Hydro-methane and methanol combined production ...

Energy Conversion and Management 79 (2014) 74–84

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Hydro-methane and methanol combined production from hydroelectricity and biomass: Thermo-economic analysis in Paraguay M. Rivarolo a,⇑, D. Bellotti a, A. Mendieta b, A.F. Massardo a a b

Thermochemical Power Group (TPG) – DIME, University of Genoa, Italy Parque Tecnologico de Itaipu (PTI), Paraguay

a r t i c l e

i n f o

Article history: Received 30 October 2013 Accepted 27 November 2013 Available online 29 December 2013 Keywords: Thermo-economic optimization Water electrolysis Bio-methanol production Renewable energy sources Synthetic Natural Gas production

a b s t r a c t A thermo-economic analysis regarding large scale hydro-methane and methanol production from renewable sources (biomass and renewable electricity) is performed. The study is carried out investigating hydrogen and oxygen generation by water electrolysis, mainly employing the hydraulic energy produced from the 14 GW Itaipu Binacional Plant, owned by Paraguay and Brazil. Oxygen is employed in biomass gasification to synthesize methanol; the significant amount of CO2 separated in the process is mixed with hydrogen produced by electrolysis in chemical reactors to produce hydro-methane. Hydro-methane is employed to supply natural gas vehicles in Paraguay, methanol is sold to Brazil, that is the largest consumer in South America. The analysis is performed employing time-dependent hydraulic energy related to the water that would normally not be used by the plant, named ‘‘spilled energy’’, when available; in the remaining periods, electricity is acquired at higher cost by the national grid. For the different plant lay-outs, a thermo-economic analysis has been performed employing two different software, one for the design point and one for the time-dependent one entire year optimization, since spilled energy is strongly variable throughout the year. Optimal sizes for the generation plants have been determined, investigating the influence of electricity cost, size and plant configuration. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction As a developing country, Paraguay faces many challenges regarding the improvement of its energy systems. Although Paraguay has a very large hydroelectric capacity, energy demand is sustained mostly by imported hydrocarbons, especially in the transport field. Considering the low cost of electricity, compared with developed countries, the production of fuels starting from exceeding renewable energy may represent an interesting solution. One of the main actors can be hydrogen, produced by water electrolysis employing the electricity of the largest renewable production plant in the world, the 14 GW hydroelectric facility of Itaipu (owned 50% by Brazil and 50% by Paraguay, meaning that 7 GW are property of Paraguay) [1]. Hydrogen is one of the most promising energy vectors in the near future [2,3]. However, its low energy to volume ratio (compared to gasoline and diesel) represents an obstacle towards ⇑ Corresponding author. E-mail address: [email protected] (M. Rivarolo). 0196-8904/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.enconman.2013.11.044

its large scale diffusion in the transportation sector. Other critical aspects are related to its distribution (pipelines and road transportation) and final use, given the high cost of fuel cells or ICE hydrogen powered vehicles [4]. Hydrogen conversion into other fuels or ‘‘chemicals’’ may represent a solution to bypass the limitations related to its low energy to volume ratio. In the Sabatier reaction hydrogen is mixed with carbon dioxide in order to produce methane, according to the Sabatier reaction:

4H2 þ CO2 ¼ CH4 þ 2H2 O

ð1Þ

The reaction is highly exothermic (DH = 165 kJ/mol) and proceeds in a temperature range between 250 and 400 °C at relatively low pressures (between 2 and 10 bar) on a catalytic bed, usually based on Nickel, Ruthenium or Aluminum [5]. Previous experimental studies [5,6] have shown that the final composition of gas, after water separation, is mainly CH4, plus H2 and CO2 not reacted. The presence of hydrogen allows for classifying the gas as hydromethane, that is a gas mixture composed mainly by methane and containing H2 in a range from 5% to 30% in volume terms. This fuel can be transported and used with technologies similar to the ones

M. Rivarolo et al. / Energy Conversion and Management 79 (2014) 74–84

75

Nomenclature

Abbreviation AEC Alkaline Electrolytic Cell ASU Air Separation Unit CDE Ciudad Del Este (Paraguay) DC Direct Costs FCI Fixed Capital Investment HHV Higher Heating Value ICE Internal Combustion Engine LHV Lower Heating Value PBP Pay Back Period PEC Purchased Equipment Cost PTI Parque Tecnologico de Itaipu TCI Total Capital Investment TPG Thermochemical Power Group WGS Water Gas Shift

Symbols C cost (€) E electricity flow (kW h) M mass flow (kg/h) Subscript acq cap cons i inst prod req var

employed for methane, bypassing the problems related to hydrogen. In particular, hydro-methane can be employed in traditional natural gas ICE powered vehicles that are significantly cheaper compared to fuel cells powered vehicles. The considerable amount of carbon dioxide (5.5 kg for each kg of H2) for the Sabatier reaction is obtained from biomass gasification, employing the large amount of pressurized O2 produced in electrolyzers as gasification agent. As mentioned in a previous study [7], the syngas obtained by biomass gasification, after cleaning and WGS section, is composed of H2 and CO2; after mixing with H2 produced in electrolyzers, the gas has the optimal composition for the Sabatier reaction, in order to synthesize hydro-methane [7]. One of the innovative aspects of this study is the combined production of hydro-methane and methanol in the same plant. Methanol production by biomass gasification is a well known process, as reported in [8,9]: during the process, large amounts of CO2 have to be separated in a Selexol unit. The remaining syngas, made of H2 and CO, is sent to methanol reactor, where the following reaction takes place:

