Optimal Design and Energy Management of stand-alone Wind/PV ...
Optimal Design and Energy Management of stand-alone Wind/PV/Diesel/Battery Using Bacterial Foraging Algorithm R.Bazyar 1, 2, Kh.Valipoor1, M.R.Javadi4, M.Valizade3 and H.Kord4 1
Department of Engineering, ECE Group, Mohaghegh Ardebili University
2
East Azarbayjan Power Generation Management Company, [email protected] 3
Department of Electrical and Computer Engineering, Ttabriz University 4
Department of Engineering, Zanjan University
Abstract Many stand-alone diesel units are Supplied Rural and remote loads all around the world. With increases oil price, energy crisis, reduce fossil fuels in the world and the concerns about global warming, the integration of diesel generators with renewable energy systems have become an attractive energy sources for supplying the load demand. This paper performs an optimal design of integrated system involving WindPV-Diesel-Battery system for supply remote and rural power demand with CO 2 emission evaluation by using BFA (Bacterial Foraging Algorithm). The system components and radiation and wind speed datasets are assumed to be fully deterministic. System costs involve investments, operation and maintenance as well as loss of load costs. Prices are all unfeigned and components are commercially available. Wind and radiation datasets are for northwest region (Ardebil, latitude: 38_170, longitude: 48_150, altitude: 1345 m) of Iran. From simulation results the proposed system is able to minimize the total annual cost of the system under study and reduce CO2 emission generated by diesel generator. Keywords: Renewable energy, Bacterial Foraging Algorithm, Optimization, Diesel generator, Emission
1. Introduction It is conjectured that two billion people in small villages in developing countries currently lack grid based electricity service. In very cases, grid extension is impractical because of dispersed populations, rugged terrain or both. So, small off grid stand alone renewable energy systems with diesel generators represent an important option for narrowing the electricity gap in rural parts of the developing world [1,2]. In many countries, power networks widely expanse for continuous and high power supply. However there still exist many isolated areas, which are not
connected to utility grid. Usually, diesel generators are utilized to supply power to those areas. Since the fuel price has been increasing in the last years, the generation cost has also increased. Moreover, the issue of global warming due to the combustion of fossil fuel could increase the temperature of the earth. Therefore, the application of renewable energy has become popular as an alternative and clean energy. Recently, the alternative energy sources such as wind and solar power are being used to reduce fuel Consumption for power generation. so, the output power generated by the renewable power sources always fluctuates depending on the weather conditions. At present, the capital cost of installing units for renewable energy system is still expensive compared with diesel generators. In order to generate the power from renewable energy continuously, there is need for backup systems such battery storage or diesel generator. Therefore, a paper methodology is required in order build Wind-PV-Diesel-Battery hybrid energy system. The block diagram of a stand-alone hybrid energy system is shown in Fig.1.To simplify the simulation, only the total capacity from diesel generator is to be used for simulation. The diesel generator is connected at AC bus so that it cans directly supplying the load demand without any converter unit.
Wind Turbines
Diesel Generator
AC/DC Converter
Bidirectional Inverter
Battrey bank
PV panels
DC/DC Converter
DC bus
Load AC bus
Figure1. The system configuration
In previous studies, the optimal sizing problem is solved for wind-fuel cell hybrid system [3], and for wind-solar-fuel cell hybrid system [4]. Furthermore the optimal
sizing of wind-solar-battery hybrid system is performed by means of genetic algorithm [5]. The hybrid Wind-PV-Diesel-Battery system has been studied extensively [6,7]. However, these refrence do not calculate the gas emission such as CO2 produced by diesel generators. Ref [7] have been introduced an applications of artificial neural network on the operation control of the PV/Battery/Diesel hybrid power generation system. Some of renewable energy software applications like HOMER have been developed in order to model the renewable energy supply with economical consideration. However, this software applications could not solve the optimal configuration of existing diesel generators system with PV and wind turbines already installed [8]. Some of the previous studies used monthly average weather condition such as wind speed and solar insolation for a year. However, it is not suitable and not accurate to design the optimal system due to the intermittent nature of renewable energy sources [9]. Hence, by configuration the detail simulation time would obtain the accurate result. Some of the researchers used the iteration techniques to determine the optimal number of renewable energy power. This paper proposes a Bacterial Foraging Algorithm based-optimal design of Wind-PV-Diesel-Battery hybrid energy system for stand-alon applicaion. Bacterial Foraging Algorithm (BFA) is utilized to minimize the ojective function in the study system. The casee study is taken from northwest region (Ardebil) of Iran. The objective function is to minimize the total annual cost. 2. DESCRIPTION OF THE HYBRID SYSTEM COMPONENTS On the design point of view, optimization of the size of a hybrid plant is very important, and leads to a good ratio between cost and performances. Before the system sizing, it's necessary to have enough information about each component of the system. Therefore they are presented in the following sections. Wind Generator Output power of wind turbine generator versus wind speed in this study is shown in Fig. 2. A wind turbine generator needs to consider the cut-in wind speed and the cutout wind speed. If wind speed exceeds the cut-in value, the wind turbine generator starts generating. If wind speed exceeds the rated wind speed, then it generates constant output, and if the wind speed exceeds the cut-out value, the wind turbine generator stops running to protect the generator [10,11]. The power of the wind turbine is described in terms of the wind speed by (1): A.
