A New Optimal Selection Method with Seasonal Flow and ... - MDPI

7 downloads 0 Views 3MB Size Report
Nov 20, 2018 - Abstract: A micro hydropower plant of the run-of-river type is considered ... variability, and fitting the best probability distribution function (PDF).
energies Article

A New Optimal Selection Method with Seasonal Flow and Irrigation Variability for Hydro Turbine Type and Size Kiattisak Sakulphan *

and Erik L. J. Bohez

Industrial and Manufacturing Engineering Department, School of Engineering of Technology, Asian Institute of Technology, Pathumthani 12120, Thailand; [email protected] * Correspondence: [email protected] Received: 19 October 2018; Accepted: 16 November 2018; Published: 20 November 2018

 

Abstract: A micro hydropower plant of the run-of-river type is considered to be the most cost-effective investment in developing counties. This paper presents a novel methodology to improve flow estimation, without using the flow direction curve (FDC) method, to determine the turbine type and size to operate consistently. A higher precision is obtained through the use of seasonal flow occurrence data, irrigation variability, and fitting the best probability distribution function (PDF) using flow data. Flow data are grouped in classes based on the flow rate range. This method will need a larger dataset but it is reduced to a tractable amount by using the PDF. In the first part of the algorithm, the average flow of each range is used to select the turbine type. The second part of the algorithm determines the optimal size of the turbine type in a more accurate way, based on minimum and maximum flow rates in each class range instead of the average flow rate. A newly developed micro hydropower plant was installed and used for validation at Baan Khun Pae, Chiang Mai Province. It was found, over four years of observation from 2014–2018, that the plant capacity factor was 82%. Keywords: run-of-river; turbine type and size; streamflow distribution; flow duration curve; hydro turbine; potential energy estimation; irrigation demand; energy efficiency

1. Introduction Global installed hydropower capacity has been growing in recent years by an average of 24.2 GW per year. The International Energy Agency has predicted the long-term deployment of hydropower. Installed hydropower capacity will increase to nearly twice the current level of 1947 GW by 2050 [1]. Hydropower plants can be classified based on the installed power generation capacity [2]. When the installed capacity of a hydropower plant is below 10 MW, it is called a small hydropower plant; when the capacity is up to 100 kW, it is called a micro hydropower plant; and when the capacity is below 10 kW, it is called a mini or pico hydropower plant. Micro and pico hydropower plants are usually in the form of run-of-river schemes [3]. Run-of-river hydropower plants are useful in remote areas where there is a need for energy for human development. At the same time, these plants can create sustainable energy with minimal impact to the surrounding area and communities. Electrical production of run-of-river hydropower plants depends on the locally available water resources. In general, the purpose of a hydropower system is to convert the potential energy of the river into mechanical energy in a turbine shaft, which is connected with an electric generator. The output of a hydropower plant is given in terms of power (kW) and electricity production (kWh). The power input is a function of water flow and net head. These variables are needed for evaluating the installed capacity and determining the appropriate turbine type and size.

Energies 2018, 11, 3212; doi:10.3390/en11113212

www.mdpi.com/journal/energies

Energies 2018, 11, 3212

2 of 16

Some researchers have presented the advantages of installing several small turbines instead of large ones. Using two or three smaller turbines allows easier control of water usage, although this requires more equipment and carries a higher operational cost [3,4]. Voros et al. [5] presented a short-cut design of small run-of-river hydroelectric plants. The minimum and maximum value of water flow was used in this approach for maximizing economic benefits of investment. Montanri [6] studied the economic compatibility of hydropower plants by regularizing the duration curve and the turbine efficiency. Hosseini et al. [7] determined the optimal installation capacity of existing run-of-river small hydropower plants (SHPs) depending on the amount of annual energy, economy, and reliability. The water flow from the flow duration curve (FDC) was used to calculate the annual energy generated. Anagnostopoulos and Papantonis [8] studied the optimal installation capacity of a run-of-river type SHP to assess the economic benefits of the investment. They optimized two parallel turbines of different types and sizes and found the net present value based on annual energy production. The water flow used in different operating ranges of the turbine is the key to more efficient operation. Santolin [9] presented a techno-economical method for predicting the installed capacity of SHPs. A model was developed on the basis of the flow availability of the site, represented by the FDC. The influence of the design operating conditions based on the FDC was used to determine the turbine size and investment cost. Monterio et al. [10] presented a novel short-term forecasting model of electric power production of existing SHPs. The available water flow was used to determine the power production, economy, and maintenance scheduling of the plants. Bortoni et al. [11] presented a novel methodology for the optimal online operation of hydropower plants provided with a single penstock. This method used the flow rate and head value under several conditions for the optimal distribution of the dispatched power amount. Additionally, Paish [12] presented plans for the future development of run-of-river hydropower plants. Future developments are toward the more accurate and rational sizing of systems to maximize their financial return. Most researchers have focused their attention on the optimal installed capacity and the optimal turbine operation of existing hydropower plants based on the turbine operation flow rate range or the potential energy evaluation from the available water resource, as shown in Table 1. They chose the design flow value from the proportion of time a certain streamflow is exceeded which is approximately between 90 to 95% of streamflow occurrence throughout the year from the FDC. Subsequently, the type and size of commercial hydro turbines were considered to optimize the energy production. The results can yield more than one type of turbine for the same operational condition. In general, designs of commercial hydro turbines are standardized, and are not customized for local water resources that are available for all seasons. These commercial turbines are not always operated consistently at the highest efficiency because the design of the turbines is usually determined by the manufacturers. This seems to guarantee that the turbine will operate nearly continuously. However, in reality the operation time is much less than 70%. This could be due to inaccurate FDCs or wrong type and size selection. Table 1. Research summary of hydropower plant turbine size and type selection. Turbine Type Selection

Author/Reference

Bakis [4]

Voros et al. [5] Montanari [6]

Q

H













Turbine Size Selection

Ns

FDC

PDF

TOR

×

×

×

×

× ×

√ √

× √

EAM Emax Teff





















×

Comments PF

×

Investigate the potential capacity size based on cost estimation

×

Maximizing the economic benefits of the plant investment

×

Method for finding the most economical plant capacity

Energies 2018, 11, 3212

3 of 16

Table 1. Cont. Turbine Type Selection

Author/Reference Hosseini et al. [7] Anagnostopoulos et al. [8]

Santolin et al. [9]

Monteiro et al. [10]

Bortoni et al. [11] Rojanamon et al. [13]

Kosa et al. [14]

K. Sakulphan et al.

Q

H

































Ns

× ×

×

FDC

√ √



TOR

×

×

×

×

×

×

×

×

×

×

×

× √





×

Turbine Size Selection

PDF

EAM Emax Teff

















×

×



× √

×

×

×

×

×









√ √





×

Comments PF



To determine the optimal installation capacity

×

×

Optimal plant capacity based on turbine operation

×

×

Method for the capacity size based on economical parameters

×

×

Novel methodology for forecasting the average power production

×

Novel methodology for the optimal turbine operation

×

×

New method for selecting feasible sites

×

×

To determine the potential capacity size based on flow and head







Novel methodology for the optimal turbine type and size

Notes: Available flow rate (Q), net head (H), specific speed (Ns ), flow duration curve (FDC), probability distribution function (PDF), turbine operation range (TOR), economic analysis method (EAM), maximum annual energy production (Emax ), turbine efficiency (Teff ), plant factor (PF).