2H2 þ CO ! CH3 OH

ð2Þ

Methanol is widely used in several current industrial applications. Although in Paraguay its demand is actually limited, Brazil is one of the largest consumers of methanol of the world: its homeland production satisfies only 30% of the internal demand, therefore it represents an ideal market for methanol exportation [10]. In the first part of the work a design point thermo-dynamic and energetic analysis of the plant for hydro-methane and methanol combined synthesis is performed. The results of this analysis (in terms of energy and mass flows) are employed to perform a time-dependent thermo-economic analysis during an entire oneyear period in order to evaluate the plant optimal size and management, for different electricity cost scenarios in Paraguay. The thermo-economic optimization is carried out using two different software, one for the design point thermodynamic and chemical analysis, named WTEMP (Web-based Thermo-Economic Modular Program) and one for time-dependent plant management optimization, named WEPoMP (Web-based Economic Poly-generative Modular Program), both developed by the TPG at University of Genoa. 2. Design point thermodynamic analysis of the process The thermodynamic plant analysis and optimization at the design point is carried out employing the WTEMP software,

acquired capital consumed ith time step installed produced required variable

developed by TPG at University of Genoa in the last 25 years [11,12]. WTEMP adopts a modular approach and a standard component interface, that allows the user to build complex cycle configuration in a short time. This approach maintains the flexibility and the extendibility of the library components (more than 90 modules are available at the moment), allowing users to add new components without modifying the core of the code [13,14]. Each component is described by three subroutines, that define its thermodynamic, exergetic and thermo-economic properties at the design point. This paper focuses on the thermodynamic analysis of the methanol and hydro-methane synthesis. The simplified plant layout, not including heat exchangers, is shown in Fig. 1. Oxygen and hydrogen are produced by 1 MWe commercially available alkaline electrolysers, operating at 80 °C and 30 bar, with an electrical consumption of 4.7 kW h/ N m3 of H2 corresponding to an efficiency of 75% based on HHV [15]. Higher efficiencies could be achieved with proton exchange membrane or solid oxide electrolysers, but they are not available on the market at competitive costs, in particular for large sizes investigated in this work: for these reasons, alkaline electrolysers (AECs) are chosen. Oxygen, produced at 30 bar, is employed for biomass gasification; the syngas exiting the gasifier at 1000 °C, is cooled and sent to a cleaning section including a scrubber and a Water Gas Shift (WGS). During the gas cooling, super-heated steam needed for WGS and biomass gasification is generated, without need for an external boiler, as reported in [8]. After the cleaning process, the gas is cooled and water is separated, therefore the syngas is made of H2, CO and CO2 mainly. In order to obtain stoichiometric conditions at the inlet of methanol reactor, CO2 must be removed in a Selexol unit, in order to obtain the composition reported in Eq. (2). The gas obtained in this way is compressed to 100 bar before entering the methanol synthesis reactor, operating at 240 °C: at the outlet, the gas is cooled and methanol is condensed downstream the reactor, while the most of the remaining syngas (85%) is sent back to the reactor. The process described above is quite well known and commercially available: it is worth noting that, in the present configuration, O2 for biomass gasification is available from water electrolysis, therefore the installation of an Air Separation Unit, normally used in this kind of plants, is avoided. The innovative aspect of this process is that CO2 separated by Selexol is mixed with H2 produced by pressurized alkaline electrolysers and then sent to Sabatier reactors to produce hydro-methane, according to Eq. (1). Since Sabatier reaction is highly exothermic, heat

76


Fig. 1. Simplified plant for methanol and hydro-methane synthesis (configuration A, with CO2 excess).

accumulation may destroy catalysts, therefore a cooling flow must be considered. In the present analysis the reactor is assumed isothermal and temperature is controlled by cooling the reactant bed in a heat exchanging mechanism with a coolant flow around the reactor. Since it is cheap and largely available close to hydroelectric facilities, water is assumed as coolant flow in the present analysis. After water separation downstream of the reactor, the gas produced is mostly composed of CH4, but also with residual H2 and CO2, which quantities depend on H2/CO2 ratio at the inlet and on reactor temperature. The thermodynamic analysis is performed considering the gas compositions as constraints at the inlet of both hydro-methane and methanol reactors, as reported in Eqs. (1) and (2). The analysis is performed at different gasifier operative conditions to find out the optimal amounts of oxygen and biomass employed in the process to obtain the optimal syngas composition [6]. The main parameters assumed for the plant simulation are reported in Table 1. Fig. 2 shows an alternative plant configuration investigated with WTEMP to reduce biomass consumption: the plant is optimized in order to produce only the CO2 needed for Sabatier reaction, without any excess vent to atmosphere. This configuration allows for a reduction of the biomass gasification section: not all the oxygen is employed in the gasifier and biomass consumption is consequently reduced; on the other hand, methanol production gets lower too. It is worth noting that hydro-methane synthesis section is not modified compared to the previous case. In both the configurations, the gas composition at the outlet of the Sabatier reactor is 77% CH4 19% H2 and 4% CO2. Table 2 compares WTEMP main results for the two different plant layouts: Configuration A refers to the process shown in Fig. 1, where all the O2 is sent to gasifier; Configuration B refers to the process shown in Fig. 2, where all the CO2 separated in the Selexol is sent to the Sabatier reaction, without any excess vent