PWG
0 3 V VC PR W VR VC PR
VW VC ,VW VF VC VW VR
(1)
VR VW VF
Where PWG and PR are wind turbine output power and rated power respectively. VW wind speed [m/s]; VC cut-in wind speed [m/s]; VF cut-out wind speed [m/s]; VR rated
or nominal wind speed [m/s]. Measured data at any height can be converted to installation height through exponent law:
Vw Vwmeasure (
hhub hmeasure
)
(2)
Where, α is the exponent law coefficient. This value varies from less than 0.10 for very flat land, water or ice to more than 0.25 for heavily forested landscapes. The one-seventh power law (0.14) is a good reference number for relatively flat surfaces such as open terrain of grasslands away from tall trees or buildings.
Figure 1. Power output characteristic versus wind speed.
PV cells output model The output power of each PV array, with respect to the solar radiation power, can be calculated through Eq. (3): B.
Ppv
G Ppv ,rated MPPT 1000
(3)
Where, G is perpendicular radiation at array’s surface (W/m2), Ppv,rated is rated power of each PV array at G= 1000W/m2, and ηMPPT is the efficiency of PV’s DC/DC converter and Maximum Power Point Tracking System (MPPT). PV systems are usually equipped with MPPT systems to maximize the power output , therefore it is reasonable to believe that the PV array working states stay around the maximum power point [12]. Using these systems, usually leads to about 30% increase in the average amount of the extracted energy from PV arrays and, as a result, it is economically reasonable to incorporate them into hybrid systems. Thus, in current study it is assumed that PV arrays are equipped with 95% efficient MPPT systems which provide a 48 V DC at DC bus side. It should be noted that, temperature effects are neglected here. The battery output model Since the output of PV cells and the turbine is a random behavior the State of Charge (SOC) of battery bank are constantly changing correspondingly in hybrid systems. When the total output power of the turbine and PV cells is greater than the load power, the battery is in the state of charging, and the charged quantity of the battery at the moment of (t) can be expressed through (4) [13]: C.
Pb t Pb (t 1).(1 ) [Pz (t ) Pl (t ) / inv ].bc
(4)
When the total output power of the turbine and PV cells is less than the load power, the battery is in the state of discharging, and the charged quantity of the battery at the moment of t can be expressed through (5):
Pb t Pb (t 1).(1 ) [Pl (t ) / inv Pz (t )]/ bf
(5)
Where: Pb (t), Battery charged quantity at time (t); Pb (t-1), Battery charged quantity at time (t-1); σ, Battery self-discharge rate per hour; Pz(t), The total output power of the turbine and PV cells in the time interval (t-1) -(t) ; Pl(t) ,The total load power in the time interval (t -1) - (t); ηinv, Inverter efficiency; ηbc, Battery charging efficiency; ηbf, Battery discharging efficiency; Diesel Generator In this simulation diesel generator is operated to meet the load demand. Hence, in order to calculate operational cost of the DG, the mathematical model D.