According to the Energy Policy and Planning Office, Thailand Ministry of Energy, Thailand has set a target to increase its number of SHPs in 2021 by 20% from the current level. Water resources in many areas of Thailand have been surveyed for the installation of SHPs [13,14]. Additionally, the Provincial Electricity Authority (PEA) has installed run-of-river SHPs in several remote communities in Northern Thailand [15]. There are problems with the hydro turbine currently used by the PEA, as it cannot be operated continuously. The turbine can only work well during the rainy season, which lasts approximately five to six months. In dry seasons, the water flow into the system is quite low and not sufficient to operate the turbine. Average data collected over 25 years for six run-of-river SHPs of the PEA are shown in Table 2, and confirm the problem that the average electrical power production value of each hydropower plant is quite different from the installed capacity. This is because the available streamflow value at each site was overestimated from the FDC in order to achieve a high installation capacity in the wet season. However, there is a problem in the dry season. Additionally, the ratio of average electrical power production to installed capacity for each hydropower plant capacity factor is less than 50%, which confirms the above problems and includes the effect of the economic returns from the project. Table 2. The annual data for hydropower plants under Provincial Electricity Authority (PEA) supervision; Maintenance Production System Division [16]. The Average over 25 Years

Mae Tian

Mae Jai

Mae Ya

Mae Pai

Mae Thoei

Baan Khun Pae

1. 2. 3. 4. 5. 6. 7.

Francis 3.13 965 × 2 4,370,199 0.000975 498.9 25.85

Francis 2.29 875 1,926,524 0.002064 219.9 25.13

Turgo 1.28 1150 4,334,214 0.000688 494.8 43.02

Turgo 4.47 1259 × 2 8,015,167 0.000448 915.0 36.34

Turgo 2.52 2,250 7,993,557 0.000614 912.5 40.56

Turgo 0.212 90 115,523 0.009875 13.2 14.67

Hydro turbine type Total investment cost (MUSD) Installation capacity (kW) Average annual energy production (kWh) Average maintenance cost (USD/kWh) Average electrical power production (kW) Plant capacity factor (%)

This paper presents a new way to obtain a more precise flow rate estimation without using the FDC method to determine the optimal turbine type and size, to allow consistent turbine operation throughout the year. This higher precision is obtained using the seasonal flow occurrence data,

Energies 2018, 11, 3212

4 of 16

irrigation variability, and fitting a probability distribution function (PDF) using the observed flow data. This approach provides a more precise estimation than the FDC method. Additionally, for the turbine size selection, the use of the minimum and maximum flow rate of operation is used to optimize Energies 2018, 11, x FOR PEER REVIEW 4 of 16 the operation time. The final results of this algorithm are optimal turbine type and size based on the maximum annual energy production. This case study uses actual streamflow data from the Baan Khun the Baan Khun Pae River, Chom Thong District, Chiang Mai Province, Thailand, including Pae River, Chom Thong District, Chiang Mai Province, Thailand, including information from the Baan information from the Baan Khun Pae micro hydropower plant. Khun Pae micro hydropower plant.

2. Potential Energy Estimation 2. Potential Energy Estimation 2.1. 2.1. Historical Historical Flow Flow Data Data Historical Historical flow flow data data were were obtained obtained from from the the main main river river gauge gauge station station data data at at the the Royal Royal Irrigation Irrigation Department (RID) in Thailand. Historical streamflow data were used to consider streamflow Department (RID) in Thailand. Historical streamflow data were used to consider streamflow trends trends and and variability. variability. Data Data for for the the closest closest gauge gauge station station to to the the study study site, site, located located at at the the Num Num Mae Mae Cheam Cheam River, available from 1953 to to 2007. 2007. The River, were were available from 1953 The catchment catchment area area of of this this gauge gauge station station covered covered the the Baan Baan Khun Pae (BKP) River; the distance between the gauge station and the BKP River is approximately Khun Pae (BKP) River; the distance between the gauge station and the BKP River is approximately 12 12 km. km. The The tendency tendency of of streamflow streamflow occurrence occurrence for for the the streamflow streamflow data data recorded recorded at at this this gauge gauge station station did not have an exact pattern. The variance of historical streamflow data can be expressed did not have an exact pattern. The variance of historical streamflow data can be expressed by by the the standard standard deviation deviation (SD). (SD). The The SD SD of of streamflow streamflow data data set set in in each each month month are are shown shown in in Figure Figure 1. 1. A A low low SD SD in in March March shows shows that that the the streamflow streamflow data data points points tended tended to to be be very very close close to to the the mean mean or or monthly monthly average value. The variance of streamflow occurrence in this month is small, which is more average value. The variance of streamflow occurrence in this month is small, which is more accurate accurate for for prediction. prediction. On On the the other other hand, hand, the the high high SD SD in in September September shows shows that that the the streamflow streamflow values values are are spread out over a large range. Therefore, the prediction of the streamflow occurrence in this month spread out over a large range. Therefore, the prediction of the streamflow occurrence in this month will will be be less less accurate. accurate. Monthly Average Streamflow (m3/s)

180 160 Mean-STD

140

Mean+STD

120

Mean

100 80 60 40 20 0 Jan

Feb

Mar

Apr

May

Jun Jul Month

Aug

Sep

Oct

Nov

Dec

Figure 1. 1. The The average average and and standard standard deviation deviation of of monthly monthly streamflow streamflow at at gauge gauge station station P14. Figure P14.

In order to predict streamflow occurrence in the study area, the variability of monthly streamflow In order to predict streamflow occurrence in the study area, the variability of monthly for various years was plotted, as shown in Figure 2. The middle line presents the mean of all values in streamflow for various years was plotted, as shown in Figure 2. The middle line presents the mean the streamflow data set, the upper line indicates the mean value plus the standard deviation, and the of all values in the streamflow data set, the upper line indicates the mean value plus the standard bottom line shows the mean value minus standard deviation; the area of both graphs is the area deviation, and the bottom line shows the mean value minus standard deviation; the area of both between the top and bottom line, and is called the probability area. This area can be used to predict graphs is the area between the top and bottom line, and is called the probability area. This area can the streamflow occurrence. Some data points are outside the probability area, however, most of the be used to predict the streamflow occurrence. Some data points are outside the probability area, data points are within it. This means that in every year, the streamflow can be greater or less than however, most of the data points are within it. This means that in every year, the streamflow can be the monthly average streamflow (middle line), but will not exceed the mean value plus or minus the greater or less than the monthly average streamflow (middle line), but will not exceed the mean value standard deviation. plus or minus the standard deviation.

Energies 2018, 11, 3212 Energies 2018, 11, x FOR PEER REVIEW

5 of 16 5 of 16

512 256

Discharge (m3/s)

128 64 32 16 8 4 2 1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2005

2010

Year P14,March

P14,September

Mean+STD

Mean-STD

Mean

Figure 2. The variability of mean streamflow in September and March for various years at gauge Figure 2. The variability of mean streamflow in September and March for various years at gauge station P14. station P14.

2.2. Observed Flow Data 2.2. Observed Flow Data After determining the trend of streamflow occurrence in this area, the actual available streamflow After determining trend oftostreamflow in this area, theand actual in Baan Khun Pae River the is measured estimate theoccurrence potential energy production plantavailable capacity. streamflow in Baan Khun Pae River is measured to estimate the potential energy production and Several techniques for measuring the flow in streams can be separated into two main classes of plant capacity. Several techniques for measuring the flow in streams can be separated into two main techniques: direct and indirect measurement methods [17,18]. Direct measurement methods consist classes of methods, techniques: indirect measurement methods [17,18]. Directdilution measurement of several suchdirect as theand velocity area method, moving boat current method, method, methods consist of severaland methods, such method. as the velocity method, moving boat method, electromagnetic method, ultrasonic Directarea measurement methods arecurrent conventionally dilution ultrasonic method. Direct measurement used formethod, mediumelectromagnetic to large rivers.method, Not all and of these methods are generally applicable: methods each willare be conventionally used for medium to large rivers. Not all of these methods are generally applicable: used in the respective power plant depending on different aspects including compatibility, economy, each will be There used are in two the indirect respective power plant depending onmethod different including and accuracy. measurement methods: the weir and aspects slope area method. compatibility, economy, and accuracy. There are two indirect measurement methods: weir In this study, the weir method is used to calculate water flow of the Baan Khun Pae Riverthe because method and slope area method. it has a small concrete weir to store water. Some studies have recommended this method for measuring In this the weir method is used to calculate flowisofthe themost Baanaccurate Khun Pae River because the flow of study, small rivers [19–21]. Furthermore, the weirwater method and reliable flow itmeasurement has a small method; concretethe weir to store water. Some studies have recommended this method accuracy should be about ±1.7 to ±3% and the flow condition should for not measuring the4flow of[21]. small rivers [19–21]. the is weir method is most accurate be more than m3 /s The principle of Furthermore, the weir method to measure thethe free surface flowand and reliable flowthe measurement method; the accuracy be about ±1.7 to ±3% and the flow condition to observe head over the weir. Frequency of should observation is important because the accuracy and 3/s [21]. The principle of the weir method is to measure the free surface should not be more than 4 m reliability of flow estimates depend on this value. Therefore, in order to check the precision of the flow to observethe thewater headlevel overover the the weir. Frequency of observation is important because flow and measurement, concrete weir should be measured twice a day every the day accuracy and reliability of flow estimates depend on this value. Therefore, in order to check the throughout the year. Results are shown in Figure 3. precision of the flow measurement, the water level over the concrete weir should be measured twice a day every day throughout the year. Results are shown in Figure 3.