Table 1 Parameters used for plant simulations in WTEMP. Alkaline Electrolyser Electrical consumption Pressure Temperature Gasifier Pressure Syngas exit temperature Oxygen to biomass mass ratio Oxygen to steam mass ratio Biomass

4.7 kW h/N m3 di H2 30 bar 80 °C 30 bar 1000 °C 1:3 1:1 Wood pellet 49% C 41% O 7% H 3% N LHV 16.0 MJ/kg

WGS Pressure Temperature Steam to syngas mass ratio

20 bar 275 °C 1:11

Selexol Working pressure Temperature

35 bar 50 °C

Methanol Reactor Working pressure Temperature Recirculation factor of unreacted syngas

100 bar 240 °C 0.85

Sabatier Reactor Gas exit pressure Operative temperature Inlet composition (volume)

10 bar 270 °C 80% H2 20% CO2

Compressor Isentropic efficiency Mechanical efficiency

86% 99%

to atmosphere. The analysis is performed considering a reference size of 172 MW of AECs, which corresponds to a production of 1 kg/s of H2.


77

Fig. 2. simplified plant for methanol and hydro-methane synthesis (configuration B, without excess CO2).

Table 2 WTEMP results (reference size of 172 MW).

H2 from AECs O2 from AECs O2 to atmosphere O2 to gasifier Biomass inlet CO2 separated CO2 to atmosphere CO2 to mixer CH3OH produced Hydro-methane produced AECs installed power Auxiliaries power

3. WEPoMP thermo-economic optimization approach

Configuration A

Configuration B

1 kg/s 8 kg/s – 8 kg/s 24 kg/s 26.8 kg/s 21.3 kg/s 5.5 kg/s 11.1 kg/s 2.2 kg/s 172 MW 6.2 MW

1 kg/s 8 kg/s 6.3 kg/s 1.7 kg/s 5.1 kg/s 5.5 kg/s – 5.5 kg/s 2.6 kg/s 2.2 kg/s 172 MW 1.4 MW

Table 2 shows a comparison of the results for the two cases: considering the same electricity input and the same AECs size, H2 and O2 produced are obviously the same. In Case A, all the O2 is employed for biomass gasification, therefore a large amount of biomass is needed. It is important to note that only a lower part of the significant amount of CO2 separated is needed in the Sabatier reactor for the mix with H2. The most of the CO2 separated is not employed and vent to atmosphere. In configuration B, the system is optimized in order to produce the CO2 needed for Sabatier reaction, reducing the size of the methanol plant. Only 22% of the oxygen is sent to gasifier, therefore both biomass consumption and methanol produced are proportionally reduced. It is worth noting that hydro-methane production does not vary; only the methanol section is modified. To sum up, configuration B allows for a reduction of biomass consumption; on the other hand, methanol production gets lower and a consistent amount of oxygen (more than 75%) is not employed: since its high purity, this amount may be sold to hospitals or industries, eventually increasing plant global revenues.

Since the electrical energy to feed electrolysers is produced by a time-dependent, not fully controllable renewable source (hydroelectric), a thermo-economic time-dependent analysis is needed to optimize both plant size and management. To study the optimal size of the plant, the software WEPoMP (Web-based Economic Poly-generation Modular Program) developed by TPG at University of Genoa for time-dependent thermo-economic analysis of poly-generation power plants has been employed [16,17]. WEPoMP is characterized by a modular approach and a standard component interface, which allows the user to build complex cycle configuration in a short time. This approach maintains the flexibility and the extendibility of the library components (41 modules are available at the moment), allowing the user to add new components without modifying the core of the code. Each component is described by three subroutines, which define mass and energy flows, off-design performance curves, variable and capital costs. WEPoMP software is provided with cost equations which evaluate the capital cost of the single components of the plant on the basis of the installed power (for gas turbines, combined cycles, fuel cells, electrolytic cells, etc.) or on the basis of the mass flow (for chemical reactors, gas cleaning, etc.). Cost functions allow to calculate the Purchased Equipment Cost for each unit, which usually depends on its size (in terms of flow rate or installed power). From the PEC, the Total Capital Investment (TCI) of the plant can be evaluated: as shown in Table 3, the most of the voices which constitute TCI are calculated as a percentage of PEC; the remaining ones are calculated as a function of Fixed Capital Investment (FCI) or Direct Costs (DC) of the plant. The economic input file contains also a section describing the economic scenario where the plant is operating, including inflation indexes, construction period, plant lifetime and depreciation.

