletnev, N. V. 2004. Method for the inhomogeneous extraction of the energy of free higher transverse lasing modes from a laser and a laser. RF Patent 2,239,921, filed November 10, 2004. [40] Apollonov, V.V. 2014. “High Power P-P Lasers.” NOVA, N-Y.

Call for Papers Dear author, This is Journal of Energy and Power Engineering (ISSN1934-8975), created in Dec. 2007. We are interested in various kinds of professional papers about energy and power. The Journal of Energy and Power Engineering, a monthly professional academic journal, is striving to provide the best platform for researchers and scholars worldwide to exchange their latest findings and results.

Current Columns Involve Thermal Science; Fluid Mechanics; Energy and Environment; Power System and Automation; Power Electronic; High Voltage and Pulse Power; Sustainable Energy.

Information for Authors 1. The manuscript should be original, and has not been published previously. Please don’t submit material that is currently being considered by another journal. 2. Manuscripts may be 3000-8000 words or longer if approved by the editor, including abstract, text, tables, footnotes, appendixes, and references. The title should not be exceeding 15 words, and abstract should not be exceeding 150 words. 3-8 keywords required. 3. The manuscript should be in MS Word format, submitted as an email attachment to our email address.

Editorial Procedures All papers considered appropriate for this journal are reviewed anonymously by at least two outside reviewers. The review process usually takes 4-5 weeks. Papers are accepted for publication subject to no substantive, stylistic editing. The Editor reserves the right to make any necessary changes in the papers, or request the author to do so, or reject the paper submitted. A copy of the edited paper along with the first proofs will be sent to the author for proofreading. They should be corrected and returned to the Editor within seven days. Once the final version of the paper has been accepted, authors are requested not to make further changes to the text.

Submitting of Manuscripts All manuscripts submitted will be considered for publication. Please visit our website at www.davidpublisher.com for our automatic paper submission systems or as an email attachment to [email protected], [email protected], [email protected] and [email protected].

Journal of Energy and Power Engineering David Publishing Company Tel: 001-323-984-7526, 001-323-410-1082; Fax: 001-323-984-7374

Integration of Energy and Fire Prevention Systems in Green Building

Integration of Energy and Fire Prevention Systems in Green Building

Suggest Documents

Solar thermal energy systems for building integration

green building rating systems - MyFlorida Green Building

green building rating systems - MyFlorida Green Building

Fire prevention – European classification of building ... - BioFach

âGreen Buildingâ â Modelling renewable building energy systems and ...

GREEN ENERGy INTEGRATION IN SOUTH AMERICA

Energy Systems Integration - NREL

hbim and fire prevention in historical building heritage ... - IAARC

TITLE 675 FIRE PREVENTION AND BUILDING ... - State of Indiana

TITLE 675 FIRE PREVENTION AND BUILDING ... - State of Indiana

Building integration of solar renewable energy ...

Building integration of solar renewable energy ...

Energy Systems Integration Partnerships: NREL + CSIRO, Energy ...

Control in Building Energy Management Systems

Modelling and optimization of multi-energy source building systems in ...

Grid Integration of Renewable Energy Systems

FIRE PREVENTION GUIDELINE and PRACTICES - NMSU Fire ...

Integration of green roof and solar photovoltaic systems - HKU

evaluation of renewable energy system in china's green building ...

Energy Saving of a Green Building in Egypt

A Green Energy Application in Energy Management Systems by an ...

Integrating Fire Systems with Other Building ... - patapsco.nist.govâ¦

Building optimisation for energy system integration

Energy Saving and Green Building Design of Libraries - IFLA.ORG

Integration of Energy and Fire Prevention Systems in Green Building