Energies 2018, 11, x FOR PEER REVIEW Energies 2018, 11, 3212 Energies 2018, 11, x FOR PEER REVIEW

6 of 16 6 of 16 6 of 16

0.60 0.70

Availiale Streamflow (m3/s)

Availiale Streamflow (m3/s)

0.70

0.50 0.60 0.40 0.50 0.30 0.40 0.20 0.30 0.10 0.20 0.00 0.10 0.00

Date Date

Figure 3. Daily available streamflow of the Baan Khun Pae River, May 2012–April 2013. Figure 3. Daily available streamflow of the Baan Khun Pae River, May 2012–April 2013. Figure 3. Daily available streamflow of the Baan Khun Pae River, May 2012–April 2013.

2.3.2.3. Annual Irrigation Demand Estimation Annual Irrigation Demand Estimation 2.3. Annual Irrigation Demand Estimation 2, including TheThe upstream and downstream irrigation areas areare approximately 56 56 kmkm thethe 2 , including upstream and downstream irrigation areas approximately 2 2 2. 2 community of 1.1 km , the agricultural area of 19 and theapproximately forest conservation areaarea 35.9 kmkm 2downstream 2 are 2. Thearea upstream and irrigation areas 56 km , of including the community area of 1.1 km , the agricultural area of km 19 km and the forest conservation of 35.9 2, thethe 2 and 2 This is the total area used in both wet and dry seasons. The dry season is from December to June community area of 1.1 km agricultural area of 19 km the forest conservation area of 35.9 km This is the total area used in both the wet and dry seasons. The dry season is from December to June. andand the wet season is from July to November. Rice is the main wet-season crop, accounting for 70% Thisthe is the area both the wet andRice dry isseasons. The dry season is from December to June wettotal season is used from in July to November. the main wet-season crop, accounting for 70% of of the area, while perennial and vegetable crops account for 30%. Dry season cropping will consist and the wet season is from July to November. Rice is the main wet-season crop, accounting for 70% the area, while perennial and vegetable crops account for 30%. Dry season cropping will consist ofof45% 45% 20%herbs, herbs, 15%ornamental ornamental and other upland crops, and 25% perennials. The schematic ofvegetables, the area, 20% while perennial and vegetable crops 30%. Dry season cropping will consist of vegetables, 15% and otheraccount uplandfor crops, and 25% perennials. The schematic simplified irrigation demand of the Baan Khun Pae area is shown in Figure 4. The upstream irrigation 45% vegetables, 20% herbs, 15% ornamental and other upland crops, and 25% perennials. The schematic simplified irrigation demand of the Baan Khun Pae area is shown in Figure 4. The upstream irrigation areas useuse thethe inflow water from the mountain for for crop growing andand irrigation TheThe remaining simplified irrigation demand ofthe the Baan Khun Pae area is shown inirrigation Figuredemand. 4.demand. The upstream irrigation areas inflow water from mountain crop growing remaining water flows to the concrete weir of the hydropower plant. The downstream irrigation areas use thethe areas use thetoinflow water from mountain for cropplant. growing irrigation demand. remaining water flows the concrete weirthe of the hydropower Theand downstream irrigationThe areas use water from thethe tail race of the hydropower plant andand the water over thethe concrete weir (bypass) forfor water flows to the concrete weir of the hydropower plant. The downstream irrigation areas use the water from tail race of the hydropower plant the water over concrete weir (bypass) 2 that receives water from cropping and irrigation demand. There is also a cropping area of 0.16 km 2 water from the tail race of the hydropower plant and the water over the concrete weir (bypass) for cropping and irrigation demand. There is also a cropping area of 0.16 km that receives water from the 2 theconcrete concrete weir for vegetable cropping. Thus, the water flow at the concrete weir is divided into croppingweir andfor irrigation demand. There is also cropping 0.16 kmweir thatisreceives from vegetable cropping. Thus, the awater flow area at theofconcrete dividedwater into three three parts—bypass, hydropower plant, and cropping area. The water used by the cropping area the concrete weir for vegetable cropping. Thus, the water flow at the concrete weir is divided into parts—bypass, hydropower plant, and cropping area. The water used by the cropping area throughout throughout the two-year period depends on the climate and streamflow occurrence. The average three parts—bypass, hydropower and cropping area.occurrence. The waterThe used by themonthly cropping area the two-year period depends on theplant, climate and streamflow average water monthly water flow requirement of the cropping area is shown in Figure 5. The observation data of throughout the two-year period depends on the climate and streamflow occurrence. The average flow requirement of the cropping area is shown in Figure 5. The observation data of the streamflow theoccurrence streamflow atweir the concrete weir and the average monthly water flow requirement of monthly water requirement theaverage cropping area iswater shown in Figure 5. The data of atoccurrence theflow concrete andofthe monthly flow requirement ofobservation the cropping area theare cropping area are used to estimate the water flow available for the hydropower plant to produce theused streamflow occurrence at the concrete weir average monthly flow requirement of to estimate the water flow available forand the the hydropower plant towater produce electrical energy. electrical energy. Based on the water flow over the concrete weir (bypass), it is estimated that less the cropping area are used to estimate the water flow available for the hydropower plant to produce Based on the water flow over the concrete weir (bypass), it is estimated that less than one percent than one of the remaining water available to sustain lifeTherefore, initthe Therefore, electrical energy. Based on the waterflow flow over the concrete (bypass), is area. estimated that less of the percent remaining water flow is available toissustain aquatic lifeweir inaquatic the area. the amount of theremaining amount of remaining water at the concrete weir will be used to estimate the maximum annual than one percent of the remaining water flow is available to sustain aquatic life in the area. Therefore, water at the concrete weir will be used to estimate the maximum annual energy production energy production of the hydropower plant. thethe amount of remaining water at the concrete weir will be used to estimate the maximum annual of hydropower plant. energy production of the hydropower plant. INFLOW INFLOW

Upstream Irrigation area Upstream Irrigation area

BAAN KHUN PAE RIVER [Concrete weir] BAAN KHUN PAE RIVER [Concrete weir]

BYPASS BYPASS

Hydropower Plant Hydropower Plant

Cropping area 0.16 km2 Cropping area 0.16 km2

Downstream Irrigation area Downstream Irrigation area

Figure 4. Assumption of water use in the Baan Khun Pae Basin. Figure 4. Assumption of water use in the Baan Khun Pae Basin.

Figure 4. Assumption of water use in the Baan Khun Pae Basin.

Average demand (m3(m /s)3/s) Average demand

Energies2018, 2018,11, 11,3212 x FOR PEER REVIEW Energies Energies 2018, 11, x FOR PEER REVIEW 0.035 0.035 0.030 0.030 0.025 0.025 0.020 0.020 0.015 0.015 0.010 0.010 0.005 0.005 0.000 0.000

of16 16 77of 7 of 16 Dry Season Dry Season

DEC DEC

JAN JAN

FEB FEB

MAR MAR

APR APR

MAY JUN MAY JUN Month Month

Wet Season Wet Season

JUL JUL

AUG AUG

SEP SEP

OCT OCT

NOV NOV

Figure 5. 5. Average Average monthly monthly water water requirement requirementfor forcrop cropproduction. production. Figure Figure 5. Average monthly water requirement for crop production.