78


Table 3 Economic input file in WEPoMP. Parameters for TCI evaluation

Economic scenario parameters

Startup costs Working capital Licensing, research and development Engineering and supervision Construction costs Civil and architectural work Service facilities Purchased equipment cost installation Piping Instrumentation and controls Electrical equipment and materials

5% of FCI 15% of FCI 7% of FCI 25% of PEC 15% of DC 15% of PEC 30% of PEC 20% of PEC 10% of PEC 6% of PEC 10% of PEC

Economic data reference year Construction starting year Construction time Plant lifetime Depreciation time Inflation rate Nominal escalation rate for revenues Nominal escalation rate for O&M Average income tax rate Maintenance/Operating factor

Two different optimization levels can be investigated in WEPoMP: low and high level, respectively. At low level, layout and plant size are considered fixed (therefore capital costs are constant) and the best operational strategy is determined to minimize variable costs throughout the year: variable costs include fuel (and biomass) consumption costs and electrical energy costs. The optimization process is performed taking into account the problem’s constraints, which are the balance equation between supply and demand of components. For example, the energy balance includes the energy produced by the prime movers (ICE, gas turbines, hydroelectric plants, etc.) in the system, the energy sold to the user and the energy consumed by system components (i.e. electrolytic cells).

Ereq

N N X X ¼ Ei;prod þ Eacq þ Ev irt Ei;cons i¼1

ð3Þ

i¼1

At high levels, it is possible to evaluate the optimal size of the plant components while minimizing total annual costs, which are the sum of variable costs Cvar calculated at low level analysis and capital plant costs Ccap.

C tot ¼ C v ar þ C cap

ð4Þ

where the total capital costs of the plant are the sum of the total capital costs of all plant components. Therefore, the optimization procedure is carried out to simultaneously obtain the plant and the component sizes, together with the plant operation. The optimal size value for the desired component is found between two limit values set by the user: at every iterative cycle, total annual costs are calculated by changing the value of the nominal size of the component. For both high level and low level optimizations, the code goal is to minimize total annual plant cost. At low levels, the code minimizes variable costs only; at high optimization levels, the software takes into account fixed costs as well, finding the optimal size for plant components. In both the cases, energy and/ or fuel demands, which represent the constraints of the problem, must be satisfied for all the representative periods of the year. Since WEPoMP is employed for time-dependent thermo-economic analysis, the calculation is usually carried by out by dividing the operational time (usually a year) in a sufficient number of representative periods (one hour or less depending on the particular application). 3.1. Cost functions In order to perform the economic analysis, the evaluation of cost functions representative of the different parts of the plant is mandatory to determinate the PEC and the TCI. New cost functions have been implemented in the code, namely for gas cleaning, Selexol and methanol synthesis sections. The cost function for gas cleaning is referred to a gasification plant for methanol synthesis from biomass gasification able to

2013 2013 1 year 20 years 10 years 3% 2.5% 2.5% 30% 1.04

process 134,000 kg/h of raw syngas [18]. Costs of cyclones, tar cracker, scrubber, WGS and Selexol are included in the function:

C cap ¼ 54:3 106

0:65 Min 134; 000

ð5Þ

The cost function for methanol production section is referred to the same gasification plant, where 54,000 kg/h of clean syngas are entering in the synthesis section [18]. The costs of compressor, synthesis reactor and distillation unit are included in the function:

C cap ¼ 14:2 106

0:65 Min 54; 000

ð6Þ

The cost function implemented in the software for pressurized AECs is extrapolated from literature data as reported in [15], where the coefficient a takes into account additional costs for auxiliaries: a conservative value of 10% has been assumed for a in the present analysis.

C cap ¼ a 37; 455 P0:4832 inst

ð7Þ

Capital costs for large size oxygen blown pressurized (30 bar) biomass gasifier are obtained from literature data, referring to a 389 MWth gasification plant as reported in [19]:

C cap ¼ 28:2 106

0:6 Pinst 389; 000

ð8Þ

4. Economic scenario Paraguay is considered as the main target of this study. The estimated population in 2012 was around 6.7 million of inhabitants. The main urban and financial regions are the metropolitan areas of Asunción and Ciudad Del Este, where around half of the population of the Country is concentrated. Paraguay has 1,063,262 registered vehicles, according to [20]: half of these are passenger cars. Considering total annual fossil fuel import data [21] around 30% of these (150,000 cars) employs gasoline fueled ICEs. This type of vehicles is considered as main consumers for the produced hydro-methane. Half of these are circulating in Asuncion (48,500 cars) and in Ciudad Del Este (24,500 cars), which represent the targets of the study. Ciudad Del Este is located in Alto Paraná, on the border between Paraguay, Brazil and Argentina: due to its proximity to the Itaipu hydroelectric power plant (less than 20 km) it represents the most strategic location for the installation of a large plant: in this way, methanol produced in the plant can be sold to Brazil, which is the main market for this chemical in South America. Three main scenarios are taken into consideration. The first scenario is represented in Fig. 3: hydro-methane and methanol are produced in CDE, all the hydro-methane necessary for CDE and Asuncion areas is produced only in CDE, employing the CO2 separated by Selexol after biomass gasification, and then transported in Asuncion by trucks.