8,760
FC C f
F( t )
(6)
t 1
Where; F(t): Hourly fuel consumption(liter/hour), based on load characteristic of the diesel generator and is calculated as follow
F ( t ) ( 0.246 PDG ( t ) 0.08415 PR ) (7) Where; PR Rated power of diesel generator,(KW) PDG(t) Power generated by diesel generator,(KW)
Cf Fuel cost per liter,(US$/1) From equation above the rated power and the power generation influence the fuel consumption of DG. Therefore, the DG should not be operated under its minimum point. Usually, manufacturer of the DG give the recommendation to set the minimum diesel operation. On the other hand, the maximum efficiency of DG corresponds to the rated power of the DG. The operation of DG has to be operated between the rated powers and specified minimum value [7]. PDGmin PDG ( t ) PDGmax
(8)
3. PROBLEM FORMULATION The aim of this study is to achieve a stand- alone hybrid generation system, which should be appropriately designed in terms of economic, and environmental measures subject to physical and operational constraints and strategies. a) System cost There are many ways to calculate the economic viability of distribution generation and energy efficiency projects. The capital and replacement costs, the operation and maintenance costs must be combined in some manner so that a comparison may be made with the costs of not doing the project. In this project we don’t need fuel cost because of not using fuel. We choose Net Present Cost (NPC) for calculation of system cost. 1) Net Present Cost The Net Present Cost (NPC) of each component is defined as (9) [13]: NPC = N×(Capital cost + Replacement cost × K +Operation maintenance cost ×
(9)
1 ) CRF(ir , R)
Where, N may be number (unit), R is the useful lifetime of the project (here, 20 years). ir is the real interest rate (here, 6%) which is a function of nominal interest rate (irnominal) and annual inflation rate (fr), defined by [15]: ir
irno min al fr 1 fr
(10)
Also, CRF and K are capital recovery factor [12] and single payment present worth [12], respectively, which are defined as follows:
CRF (ir , R)
ir (1 ir ) R (1 ir ) R 1
(11)
yi
1 n Li n 1 (1 ir )
K i (ir , Li , yi )
(12)
Where, L and y are useful lifetime and number of replacements of the component during useful lifetime of the project, respectively. Number of replacements of each component is a simple function of useful lifetimes of the component and the project, it can be calculated by:
R yi 1 if R is dividable to Li Li
(13)
R yi if R is not dividable to Li Li
(14)
2) Objective function The objective function is the sum of all net present costs. NPC = NPCwg + NPCpv + NPCbat + NPCinv +NPCDG
(15)
Constraints 1) Power balance constraint For any period t, the total power supply from the hybrid generation system must supply the total demand PLOAD with a certain reliability criterion. This relation can be represented by : PPV +PWG +PBAT PLOAD (16) 2) The constraints of the number of turbines, PV cells and batteries; N WG , N PV , N BAT 0
(17)
3) The constraints of the capacity of batteries;
Pbmin Pb Pbmax
(18)
Where, Pbmax means the maximum allowable capacity of batteries, which is generally set to rated battery capacity, Pbmin means the minimum allowable battery capacity, which is determined by the maximum depth of discharging DOD , that calculated by (19):
Pbmin = (1-DOD).Pbmax
(19)
4. Bacterial Foraging Algorithm Natural selection tends to eliminate animals with poor foraging strategies and favor the propagation of genes of those animals that have successful foraging strategies since they are more likely to enjoy reproductive success. After many generations, poor foraging strategies are either eliminated or shaped into good ones. The E. coli bacteria that are present in our intestines also undergo a foraging strategy. The control system of these bacteria that dictates how foraging should proceed can be
subdivided into four sections namely Chemotaxis, Swarming, Reproduction and Elimination and Dispersal [16-18]. a) Chemotaxis This process is achieved through swimming and tumbling via Flagella. Depending upon the rotation of Flagella in each bacterium, it decides whether it should move in a predefined direction (swimming) or altogether in different directions (tumbling), in the entire lifetime. To represent a tumble, a unit length random direction, sayψ, is generated; this will be used to define the direction of movement after a tumble in a tumble, the position of the the bacterium is updated as:
i ( j 1, k ,l ) i ( j , k ,l ) c
(20)
b) Swarming E-coli bacterium has a specific sensing, actuation and decision-making mechanism. As each bacterium moves, it releases attractant to signal other bacteria to swarm towards it. Meanwhile, each bacterium releases repellent to warn other bacteria to keep a safe distance from it. BFA simulates this social behavior by representing the combined cell-to-cell attraction and repelling effect as:
s
jcc ( i ( j , k ,l ), ( j , k ,l )) jcct ( i ) t 1
d attract exp( attract ( mi mt )2 ) t 1 m 1 p s d repellant exp( repellant ( mi mt )2 ) t 1 m 1 p
s
(21)
dattract , ωattract , hrepellant , ωrepellant are different coefficients that are to be chosen properly.