2.4. 2.4. Estimation Estimation of of the the Available AvailableStreamflow Streamflow 2.4. Estimation of the Available Streamflow There Thereisisno noavailable availableflow flowrecord recordand andthe themeasured measuredstreamflow streamflowfor foronly onlyone oneyear yearisisnot notaccurate accurate There is no available flow record and the measured streamflow for only one year is not accurate enough evaluatethe theavailable availableenergy. energy. The measurement error missing flow data, enough to evaluate The measurement error andand somesome missing flow data, when enough to evaluate theaffect available energy. The measurement error andThis some missingcan flowbe when when rain occurred, affect accuracy of the energy estimation. Thisproblem problem can bedata, by heavyheavy rain occurred, the the accuracy of the energy estimation. solved by heavy rain occurred, affect the accuracy of the energy estimation. This problem can be solved by using using aa PDF. PDF.Various Variousdistribution distributionfunctions functionshave havebeen beenused usedto toincrease increasethe thereliability reliabilityof ofobservation observation using a PDF. distribution functions have been used to and increase the reliability of observation data, as the distribution, lognormal distribution, distribution. One data, such such asVarious the Weibull Weibull distribution, lognormal distribution, andGamma Gamma distribution. One of of the the data, such as the Weibull distribution, lognormal distribution, and Gamma distribution. One ofand the distribution functions will used to to obtain thethe probability of the observed values. Many textbooks distribution functions willbe be used obtain probability of the observed values. Many textbooks distribution functions be used to obtain of distribution theparameters observedparameters values. Many textbooks researchers have recommended a good ofprobability the distribution by using the and researchers have will recommended aestimate goodthe estimate of the bymaximum using the and researchers have recommended a good estimate of the distribution parameters by using are the likelihood [22–24]. This study This considers different probability distributions which maximummethod likelihood method [22–24]. studythree considers three different probability distributions maximum likelihood method [22–24]. This study considers three different probability distributions widely usedwidely in hydrology analysis. analysis. which are used in distribution hydrology distribution which are widely used in hydrology distribution analysis. The to The three three different different distributions distributions are are fitted fitted to the the observation observation values, values, and and parameters parameters of of The three different distributions are fitted to the observation and parameters of distributions method. “statistic toolbox” function distributions are are estimated estimated using using the themaximum maximumlikelihood likelihood method. The Thevalues, “statistic toolbox” function distributions are estimated the Results maximum method. Theestimated “statistic parameters toolbox” function in used for purpose. shown of in MATLAB MATLABisis used forthis thisusing purpose. Results are arelikelihood shown in in Table Table3.3.The The estimated parameters of the the in MATLAB is usedon forthe this purpose.likelihood Results aremethod shownshow in Table The estimated parameters of the distributions based maximum that Gamma distribution performs distributions based on the maximum likelihood method show that3.the the Gamma distribution performs distributions based on the maximum likelihood show that the Gamma performs better other distributions. Figure 6 shows ashows comparison between the fitteddistribution Gamma distribution betterthan thanthethe other distributions. Figure 6 method a comparison between the fitted Gamma better than the other distributions. Figure 6 shows a comparison between the fitted Gamma and the frequency of the observation from Figure The Figure observed flowobserved rates were grouped into distribution and the frequency of thevalues observation values5.from 5. The flow rates were distribution and frequency of the observation 3 /s. 20 classes into with a the range ofwith 0.05am grouped 20 classes range of 0.05 m3/s.values from Figure 5. The observed flow rates were grouped into 20 classes with a range of 0.05 m3/s.

(m3/s) (m3/s)

Figure 6. Comparison of fitted distributions with the observation flow data grouped into 20 classes. Figure 6. Comparison of fitted distributions with the observation flow data grouped into 20 classes. Figure 6. Comparison of fitted distributions with the observation flow data grouped into 20 classes. Table 3. Estimated parameters for selecting distributions. Table 3. Estimated parameters for selecting distributions. Table 3. Estimated parameters for selecting distributions. Distribution Weibull Lognormal Gamma Distribution Weibull Lognormal Gamma Scale parameters α = 0.1023 µ = − 2.9144 α = 0.8682 Distribution Weibull Lognormal Gamma ScaleShape parameters αβ = 0.1023 −2.9144 α = 0.8682 parameters = 0.9157 σ μ= =1.3641 β = 0.1228 Scale parameters α = 0.1023 μ = −2.9144 α = 0.8682 Max. likelihood estimation β = 459.146 437.749 459.386 Shape parameters 0.9157 σ = 1.3641 β = 0.1228 Shape parameters β = 0.9157 σ = 1.3641 β = 0.1228 Max. likelihood estimation 459.146 437.749 459.386 Max. likelihood estimation 459.146 437.749 459.386

Energies 2018, 11, 3212

8 of 16

Energies 2018, 11, x FOR PEER REVIEW

8 of 16

The scale (mean) and shape (variance) parameters of the Gamma distribution function were used The scale (mean) and shape (variance) parameters of the Gamma distribution function were used to generate the streamflow data setset based flowdata datainin one year. streamflow to generate the streamflow data basedon onthe theobserved observed flow one year. TheThe streamflow data data set forseteach dayday within one year usedtoto estimate potential energy production for each within one year(365 (365samples) samples) isisused estimate the the potential energy production of of thethe BKP micro hydropower comparisonofofobservation observation data in one generated BKP micro hydropowerplant. plant. The The comparison data in one yearyear andand generated streamflow datadata set from thethe Gamma distribution isshown shownininTable Table 4. The probability streamflow set from Gamma distributionfunction function is 4. The probability of theof the generated streamflow data is representedby bythe the histograms histograms shown Figure 7. Note thatthat therethere is is generated streamflow data setset is represented showninin Figure 7. Note less variability of the compared withFigure Figure6. 6. less variability of the datadata compared with Table 4. Comparison of observation data and streamflow data form the Gamma distribution function.

Table 4. Comparison of observation data and streamflow data form the Gamma distribution function. Observation Data in

Description Description Observation Data in One Year One Year number of samples Total numberTotal of samples 3 /s) Average streamflow Average(m streamflow (m3/s) 3 /s) Minimum streamflow (m Minimum streamflow (m3/s) 3 /s) Maximum streamflow Maximum(m streamflow (m3/s) Standard deviation Standard deviation

365365 0.1067 0.1067 0.0048 0.0048 0.6024 0.6024 0.1132 0.1132

Gamma Distribution Gamma Distribution Function Function 365 0.1041 0.0001 0.6314 0.1117

365 0.1041 0.0001 0.6314 0.1117

Probability of occurrence

0.275 0.250 0.225 0.200 0.175 0.150 0.125 0.100 0.075 0.050 0.025 0.000

Flow rate range (m3/s)

Figure 7. Probability ofofoccurrence samplesinin2020flow flow rate ranges. Figure 7. Probability occurrenceof of 365 365 samples rate ranges.