79

Fig. 3. Scenario 1 (hydro-methane production only in Ciudad Del Este).

The time-dependent analysis is performed considering different electricity costs to feed AECs, depending on the use of lower cost spilled energy from Itaipu hydroelectric plant or higher cost grid energy. In order to consider different viable hydro-methane consumer population, four different system sizes are considered (45, 90, 135 and 180 MW, referred to AECs installed power). The previous analysis in WTEMP has evidenced two possible approaches for the system. In this scenario, two different configurations are considered for all the plant sizes: (A) Employing all the oxygen produced in CDE in order to design a larger gasifier, that will increase biomass consumption but also will improve the production of methanol (as shown in Fig. 1). (B) Optimizing the size of the gasifier according to the total required hydro-methane (Fig. 2), a part of the O2 produced by the AECs installed in Ciudad Del Este is vent to atmosphere but there is no surplus of CO2.

The second scenario is represented by Fig. 4: hydro-methane and methanol are main products in Ciudad Del Este (CDE), CO2 is transported by ships [22,23] to Asuncion, where it is mixed with H2 produced by electrolysers to synthesize hydro-methane. O2 co-produced in Asuncion is not employed. WEPoMP approach is used to investigate different sizes in CDE and Asunción as well, considering that the cars circulating in Asuncion are about two times compared to CDE. An interesting concept investigated is that the CO2 needed for the hydro-methane synthesis (Eq. (1)) is separated in Selexol (see Fig. 1) and transported to Asuncion, where only the electrolysers are considered to be installed. The influence of electricity cost on economic results is investigated, considering different electricity costs to feed AECs installed in Asuncion. It is worthy to observe that energy is a time-dependent parameter in CDE, while in Asuncion, where spilled energy from Itaipu is not available, it is constant throughout the year. Also in this scenario both configurations A and B are investigated.

Fig. 4. Scenario 2 (CO2 transportation and hydro-methane production in Asuncion).

80


Finally, a third scenario is considered to investigate the influence of selling the oxygen produced in Asuncion, which is not employed for biomass gasification as in CDE: the analysis is performed for all the scenarios previously investigated. It is worthy to note that O2 excess is strongly variable with the configuration adopted, as can be observed from WTEMP results in Table 2. Table 4 shows all the different scenarios investigated in the present paper. WEPoMP software needs a large amount of data, related to the technical and economic scenario where the plant is going to operate. The main plant data are as follows: (a) River water flow represents the main input to the hydroelectric plant, strongly influencing the net electrical power and the spilled power. Values are highly time-dependent, therefore spilled power also varies with time. Paraná River water flows from Itaipu data [24] have been considered in the present analysis. (b) Electricity load curves are very important, since they represent one of the optimization problem constraints. The software receives the time dependent electrical load demand as input, which must be satisfied at each period of the year using electricity produced by the Itaipu plant. Excess electrical energy (spilled energy), when available, is employed by alkaline electrolysers (AECs) to produce H2. In this analysis, since Brazil and Paraguay are the two countries which own the hydraulic plant, load curves relative to electricity amounts that must be satisfied by Itaipu are considered: the hourly load distribution is assumed based on data available at PTI [24,25]. From the combination of water flows and electrical demands, time-dependent spilled power average monthly values can be evaluated: Fig. 5 reports average values of spilled power for some representative days of February, November and June, when the average flow rates of Paraná river are 15,000 11,500 and 10,000 m3/s respectively [24]: it is worth noting that spilled power value depends on both available water flow and electrical demand curves: water flow is affected by the season, electrical load demands are strongly variable throughout the day, therefore spilled power available is a strongly time-dependent parameter and a one year hour by hour optimization is mandatory. It is worth emphasizing spilled power distribution time dependent nature, since it is affected by both electrical demand and available water flow which vary greatly throughout the period. Therefore, hydrogen production that uses only spilled power would be strongly time-dependent too, as reported in previous analysis performed in [1,6]. (a) Energy cost represents a term of primary importance to find the optimal system size. Regarding Itaipu (Ciudad Del Este plant), two different energy costs are considered; the first one is the spilled energy cost that assumes a value of 0.005 €/kW h since spilled water is a surplus energy usually widely available at low cost. When spilled energy is not available, electrical energy to power