C)Reproduction In BFA, a fixed total number of reproduction steps, Nre is given. Only the first half of populations survive in each reproduce step a surviving bacterium splits into two identical ones, which occupy the same position in the environment as the one in previous step. Thus the population of bacteria keeps constant in each chemotactic step. After Nc chemotactic steps, the fitness values for the it bacterium in the chemotaxis loop are accumulated and calculated by: i health
j
Nc 1
j ( j , k ,l ) j 1
i
(22)
start
Y
Initialization of variables
Compute value of Fitness Function for Each Bacteria J(I,j)
Elimination and dispersal loop counter
J(I,j)<J(I,j+1) i=i+1
L>Ned
Yes
Yes
No
No Reproduction loop counter K=k+1
No
Swiming No Ns> Swiming
No K>Nre
Tumble Yes
No X
Chemotactic loop counter J=j+1 No
I>S Yes X
j>Nc No Y Figure 2. Bacterial Foraging Algorithm flowchart
The smaller the ih jealth is, the healthier the bacterium is. To simulate the reproduction character in nature and to accelerate the swarming speed, all the bacteria are sorted according to their health values in an ascending order and each of the first Sr ( Sr = S / 2 ) , for convenience S is assumed to be a positive even integer) bacteria splits into two bacteria. The characters including location and step length of the mother bacterium are reproduced to the children bacteria. Through this selection process the remaining Sr unhealthier bacteria are eliminated and discarded. To simplify the algorithm, the number of the bacteria keeps constant in the whole process. d)Elimination and Dispersal For the purpose of improving the global search ability, elimination-dispersal event is defined after Nre steps of reproduction. The bacteria are eliminated and dispersed to
random positions in the optimization domain according to the probability Ped . This elimination-dispersal event helps the bacterium avoid being trapped into local optima. The number of the event is denoted as Ned. The flowchart of the algorithm is shown in Fig 3.
5.
Results and Discussions An optimum configuration data, load performed in this simulation. Meteorological data, load profile and the other factors were taken from northwest region (Ardebil) of Iran. Table 1 shows data of the diesel generator and tables 2 until 5 show the data of renewable energy units. The installation cost has been included in the capital cost of the devices and maintenance cost of each unit per year. The maintenance cost is expressed as a fraction of the component cost. The available data consist of hourly averages of wind speed and solar radiation in one of the northwestern provinces of Iran, i.e. Ardebil (latitude: 38_170, longitude: 48_150, altitude: 1345 m). The chosen load profile is the IEEE household consumptions with a peak of 1 kW. For the sake of simplicity, we have considered the weekly mean in input data in our simulation. The data are the wind velocity, solar radiation and the demand in every one hour in a day. So, an average of the input data in each hour is calculated during a week. In a year, we have 1248 (52 × 24) data about the wind velocity and demand. These data are shown in Figs. 4–7. The simulation time is done in one-hour time step until 8760 hour in one year. The project lifetime in this study is 20 years. Batteries need to be replaced throughout the project lifetime .The fuel price and O&M for DG shows in table 1. The simulation result is shown in Tables 5 and 6. From simulation results by Bacterial Foraging Algorithm it can be seen that the optimal design of hybrid system consist of one wind turbine, 14 of PV panels , 7 batteries storage, one diesel generator. From Table 7 it can be seen that the annual cost of hybrid Wind-PVDiesel-Battery is equal to 110329.28 US$ .The annual cost of DG only is 113671.08 US$. Also convergence curves of the Bacterial Foraging algorithm , are depicted in Fig. 8. Therefore, from economical of view the hybrid Wind- PV-Diesel-Battery is suitable to be implemented in Ardebil. The proposed configuration also compared with existing system only used DG. Table 7 shows, the operation of Wind-PVDiesel-Battery system could decrease the total annual cost of the system. The operation of diesel only consumes a lot of fuel so that the operation cost is expensive that shows in table 6. Moreover, the oil prices always increase year by year. Hence, it is cost effective to make combination between renewable energy with DG. In addition, the operation of DG will generate CO 2 emission higher than the proposed system.