3. Turbine Type Selection

3. Turbine Type Selection

The selection of the turbine depends characteristics.The The main characteristics The selection of the turbine dependsupon uponthe the site site characteristics. main characteristics that that should be considered to select the turbine type are the head and flow available, the desired running should be considered to select the turbine type are the head and flow available, the desired running speedspeed of theofgenerator, and whether the turbine will bewill expected to operate in variable flow conditions. the generator, and whether the turbine be expected to operate in variable flow conditions. can be classified into two mainbased groupson based thepattern flow pattern in the Hydro turbinesHydro can beturbines classified into two main groups the on flow in the turbine turbine and speed. the specific speed. Based on pattern, the flow pattern, hydro turbines are categorized three and the specific Based on the flow hydro turbines are categorized into into three groups: high, medium, and These low head. are into grouped two categories: impulse and reaction high, groups: medium, and low head. are These grouped twointo categories: impulse and reaction turbines. turbines. The difference between impulse and reaction turbines can be explained by water pressure. The difference between impulse and reaction turbines can be explained by water pressure. When the When the water pressure applies a force on the face of the runner blades, which decreases as it water pressure applies a force on the face of the runner blades, which decreases as it proceeds through proceeds through the turbines, these are called reaction turbines. A reaction turbine needs a higher the turbines, these are called reaction turbines. A reaction turbine needs a higher volume flow rate to volume flow rate to operate than an impulse turbine. There are three main types of reaction turbine operate than impulseKaplan, turbine. There are turbines. three main of reaction used:pressure the propeller, used: thean propeller, and Francis For types impulse turbines, turbine all the water is Kaplan, and Francis turbines. impulse turbines, all the water pressure converted into kinetic converted into kinetic energyFor before entering the runner. Kinetic energy is in theisform of a high-speed energy entering the runner. Kinetic energy is inofthe of aThere high-speed thattypes strikes jet before that strikes the buckets, mounted on the periphery theform runner. are threejet main of the impulse turbines the Pelton,of Turgo, and cross-flow turbines. variables related toturbines the buckets, mounted onused: the periphery the runner. There are threeAll main types are of impulse specific speed. The specific speed of turbine types is theare most suitable parameter which to used: turbine the Pelton, Turgo, and cross-flow turbines. All variables related to the turbineonspecific speed. base the selection of a turbine. In general, low turbine specific speed of an impulse turbine of a The specific speed of turbine types is the most suitable parameter on which to base the selection corresponds to low flow rates and high heads, whereas high turbine specific speed of a reaction turbine. In general, low turbine specific speed of an impulse turbine corresponds to low flow rates and turbine corresponds to high flow rates and low heads. The turbine specific speed is given by the high heads, whereas high turbine specific speed of a reaction turbine corresponds to high flow rates dimensionless parameter [25]: and low heads. The turbine specific speed is given by the dimensionless parameter [25]: NS =

√ NR Q D ( gHN )3/4

(1)

Energies 2018, 11, 3212

9 of 16

where NS is the turbine specific speed, NR is the angular velocity (rad/s), QD is a design flow (m3 /s), and HN is an effective head (m). Commonly, the micro hydro turbine should be operated at the normal speed of a standard generator. Therefore, the specific speed equation is used in this algorithm for selecting turbine type by fixing the turbine shaft speed. The normal operating ranges for the turbine specific speed and the effective head for each turbine type are shown in Table 5. Table 5. Operating ranges of hydraulic turbines; Dixon and Hall [19]. Turbine Type

Turgo

Pelton

Francis

Kaplan

Specific speed Effective Head (m)

0.02–0.8 50–250

0.05–0.4 50–1300

0.4–2.2 10–350

1.8–5.0 2–40

4. Annual Energy Production Estimation The energy potential of a run-of-river power plant is proportional to the flow and the head. Here, the operation range is used to estimate the annual energy production. The operation range of each turbine type is different. The turbine can operate when the water flow rate is between a minimum (Qmin ) and a maximum (Qmax ) value. Penche [3] presented the minimum technical flow for different types of turbines. The minimum flow for each turbine type is given as a percentage of design flow as shown in Table 6. The design flow (QD ) or water flow rate through the turbine is determined by the following relationship as a function of the available streamflow (Qav ):    0 QD = Q av   Q max

Q av < Qmin Qmin < Q av < Qmax Qmax < Q av

(2)

If the available streamflow is less than the minimum flow rate of the turbine, the turbine will shut down. On the other hand, if the available flow rate is more than the maximum flow rate of the turbine, the maximum flow rate or design flow will be used to operate the turbine because the intake gate and the difference of water level at the small weir of run-of-river will not affect the turbine performance. Therefore, the decision to select the design flow of each turbine type and size will be limited by Equation (2). In order to increase the reliability of micro hydro turbine operation and increase the accuracy of the potential energy estimation, the probability of streamflow occurrence is used to evaluate the maximum annual energy production, given by nc

E=

∑ [ρgQD,i HN Pri h]

(3)

i =1

where E is the maximum annual energy production (kWh/year), QD is the design flow (m3 /s), HN is the effective head or net head (m), Pri is the probability of streamflow occurrence within the range, h is the number of operation hours within a year (h/year), i is the sequence of streamflow range, and n is the total number of streamflow ranges. Table 6. Minimum operating flow of turbines; Penche [3]. Turbine Type

Qmin (% of QD )

Francis Kaplan Cross-flow Pelton Turgo Propeller

30 15 15 10 20 65

Energies 2018, 11, 3212

10 of 16

5. Optimal Turbine Type and Size Selection Algorithm The selection algorithm is separated into two main processes. The first process is selecting the turbine type and minimum flow variable for the operation of each hydro turbine type, as shown in Figure 8. The results of the first process will be used for the second process, which is to find the optimal turbine size after type selection, as shown in Figure 9. The first two steps of the first process are the collection and analysis of observation flow data, including the irrigation demand. The Gamma distribution function is used to generate the streamflow data set for one year (365 samples). Thence, the streamflow data set are arranged from larger values to smaller values. The next steps in the flowchart are finding the maximum streamflow (Qmax ) and minimum streamflow (Qmin ) from the generated streamflow data (Qav ) (365 samples). The class range index is used to select the streamflow range (SR) to calculate the number of class ranges (nc). In each class range (i) the mean streamflow (Qmean ) is used to count the streamflow occurrence number of days to evaluate the probability of occurrence (Pr,i ). Then, each streamflow range is associated with this probability to evaluate the annual energy production and turbine specific speed. The first iteration starts from the minimum flow rate value (Qmin ), and it has a high probability of occurrence. This iteration process continues until the last class range (nc). After that, the turbine specific speed at maximum annual energy production (Equation (3)) is used to select the turbine type and minimum flow rate to operate. If the turbine specific speed value is lower than 0.02 the algorithm is stopped because the lowest streamflow and net head are not suitable; otherwise, it will recommend the best type. The type results in the first process are then used in the second process to select the size (Figure 9). The second process finds the turbine size based on a more accurate way of calculating maximum potential energy production. The streamflow data (Qav ), mean flow rate (Qmean ), probability of streamflow occurrence (Pr,i ), and number of class ranges (nc) are used. Especially, the minimum flow rate variable of the turbine type is used to check the operation condition of the turbine. The first iteration begins from the mean streamflow (Qmean ) in the first-class range. The streamflow data between mean streamflow (Qmean ) and turbine minimum streamflow (Qturbine min flow ) are used to check the operation condition before computing the annual energy production. If the available streamflow is lower than the minimum streamflow, the turbine will be shut down and the energy is equal to zero.

Energies 2018, 11, 3212 Energies 2018, 11, x FOR PEER REVIEW

11 of 16 11 of 16

Start

Collection the available streamflow data (Qav), twice a day and everyday throughout the year Fitting the observation streamflow data by Gamma distribution function and generate the streamflow data set (Qav = Q av,k ); k= 365 Find the minimum streamflow (Qmin) and maximum streamflow (Qmax) from the generated streamflow data (Qav) Identify of streamflow data in each range

Calculate the streamflow range (SR) SR = Q min × Class range index Calculate the number of class ranges (nc) nc = [Q min+Qmax]/SR Calculate the mean flow rate (Qmaen) in each range Count the number of days of the flow rate occurs in each range (Qmean)

Evaluate the annual energy production

Calculate the occurrence probability (Pr,i) of flow rate for each range Calculate the annual energy production (E) and turbine specific speed (Ns) for each range Find the turbine specific speed at maximum annual energy production

Ns < 0.02

Yes

Hydro turbine cannot operate under the flow rate and net head of this case

Yes

Turgo turbine minimum flow rate variable (Qturbine min flow = 0.2 × Qmean)

Yes

Pelton turbine minimum flow rate variable (Qturbine min flow = 0.1 × Qmean)

Yes

Fancis turbine minimum flow rate variable (Qturbine min flow = 0.3 × Qmean)

Yes

Kaplan turbine minimum flow rate variable (Qturbine min flow = 0.15 × Qmean)

Stop

No

Ns ≤ 0.3 No

Ns ≤ 0.4 No

Ns ≤ 2.5 No

Ns > 2.5

Figure 8. Optimal turbine type selection algorithm.