AECs must be drawn from the electrical grid directly from Itaipu power plant at a higher cost of 0.007 €/ kW h. For the plant located in Asuncion however, three different energy costs are also considered: – 0.014 €/kW h, which is the current cost for plants of 5 MW maximum installed power [24], is assumed as a possible cost considering the renewable philosophy of the plant; – 0.020 €/kW h, assumed as a medium cost scenario; – 0.025 €/kW h, which is the normal grid electricity for electro intensive plants, as reported in [26]. (b) Biomass cost is assumed equal to 40 €/ton, which represents a typical market value, as reported in [9]. (c) Oxygen selling price highly depends on its diffusion on the market, which is related to local conditions, applications, etc. In the case under analysis, O2 price is assumed 150 €/ton, which represents the minimum selling price for industrial use of oxygen (rates are higher for medical use) [25]. (d) Hydro-Methane selling price is assumed to be 1.00 €/ kg, which is a realistic value based on methane selling price at refueling stations. (e) Methanol selling price is assumed to be 400 €/ton, which is the 2013 market price, as reported in [27]. 5. Thermo-economic results Time-dependent plant analysis is performed considering an entire one-year period (8760 hourly periods) for different plant sizes. The analysis is carried out employing spilled energy, when available, to power AECs; in the remaining periods, the possibility of taking energy from the grid is chosen in order to produce constant H2, O2, methanol and hydro-methane mass flows. Fig. 6 reports the power from electrical grid to feed AECs vs time, for the different sizes considered (45, 90, 135 and 180 MW). It is worthy to observe that periods where electrical energy has to be fully bought are constant and equal to 33% of the year (when spilled power is equal to zero), while number of periods that partially require energy from the Paraguay grid are affected by system size: therefore, increasing AECs size determines an increase in variable costs for electricity taken from the grid at higher price. However, as Fig. 6 shows, the increase is not very strong for the sizes considered in the present analysis: at the lowest and highest sizes of 45 and 180 MW, spilled energy is sufficient to power all the installed AECs for respectively 60% and 57% of the year. The thermo-economic time-dependent analysis has been performed considering the three scenarios described in the previous section: the main economic assumptions for WEPoMP analysis are reported in Fig. 3, they are obviously the same for all the configurations investigated. Since plant lifetime is 20 years, solutions characterized by a Pay Back Period higher than 10 years are considered not economically feasible in this analysis. Fig. 7 reports economic results related to scenario 1, where hydro-methane and methanol are produced in a large size plant

Table 4 Scenarios investigated in WEPoMP. Scenario

Electricity cost

Plant size

Configuration

O2 price

1

CDE energy 5/7 €/MW h

CDE 45-90-135-180 MW

–

2

CDE ASU CDE ASU

CDE ASU CDE ASU

A (CO2 excess) B (O2 excess) A (CO2 excess) B (O2 excess) A (CO2 excess) B (O2 excess)

3

energy energy energy energy

5/7 €/MW h 14, 20, 25 €/MW h 5/7 €/MW h 14, 20, 25 €/MW h

15-30-45-90 MW 30-60-90-120 MW 15-30-45-90 MW 30-60-90-120 MW

– 150 €/ton


81

Fig. 5. Spilled power distribution.

Fig. 6. Power from electrical grid to power AECs vs time (for different sizes).

located in CDE, according to plant lay-out shown in Fig. 4. Configurations A and B previously investigated with WTEMP are compared from an economic point of view: since the strong amount of methanol produced, configuration A allows for significantly higher revenues; annual costs are increased as well, due to larger size of gasification and cleaning sections and consequently higher biomass consumption. However, the effect of higher methanol production is preponderant, therefore configuration A appears to be the better solution from an economic point of view, as confirmed by lower PBP values. In general, it is worth observing from function costs (Eqs. (5)–(8)) that capital costs for installed kW decrease as plant size increases, therefore larger sized plants have a lower PBP and represent the best economic solution. Since the plant is located close to Itaipu, the availability of low cost electricity allows for reduced values of PBP, in each case inferior than 10 years, which represents the minimum acceptable value. Fig. 8 shows Pay Back Period related to scenario 2 for the different plant sizes investigated: since the ratio of cars circulating in

Asuncion and CDE is about 2, the same ratio is considered for plant sizes as well. Electricity cost in CDE is the same considered for the previous scenario analysis. The strong influence of electricity grid cost in Asuncion is evident: assuming a size of 30 MW in CDE and 60 MW in Asuncion, the plant would not be economically feasible for the ‘‘standard’’ electricity cost of 0.025 €/kW h (PBP equal to 11 years); decreasing electricity cost to 0.020 and 0.014 €/kW h would lead to a significant reduction of PBP to 10 and 9 years respectively, therefore the plant would represent a viable economic solution. In Fig. 8 the influence of the gasifier reduction (configuration B) on PBP is also represented at 0.025 €/ kW h electricity cost for different plant sizes. As results suggest, reducing gasifier and methanol section affects negatively the PBP: for scenario 2, it does not represent a viable economic option. Fig. 9 compares configurations 2A and 2B from an economic point of view, showing annual costs and revenues for the two cases. Although the reduction of gasifier size allows for lightly reducing biomass consumption and equipment capital costs,

Fig. 7. Economic results comparison for scenario 1 (configurations A and B).