Figure 3. ourly IEEE household
Figure 4. Solar radiation in the Ardebil city.
consumptions load profile.
Insulation (W/M2)
1000 800 600 400 200 0 0
5
10 15 Time (Hour of Day)
Figure 5. Wind speed in a year in the Ardebil city.
20
25
Figure 6. Solar radiation in 24 hour
6
3
x 10
Total cost (US$)
2.5 2
1.5 1
0.5 0 0
50
100
150
200
250
Number of iteration Figure 7. Convergence of optimization of algorithme. TABLE 1. DIESEL GENERATOR SPECIFICATIONS (WILSON P75P1 )
Power rating (KW)
Fuel consumpt (L/hr)
Fuel price (US$/Liter)
Capital cost (US$)
Maintenance cost (US$/year)
Lifetime (hour)
2
17.5
0.66
5600
5% of capital cost
15000
TABLE 2. WIND GENERATOR SPECIFICATIONS
Power rating (W)
Cut-in wind speed m/s
Rated wind speed m/s
Cut-out wind speed m/s
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
1000
2.5
11
24
3.000
3% of capital cost
20
TABLE 3. PV MODULES SPECIFICATIONS
Power rating (W)
Open circuit voltage (v)
Short circuit current (A)
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
110
21
7.22
4.86
1% of capital cost
20
TABLE 4. BATTERY BANK SPECIFICATIONS
Nominal capacity (A h)
Voltage (v)
230
12
DOD (%)
Battery charging and discharging efficiency
σ (%)
85%
0.02
80
Capital cost (US$/W)
Maintenance cost (US$/year) Lifetime (year)
0.171
1% of capital cost
4
DC/AC INVERTER SPECIFICATIONS
TABLE 5.
Power rating (W)
Efficiency (%)
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
80
0.707
1% of capital cost
10
1500
TABLE 6.
OPTIMAL SIZING RESULTS FOR DIESEL GENERATOR-ONLY POWER SOURCE Diesel generator capacity (KW)
Total cost (US$)
CO2 emission(Ton /year)
2
113671.08
1398
TABLE 7. HYBRID SYSTEM OPTIMAL SIZING RESULTS
Optimization technique
Populatio n size
Iteration number
Time (s)
BFA
30
150
563. 5
Optimum values of parameters NWG NPV NBAT DG 1
14
7
1
Total cost (US$)
110329.28
CO2 emission(To n/year) 754.92
Conclusion
This paper presents an effective approach for optimal design, energy management, modeling and analysis of stand-alone hybrid Wind-PV-Diesel-Battery energy systems. The main purpose for designing hybrid energy systems is reliable supply of the load, under varying weather conditions, with minimum cost. In this paper, a stand-alone hybrid system is designed for a 20-year period of operation. The 20-year round total system cost is equal to the sum of the respective components capital and maintenance costs. The optimal design, energy management and cost function minimization is implemented using Bacterial Foraging algorithm. The simulation results of algorithm verify that the hybrid Wind-PV-Diesel-Battery systems result in lower system cost compared to diesel generators are used. The total cost of the optimized system showed that the system can deliver energy in a stand-alone installation with an acceptable cost. Hence, the hybrid proposed system is better from economical point of view and environmental effect. References [1] CD. Barley, "Optimal dispatch strategy in remote hybrid power systems". Solar Energy; (1996). 58(4–6):165–79. [2] J. Bryne , B. Shen, W. Wallace, "The economics of sustainable energy for rural development" a study of renewable energy in China. Energy Policy; (1998). 