Next process for Turbine size selection

Energies 2018, 11, 3212

12 of 16

Energies 2018, 11, x FOR PEER REVIEW

12 of 16

Available streamflow data (Qav)

Turbine type selection process

Number of class range (nc)

Probability (Pr) in each range

Turbine minimum flow rate variable for type selected before i = i+1

j = j+1

Mean flow rate (Qmean) in each range

For i = 1 : 1 : nc

Nav = number of available streamflow data (365)

For j = 1 : 1: Nav

Qav j ≤ Qturbine min flow

Yes

Power = 0

Yes

Power = ρgHNQav,j

Yes

Power = ρgHNQmean,i

No

Qav j ≤ Qmean, i No

Qav j ≥ Qmean, i

Calculate the average power capacity (Pavg,j)

Calculate the annual energy production (Ej,i)

Find the maximum annual energy production (Emax)

i = nc

Calculate the payback period

Final results Turbine type and size, design flow, plant capacity, payback period, and maximum annual energy production

Stop

Figure 9. Optimal turbine size selection algorithm. Figure 9. Optimal turbine size selection algorithm.

OnOn thethe other hand, if the available streamflow (Q(Q less than oror equal toto the mean streamflow av )avis other hand, if the available streamflow ) is less than equal the mean streamflow (Q(Q ) of the turbine condition, the available streamflow is used to evaluate the power capacity. If the mean mean) of the turbine condition, the available streamflow is used to evaluate the power capacity. If the available streamflow (Q ) is more than the mean streamflow (Q ) of the turbine condition, the mean available streamflow av (Qav) is more than the mean streamflow mean (Qmean) of the turbine condition, the mean streamflow is used to calculate the power capacity.capacity. This loopThis stopsloop withstops the last sequence of streamflow streamflow is used to calculate the power with the last sequence of data (j = 365). After that, the average power capacity (P ) in this range is used to compute the avg streamflow data (j = 365). After that, the average power capacity (Pavg) in this range is used to annual compute energy production The mean(E). streamflow changed toisthe next class range lastuntil classthe the annual energy(E). production The meanisstreamflow changed to the next until class the range range = nc)range is reached. maximum annual energyannual production is production used to determine size. last (i class (i = nc)The is reached. The maximum energy is usedturbine to determine The last process is to check the cost–benefit ratio of the proposed turbine type and size and calculate turbine size. The last process is to check the cost–benefit ratio of the proposed turbine type and size theand design flow. the Information about the maintenance costs of the hydrocosts turbines of calculate design flow. Information about the maintenance of theunder hydrosupervision turbines under

supervision of PEA, the average annual income from electricity sales of PEA, and the investment

Energies 2018, 11, 3212

13 of 16

Energies 2018, 11, x FOR PEER REVIEW

13 of 16

PEA, the average annual income from electricity sales of PEA, and the investment costs of turbines are costs areThe used foroutputs this step.ofThe outputs ofathis algorithm are asize, turbine type and size, used of forturbines this step. final thisfinal algorithm are turbine type and maximum annual maximum annual energy flow, capacity, andThe payback The design energy production, designproduction, flow, plant design capacity, andplant payback period. designperiod. flow value is then flow value is then used to design the detailed turbine geometry. used to design the detailed turbine geometry.

6. 6. Results Results and and Discussion Discussion

Turbine specific speed

0.500

250,000

0.450

225,000 193,251 kWh

0.400

200,000

0.350

175,000

0.300 0.250

0.2051

Specific Speed

150,000

Annual Energy

125,000

0.200

100,000

0.150

75,000

0.100

50,000

0.050

25,000

0.000

Annual Energy Evaluation (kWh)

Based Based on on the the histogram histogram of of probability probability of of streamflow streamflow occurrence occurrence from from Figure Figure 7, 7, the the streamflow streamflow 3 3 range forfor a total of 20 ranges. Therefore, the first-class range of flowof rate is 0.0001– range isis0.0175 0.0175mm/s/s a total ofclass 20 class ranges. Therefore, the first-class range flow rate is 3/s, which 3 0.0175 m encompasses the flow rate values for 289 days. The mean flow rate of this range, 0.0001–0.0175 m /s, which encompasses the flow rate values for 289 days. The mean flow rate of this 3/s, has 3a high probability of occurrence of approximately 80%. The power capacity, annual 0.0088 range, m 0.0088 m /s, has a high probability of occurrence of approximately 80%. The power capacity, energy production, and turbine specific speedspeed in this range are are 6.736.73 kW, 39,015 annual energy production, and turbine specific in this range kW, 39,015kWh, kWh,and and0.0567, 0.0567, respectively. For the next class range, the flow rate range is greater than that of the first class, respectively. For the next class range, the flow rate range is greater than that of the first class, and and will will be used to calculate next annual energy production untilthe thelast lastclass classrange range with with the lowest be used to calculate thethe next annual energy production until lowest probability of occurrence. The inin each class. A higher value of The class class range rangeindex index(see (seeFigure Figure8)8)does doeschange changethe theflow flowrange range each class. A higher value class range index willwill result in a lower number of classes. The class selected to be as of class range index result in a lower number of classes. Therange classindex rangewas index was selected small as possible so that further increases did not change the resulting maximum energy production to be small as possible so that further increases did not change the resulting maximum energy in the turbine typeturbine selection This was obtained 200 classes forclasses a classfor range index of production in the typealgorithm. selection algorithm. This waswith obtained with 200 a class range 36 through and trial error. The comparison of annualofenergy results and turbine index of 36 trial through and error. The comparison annualproduction energy production results and specific turbine speeds for 200 classes are shown in Figure 10. High flow rates have a lower probability of occurrence specific speeds for 200 classes are shown in Figure 10. High flow rates have a lower probability of and produce annual energy. the other flow have have a high probability of occurrence andlow produce low annualOn energy. On thehand, other low hand, lowrates flow rates a high probability occurrence, and allow thethe turbines toto produce of occurrence, and allow turbines producelarger largeramounts amountsofofannual annualenergy. energy.The Theimpulse impulse turbine turbine is is more popular popular in micro hydropower plant selection selection than than the the reaction reaction turbine turbine because because it it is is cheaper cheaper and easiertotomaintain. maintain. Moreover, impulse turbines frictional losses than turbines, reaction and easier Moreover, impulse turbines have have lowerlower frictional losses than reaction turbines, which the flow range tothe operate the From turbine. thethe results, thespecific turbine speed specific which affects theaffects flow rate rangerate to operate turbine. theFrom results, turbine is speed is 0.2051 at maximum annual energy production, which is in the range of the Turgo and Pelton 0.2051 at maximum annual energy production, which is in the range of the Turgo and Pelton turbines. turbines. Bothare turbines areinsuitable inbecause this caseofbecause therate lowand flowhigh rateeffective and highhead. effective head. Both turbines suitable this case the lowof flow The Turgo The Turgo turbine will have a smaller diameter than the Pelton turbine. Therefore, it can be concluded turbine will have a smaller diameter than the Pelton turbine. Therefore, it can be concluded that the that theturbine Turgo is turbine is choice a betterfor choice for this case study theturbine. Pelton turbine. Turgo a better this case study than the than Pelton

0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

0.65

Streamflow occurrence (m3/s)

Figure 10. Results of annual energy evaluation and turbine type selection from the first phase of the Figure 10. Results of annual energy evaluation and turbine type selection from the first phase of the algorithm (Figure 9) in 200 classes. algorithm (Figure 9) in 200 classes.