82


Fig. 8. PBP vs plant size, influence of electricity cost in Asuncion (scenario 2).

revenues significantly decrease due to the reduction in methanol production (see Table 2). It is worth observing that, since oxygen is not sold to any user, the exceeding amount of O2 does not represent a benefit in the present scenario. In Fig. 9 the impact of each equipment is also represented (for case A); for the case B, gasifier, gas cleaning and methanol reactor are reduced, but AECs are the same; since the high influence of electrolysers, capital costs are not strongly reduced. Fig. 10 compares the results of configuration A for both scenarios 1 and 2. From an economic point of view, the superiority of solution 1A, due to higher methanol revenues, is evident, as confirmed by PBP values, which are the half compared to scenario

2A. However, it is worth comparing mass flows as well. In solution 1A the whole plant is located in CDE, meaning that 180 MW of AECs are installed in the same place, while in case 2A 60 MW are installed in CDE and 120 MW in Asuncion. Therefore, configuration 1A allows for larger amounts of O2: gasifier section is larger, resulting in higher amounts of biomass consumed (about three times more). In configuration 1A about 80 ton/h are needed, equivalent to 700,000 ton/year. Since in Paraguay the average growth of wood biomass is 100 ton/ha [24], and considering a land rotation of 5 years for biomass growth [24], about 35,000 ha would be required. In scenario 2A about 12,000 ha would be sufficient. Biomass are largely available in Paraguay scenario, but, since the high

Fig. 9. Annual costs and revenues, configurations A and B (scenario 2).

Fig. 10. economic and mass flows comparison (scenarios 1A and 2A).


83

Fig. 11. Economic results for scenario 3 (configurations A and B).

Fig. 12. Economic results comparison.

amount required, not the biomass may be provided close to the hydroelectric facility, therefore transportation costs of biomass may increase, causing adjunctive costs for the plant management. Fig. 11 compares configurations 3A and 3B from an economic point of view, showing the influence of electricity cost on PBP: reducing electricity cost, PBP decreases. It is worth noting that, considering the same electricity cost, scenarios 3A and 3B present the same PBP, meaning that the two solutions are equivalent from an economic point of view. The equivalence is more evident comparing annual costs and revenues for the two cases. In this scenario O2 selling is allowed: therefore configuration B, characterized by higher amount of produced oxygen (vent to atmosphere in the previous scenarios), becomes more economically attractive. Fig. 12 shows that revenues of case 3B are very close to revenues of case 3A, since O2 sale balances the reduction of revenues from methanol sale. Moreover, in case 3B capital costs and biomass consumptions are lower, since size of the gasifier reduces. Although solutions 3A and 3B are economically equivalent, solution 3A is the best one from a strategic point of view, since methanol market in South America is more developed, therefore this product is easier to be sold, compared to O2, whose market is limited at the present moment. In Fig. 12 the best configurations for the different scenarios are compared: scenario 2A presents the worst results, as confirmed by values of PBP higher than 10 years, in particular for low and medium sizes, therefore it is not acceptable. In case 3A, considering oxygen selling, results get significantly better: PBP gets lower, varying from 10 years for low sizes to 6 years for large sizes. Oxygen selling allows for an increasing of annual revenues of about

20 M€ for year, considering to sell it at 150 €/ton, which represents an average value for industrial applications. Scenario 1A presents the best results: thanks to larger amounts of methanol, sold at an average price of 400 €/ton [27], revenues gets significantly higher, leading to a strong reduction of PBP. Configuration 1A seems to represent a worthy solution also from a strategic point of view, considering that methanol would be sold to Brazil, which represents the largest methanol consumer in South America. However, it is worth remembering that biomass consumption to produce the high amounts of CH3OH also increases in configuration 1A, about 75 ton/h of biomass are required: therefore, biomass availability close to the plant should be verified. Biomass cost is increased, if they have to be transported for long distances throughout the Country.

6. Conclusions In this paper, the optimization of a hydro-methane and methanol generation plant was carried out based on a one-year time-dependent analysis, using two different thermo-economic optimization codes, one for the thermodynamic analysis, named WTEMP, and one for the time-dependent thermo-economic optimization, named WEPoMP, both of them developed at TPG. Two different configurations were investigated in WTEMP analysis, determining the optimal thermodynamic and chemical parameters of the process. The results were compared for the two plants by analyzing the amounts of methanol and biomass, keeping constant the hydro-methane produced.

84


Throughout WEPoMP optimization, three different scenarios were investigated. In scenario 1 Hydro-methane is produced in CDE only; in scenario 2 CO2 separated in CDE is transported to Asuncion, where it is mixed with H2 produced by electrolysers to synthesize hydro-methane; in scenario 3 Oxygen co-produced by alkaline pressurized electrolysers in Asuncion is sold to Argentina. Each scenario is investigated in both the configurations previously analyzed in WTEMP. The results allow the following conclusion to be drawn: For the configurations designed with WTEMP, results suggest that increasing gasifier size (configuration A) allows for higher revenues due to methanol sale, however reducing the oxygen amount sent to gasifier (configuration B), allows for lower biomass consumption, therefore the impact of biomass cost on the PBP is reduced as well. According to WEPoMP simulations, the higher revenues are obtained in scenario 1A, where the production plant is centered in Ciudad Del Este: in this case using low cost spilled electricity from Itaipu impacts positively on PBP as well. Results of scenario 3A suggest that incorporating oxygen selling, for example to Argentina, represents a viable option that also allows for a wide variety of gas produced (O2, CH3OH, hydro-methane, etc.), making Paraguay a gas producer in South America. Thought results suggest that scenario 1A is the most beneficial in economic terms, about 35,000 ha would be required since high biomass consumption, land and biomass transportation may affect negatively plant costs. Since for every scenario hydro-methane is generated from renewable electricity and biomass, it can be considered a renewable fuel: CO2 emitted in its combustion in natural gas ICEs is the same related to biomass growing, therefore this fuel results totally ‘‘CO2 free’’. In conclusion it is important to consider the general validity of the results of this paper. Even though these simulations correspond to socio-economical aspects of Paraguay scenario, the same approach can be applied to different types of hydroelectric facilities in the world, also considering that usually biomass is largely available close to these kinds of renewable plants. References [1] Rivarolo M, Bogarin J, Magistri L, Massardo AF. Time-dependent optimization of a large hydrogen generation plant using ‘‘spilled’’ water at Itaipu 14 GW hydraulic plant. Int J Hydrogen Energy 2012;37:5434–43. [2] Nejat Vezirog˘lu T, Sümer S ß ahi_n⁄⁄. 21st Century’s energy: hydrogen energy system. Energy Conv Manage 2008;49:1820–31.