As shown in Figure 8, the optimal turbine type selection algorithm uses the mean flow rate in shown in Figurethe 8, estimation the optimalofturbine selection algorithm usesproduction. the mean flow rate in eachAs range to compute powertype capacity and annual energy As shown each range to compute the estimation of power capacity and annual energy production. As shown in in Figure 10, the optimal turbine size selection algorithm will use the Qmin cut of values to calculate cut of production values to calculate Figure 10, the optimal turbine size selection algorithm will use annual the Qminenergy the power of the turbine operation. Considering the maximum (Figure the 11), power of the turbine operation. Considering the maximum annual energy production (Figure 11), the the probability of streamflow occurrence is approximately 30%, which can produce a power capacity probability of 87.95 kW.of streamflow occurrence is approximately 30%, which can produce a power capacity of 87.95 kW.

1.000 1.000 0.900 0.900 0.800 0.800 0.700 0.700 0.600 0.600 0.500 0.500 0.400 0.400 0.300 0.300 0.200 0.200 0.100 0.100 0.000 0.0000.00 0.00

14 of of 16 16 14 14 of 16

0.3041 0.3041

87.95 kW 87.95 kW 0.05 0.05

0.10 0.10

0.15 0.15

0.20 0.20

500 500 450 450 400 400 350 350 300 300 250 250 Power Capacity 200 Power Capacity 200 Probability of Occurrence 150 Probability of Occurrence 150 100 100 50 50 0 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0 0.25 0.30 occurrence 0.35 0.40 0.45 0.50 0.55 0.60 0.65 Streamflow (m3/s) Streamflow occurrence (m3/s)

Power Capacity Evaluation (kW) Power Capacity Evaluation (kW)

Probability Occurrence Probability of of Occurrence

Energies 2018, 2018, 11, 11, x3212 Energies FOR PEER REVIEW Energies 2018, 11, x FOR PEER REVIEW

Figure 11. Relationship between probability of occurrence and power capacity evaluation based on Figure and power capacity evaluation based on the Figure 11. 11. Relationship Relationshipbetween betweenprobability probabilityofofoccurrence occurrence and power capacity evaluation based on the first phase of the algorithm (Figure 9)200 in 200 classes. first phase of the algorithm (Figure 9) in classes. the first phase of the algorithm (Figure 9) in 200 classes.

200,000 200,000 180,000 180,000 160,000 160,000 140,000 140,000 120,000 120,000 100,000 100,000 80,000 80,000 60,000 Power Capacity 60,000 Power Capacity 40,000 Annual Energy Production 40,000 Annual Energy Production 20,000 20,000 0 0.070 0.080 0.090 0.100 0 0.070 0.080 0.090 0.100

177,100 kWh 177,100 kWh

Power Capacity (kW) Power Capacity (kW)

100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 00.000 0.000

50.530 kW 50.530 kW

0.010 0.010

0.020 0.020

0.030 0.030

0.040 0.050 0.060 0.040 Design0.050 flow (m30.060 /s) Design flow (m3/s)

Annual Energy Production (kWh) Annual Energy Production (kWh)

In order to increase the accuracy, the operation range of the Turgo turbine, turbine, as calculated calculated using In In order order to to increase increase the the accuracy, accuracy, the the operation operation range range of of the the Turgo Turgo turbine, as as calculated using using Equation (2),was wasused used to estimate the annual energy production. The minimum rate for Equation to estimate the annual energy production. The minimum flow rateflow for operating Equation (2), (2), was used to estimate the annual energy production. The minimum flow rate for operating hydro turbine (Table 6) and information the maintenance costs of the Turgo the hydro the turbine (Table 6) and information regardingregarding the maintenance costs of the Turgo turbines operating the hydro turbine (Table 6) and information regarding the maintenance costs of the Turgo turbines under supervision of the PEA are used in the optimal turbine size selection algorithm, as under supervision of the PEA are PEA usedare in the optimal turbine size selection algorithm, as shown turbines under supervision of the used in the optimal turbine size selection algorithm, as shown in Figure The average annual income electricity the PEA is around in Figure The 9. average annual income from from electricity sales sales of theof is around 0.1050.105 USDUSD per shown in 9. Figure 9. The average annual income from electricity sales ofPEA the PEA is around 0.105 USD per kW. investment determined in this study not include work, mechanical kW. The The investment cost cost determined in this doesdoes not include civilcivil work, onlyonly mechanical and per kW. The investment cost determined in study this study does not include civil work, only mechanical and electrical works of the Turgo turbine type, the cost of which is approximately 1250 USD per kW. electrical works of the Turgo turbine type, the cost of which is approximately 1250 USD per kW. and electrical works of the Turgo turbine type, the cost of which is approximately 1250 USD per kW. The conditionofofturbine turbine operation based on turbine minimum flow rate important is very important to The condition operation based on turbine minimum flow rate is very to accurately The condition of turbine operation based on turbine minimum flow rate is very important to accurately estimate the precise turbine size, since if the available flow rate is lower than the minimum estimate the precise the turbine size, since if thesince available rate isflow lower the minimum flow rate accurately estimate precise turbine size, if theflow available ratethan is lower than the minimum flow rate value thewill turbine willdown. be shutIndown. In contrast, if the available flow rate is than higher than the value the turbine be shut contrast, if the available flow rate is higher the design flow rate value the turbine will be shut down. In contrast, if the available flow rate is higher than the design flow value, the turbine will use the design flow rate value to produce electricity. The accuracy flow value, the turbine will use the design flow rate value to produce electricity. The accuracy of this design flow value, the turbine will use the design flow rate value to produce electricity. The accuracy of this algorithm will depend on the amount of available streamflow is amount a large amount algorithm will depend on the amount of available streamflow data. If data. there If isthere a large of data, of this algorithm will depend on the amount of available streamflow data. If there is a large amount of data, the accuracy of the result will be higher. Furthermore, energy production depends on the the accuracy of the result will be higher. Furthermore, energy production depends on the operational of data, the accuracy of the result will be higher. Furthermore, energy production depends on the operational turbine performance. The final results show that the Turgo turbine can produce a turbine performance. The final results the Turgo produce a maximum annual operational turbine performance. Theshow finalthat results show turbine that thecan Turgo turbine can produce a maximum annual energy of 0.1771 GWh with a payback period of 1197 days. The minimum flow rate energy of 0.1771 GWh withofa0.1771 payback period ofa1197 days.period The minimum flowThe rateminimum required toflow operate maximum annual energy GWh with payback of 1197 days. rate 3/s, with a design3flow of 0.0526 m3/s at a net head of 98 m and a turbine 3operate required 0.0105 mflow is 0.0105 to m /s, withis design of 0.0526 m /s at of a net head and a turbine of required to operate isa0.0105 m3/s, with a design flow 0.0526 m3of /s 98 at amnet head of 98 mshaft and aspeed turbine shaft speed of 1000 rpm. The plant capacity is 50 kW. It is possible to compare the more accurate 1000 The capacity 50 kW. It is possible to compare the more accurate the maximum energy shaftrpm. speed of plant 1000 rpm. Theisplant capacity is 50 kW. It is possible to compare more accurate maximum energy estimate by comparing Figures 10results and 11from withthe thesecond results phase from the second phase estimate by comparing Figures 10 and 11 with the of the algorithm maximum energy estimate by comparing Figures 10 and 11 with the results from the second phase of the algorithm (Figure 12). (Figure 12). of the algorithm (Figure 12).

Figure 12. 12. Comparison Comparison between between the the power power capacity capacity and and annual annual energy energy production production of of the the result result of of the the Figure Figure 12. Comparison between the power capacity and annual energy production of the result of the second phase of the algorithm (Figure 10) in 200 classes. second phase of the algorithm (Figure 10) in 200 classes. second phase of the algorithm (Figure 10) in 200 classes.

Energies 2018, 11, 3212

15 of 16

7. Conclusions A novel methodology for estimating the energy production of a hydro plant based on observation, irrigation demand, and historical flow data was developed and implemented. The results were used to determine optimal turbine type and size. Streamflow was measured and fitted with the best PDF. The flow data were grouped into classes based on flow rate range. Flow occurrence curves were used to compute the total annual energy production for various turbine parameters. An algorithm was developed to find the optimal turbine type and size based on the maximum annual energy production. In the first part of the algorithm the average flow of each range was used. The average flow of the class that gives the highest annual energy was used to select the turbine type. The number of classes used is important: the higher the number of classes, the more accurate the results. The class number was selected to be as small as possible so that a further increase did not change the resulting maximum energy production. This was obtained for 200 classes with a class index of 36. The second part of the algorithm starts with the resulting turbine type as an outcome of the first part of the algorithm. Instead of using the class average flow range, the minimum and maximum flow rate of the selected nominal size was used to compute a more accurate estimate of the annual energy production. Additionally, the information of the maintenance cost, investment cost, and average income for electricity sales from the PEA was used to estimate the payback period. This algorithm was validated by a real case study of the Baan Khun Pae micro hydropower plant in Chiang Mai Province, Thailand. The result of the algorithm showed that the Turgo turbine is a better choice for this case study than other turbines; it can produce a maximum annual energy of 0.1771 GWh with a payback period of 3.28 years. The minimum flow rate needed to operate the turbine is 0.0105 m3 /s, with a design flow of 0.0526 m3 /s, at a maximum power capacity of 50 kW. The final turbine design was manufactured and installed to replace the existing turbine, the size of which was selected based on the standard FDC method. The average annual energy production of the new Turgo turbine for four years (2014–2018) was 0.146 GWh at a plant capacity of 82%. Comparing this result to the historical data of the Baan Khun Pae micro hydropower plant, as shown in Table 2, it can be seen that the new Turgo turbine produced electricity similar to the prediction of our algorithm and the plant capacity is better than for the previous turbine. Although the maximum power of the new Turgo turbine is lower than that of the existing turbine, it can produce more annual energy. Therefore, this algorithm is more accurate than others at determining the type and size of hydro turbine to install in the run-of-river micro hydropower plants with highly variable flow conditions. In case the flow rate is deterministic there will be no need to use the PDF but the two-phase type and size selection can still be used to maximize the annual energy production. Author Contributions: K.S. contributed to the conceptualization, methodology, validation, and data collection, and wrote the paper; E.L.J.B. made corrections to the paper and gave some useful recommendations. Funding: This research was funded by the Department of Research Funding and Technology Development, Provincial Electricity Authority (PEA), Thailand. Acknowledgments: The authors would like to thank the Royal Project Foundation at Baan Khun Pae branch for their annual cropping report in this area. The authors also express their gratitude to the Hydrology and Water Management Center for upper northern region, Royal Irrigation Department for the historical flow data.

References 1. 2. 3. 4.

International Energy Agency. Technology Roadmap Hydropower; IEA: Paris, France, 2012. Spanhoff, B. Current status and future prospects of hydropower in Sexony (Germany) compared to trends in Germany the European Union and the World. Renew. Sustain. Energy Rev. 2014, 30, 518–525. [CrossRef] Penche, C. Layman’s Handbook on How to Develop a Small Hydro Site; European Small Hydropower Association (ESHA): Brussels, Belgium, 1998. Bakis, R. Electricity production opportunities from multipurpose dams (case study). Renew. Energy 2007, 32, 1723–1738. [CrossRef]

Energies 2018, 11, 3212

5. 6. 7.

8. 9. 10. 11. 12. 13.

14.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

16 of 16

Voros, N.G.; Kiranoudis, C.T.; Maroulis, Z.B. Short-cut design of small hydroelectric plants. Renew. Energy 2000, 19, 545–563. [CrossRef] Montanari, R. Criteria for economic planning of a low power hydro electrical plant. Renew. Energy 2003, 28, 2129–2145. [CrossRef] Hosseini, S.M.H.; Forouzbakhsh, F.; Rahimpoor, M. Determination of the optimal installation capacity of small hydro-power plants through the use of technical economic and reliability indices. Energy Policy 2005, 33, 1948–1956. [CrossRef] Anagnostopoulos, J.S.; Papantonis, D.E. Optimal sizing of run-of-river small hydro power plant. Energy Convers. Manag. 2007, 48, 2663–2670. [CrossRef] Santolin, A.; Cavazzini, G.; Pavesi, G.; Ardizzon, G.; Rossetti, A. Techno-economical method for the capacity sizing of a small hydropower plant. Energy Convers. Manag. 2011, 52, 2533–2541. [CrossRef] Monteiro, C.; Ramirez-Rosado, I.J.; Alfredo Fernandez-Jimenez, L. Short-term forecasting model for electric power production of small hydropower plants. Renew. Energy 2013, 50, 387–394. [CrossRef] Bortoni, E.; Bastos, G.; Abreu, T.M.; Kawkabani, B. Online optimal power distribution between units of a hydro power plant. Renew. Energy 2015, 75, 30–36. [CrossRef] Paish, O. Micro-hydropower: Status and prospects. Proc. Inst. Mech. Eng. Part A J. Power Energy 2002, 216, 31–40. [CrossRef] Rojanamon, P.; Chaisomphob, T.; Bureekul, T. Application of geographical information system to site selection of small run-of-river hydropower project by considering engineering/economic/environmental criteria and social impact. Renew. Sustain. Energy Rev. 2009, 13, 2336–2348. [CrossRef] Kosa, P.; Kulworawanichpong, T.; Rivoramas, R.; Chinkulkijniwat, A.; Horpibulsuk, S.; Teaumroong, N. The potential micro-hydropower project in Nakhon Ratchasima province, Thailand. Renew. Energy 2011, 36, 1133–1137. [CrossRef] Provincial Electricity Authority. The Feasibility Study for Construction of Small Hydropower Plants; Water Resources Department, Faculty of Engineering, Kasetsart University: Bangkok, Thailand, 2009. Provincial Electricity Authority. Maintenance Production System Division, Maintenance Report of Hydropower Plant; Provincial Electricity Authority: Bangkok, Thailand, 2015. United States Department of Agriculture, Natural Resources Conservation Service (NRCS). Stream Restoration Design Part 654, National Engineering Handbook; USDA: Washington, DC, USA, 2007. Loucks, D.P.; Van Beek, E.; Stedinger, J.R.; Dijkman, J.P.M.; Villars, M.T. Water Resources Systems Planning and Management; UNESCO: Paris, France, 2005; ISBN 92-3-103998-9. Dixon, S.L.; Hall, C.A. Fluid Mechanics and Thermodynamics of Turbo Machinery, 6th ed.; Elsevier Inc.: Amsterdam, The Netherlands, 2010; ISBN 978-1-85617-793-1. Yunus, A.C.; John, M.C. Fluid Mechanics, 1st ed.; McGrew-Hill: New York, NY, USA, 2006; ISBN 0-07-247236-7. White, F.M. Fluid Mechanics, 4th ed.; McGrew-Hill: New York, NY, USA, 1999; ISBN 0-07-069716-7. Axami, Z.; Siti, K.; Ahmad, M.R.; Hamaruzzaman, S. The suitability of statistical distribution in fitting wind speed data. WSEAS Trans. Math. 2008, 12, 718–727. Sabereh, D.; Mohammad, T.A.; Hakimeh, A. Comparison of four distributions for frequency analysis of wind speed. Environ. Nat. Resour. Res. 2012, 2, 96–105. Booker, D.J.; Snelder, T.H. Comparing methods for estimating flow duration curves at ungauged sites. J. Hydrol. 2012, 434–435, 78–94. [CrossRef] Thake, J. The Micro-Hydro Pelton Turbine Manual: Design, Manufacture and Installation for Small-Scale Hydropower; ITDG Publishing: London, UK, 2000; ISBN 978-1-85339-4607. © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).