[3] Dincer I, Ratlamwala TAH. Development of novel renewable energy based hydrogen production systems: a comparative study. Int J Hydrogen Energy 2013;72:77–87. [4] Ogden JM. Developing an infrastructure for hydrogen vehicles: a Southern California case study. Int J Hydrogen Energy 1999;24:709–30. [5] Hoekman SK et al. CO2 recycling by reaction with renewably-generated hydrogen. Int J Greenhouse Gas Control 2010;4(2010):44–50. [6] Brooks KP, Hu J, Zhu H, Kee RJ. Methanation of carbon dioxide by hydrogen reduction using the Sabatier process in microchannel reactors. Chem Eng Sci 2007;62:1161–70. [7] Rivarolo M, Marmi S, Riveros-Godoy G, Magistri L. Development and assessment of a distribution network of hydro-methane, methanol, oxygen and carbon dioxide in Paraguay. Energy Conv Manage 2014;77:680–9. [8] Clusen LR, Houbak N, Elmegaard B. Technoeconomic analysis of a methanol plant based on gasification of biomass and electrolysis of water. Energy 2010;35:2338–47. [9] Huisman GH, Van Rens GLMA, De Lathouder H, Cornelissen RL. Cost estimation of biomass-to-fuel plants producing methanol, dimethylether or hydrogen. Biomass Bioenergy 2011;35:S155–66. [10] Confederação Nacional da Indústria, 2010, A indústria e o Brasil – gás natural: uma proposta de política para o Brasil/Confederação Nacional da Indústria. – Brasília, 2010. 89 p. [11] Agazzani A, Massardo AF. A tool for thermoeconomic analysis and optimization of gas, steam, and combined plants. J Eng Gas Turbines PowerTrans ASME 1997;120:557–65. http://dx.doi.org/10.1115/1.2818183. [12] Agazzani A, Massardo AF, Frangopoulos CA. Environmental influence on the thermoeconomic optimization of a combined plant with NOx abatement. ASME Trans J Eng Gas Turbine Power 1998;120:557–65. [13] Galanti L, Franzoni A, Traverso A, Massardo AF. Existing large steam power plant upgraded for hydrogen production. Appl Energy 2011;88:1510–8. [14] Galanti L, Massardo AF. Micro gas turbine thermodynamic and economic analysis up to 500 kWe size. Appl Energy 2011;88:4795–802. [15] Jensen JO, Bandur V, Bjerrum NJ, Jensen SO, Ebbesen S, Mogensen M, Tophøj N, Yde L. Pre investigation of water electrolysis, PSO-F&U 2006-1-6287; 2006. [16] Rivarolo M, Greco A, Massardo AF. The thermo-economic optimization of the impact of renewable generators on poly-generation smart grids including hot thermal storage. Energy Conv Manage 2013;65:75–83. [17] Torti C. Development of time depending calculation tools for plant optimization, PhD thesis, University of Genoa, Italy; 2009. [18] Hamenlinck CN, Faaij APC. Future prospects for production of methanol and hydrogen from biomass, Report NWS-E-2001-49, Utrecht University, Copernicus Institute, Department of Science Technology and Society; 2001. [19] Jin H, Larson ED, Celik FE. Performance and cost analysis of future, commercially mature gasification-based electric power generation from switchgrass. Biofuels Bioprod Bioref 2009;3:142–73. [20] [accessed 06.30.13]. [21] Vice-Ministerio de Minas y Energía (VMME). Balance Energético Nacional del Paraguay; 2012 . [22] Klokka Ø, Schreinera PF, Pagès-Bernaus A, Tomasgard A. Optimizing a CO2 value chain for the Norwegian continental shelf. Energy Pol 2010;38:6604–14. [23] Gao L et al. Cost analysis of CO2 transportation: case study in China. Energy Procedia 2011;4:5974–81. [24] [accessed 06.30.13]. [25] Private communication from Parque Tecnologico de Itaipu (PTI). [26] Pliego de Tarifas No. 20 – ANDE Paraguay [accessed 06.30.13]. [27] [accessed 06.30.13].