26(1):45–54. [3] Z. Samaras, D. Zafeiris , “Optimization of a wind power fuel cell hybrid system in an autonomous electrical network environment,” International Journal of Renewable Energy, (2007). Vol. 32, pp. 57-79. [4] C. Wang, and M.H. Nehrir, “Power Management of a Stand-Alone Wind/Photovoltaic/Fuel Cell Energy System,” IEEE Trans. on Energy Conversion, (2008). Vol. 23, No.3. [5] D. Kolokotsa et al, “Methodology for optimal sizing of stand-alone photovoltaic/windgenerator systems using genetic algorithms,” International Journal of Solar Energy, (2006).Vol. 80(9), pp. 1072-1088. [6] T. Senjyu, D. Hayashi, A. Yona, N. Urasaki, and T. Funabashi: 2007 " optimal configuration or power generating in isolated island with renewable energy", Renewable Energy, Vol.32, pp.1917-1933 [7] R. Dufo and J.L. Bernal, "Design and control strategies of PV-Diesel system using genetic algorithms, Solar Energy,( 2004). Vol. 79, pp33-46 [8] https://analysis.nrel.gov/homer/ [9] B. Nelson, M. H. Nehrir, and C. Wang, " unit sizing and cost analysis of standalone hybrid wind/pv/feul cell power generation systems", Renewable Energy,( 2006). Vol.31, pp.1641-1656 [10] Y. Xingjia, L. Yingming, B. Jieqiu, and Xing Zuoxia, “Research and Simulation of Direct Drive Wind Turbine VSCF Characteristic” Proc. Int Conf. on Automation and Logistics Qingdao, China, September (2008). PP.1683-1687.
[11] M. S. Alam, and D. W. Gao, “Modeling and Analysis of a Wind/PV/Fuel Cell Hybrid Power System in HOMER” Proc. Int. Conf. on Industrial Electronics and Applications, (2007). PP.1594-1599. [12] H. Yang, W. Zhou, Z. Fang, “ Optimal sizing method for stand-alone hybrid solar–wind system with LPSP technology by using genetic algorithm ,” Sol Energy; (2008). 82:354–67. [13] S.M. Hakimi , S.M. Tafreshi , A. K. Kaviani. August, "Unit sizing of a standalone hybrid power system using particle swarm optimization (PSO). In: Proceeding of the international conference on automation and logistics, Jinan, China. (2007). [14] S.M. Hakimi, S.M. Tafreshi, Aug "Unit sizing of a stand-alone hybrid power system using particle swarm optimization (PSO)". (2007). IEEE ICAL; 18– 21:3107–12. [15] A.H. Shahirinia, S.M. Tafreshi, A.H. Gastaj, A.R. Moghaddomjoo, November 2005 “Optimal sizing of hybrid power system using genetic algorithm”, In: International conference on future power systems (2005); 16–18. [16] K. M. Passino, Jun. " Biomimicry of bacterial foraging for distributed optimization and control",. IEEE Control Syst. Mag., (2002). vol. 22, no. 3, pp.52.67. [17] W. Morgantown, "Emission rates for new DG technologies,. The Regulatory Assistance Project, [Online]". Available,http://www.raponline.org/ProjDocs/DREmsRul/Collfi le/DGEmissionspdf. May(2001). [18] H. Noormohamadi, S. Niknafs, A. Gharaveisi , "Automatic Generation Control of a Multi-Area Power System Using Bacterial Foraging Algorithm, International Review of Electrical Engineering (IREE), May (2010). vol. 5 n. 3, pp. 1183 . 1188.
Department of Engineering, ECE Group, Mohaghegh Ardebili University
2
East Azarbayjan Power Generation Management Company, [email protected] 3
Department of Electrical and Computer Engineering, Ttabriz University 4
Department of Engineering, Zanjan University
Abstract Many stand-alone diesel units are Supplied Rural and remote loads all around the world. With increases oil price, energy crisis, reduce fossil fuels in the world and the concerns about global warming, the integration of diesel generators with renewable energy systems have become an attractive energy sources for supplying the load demand. This paper performs an optimal design of integrated system involving WindPV-Diesel-Battery system for supply remote and rural power demand with CO 2 emission evaluation by using BFA (Bacterial Foraging Algorithm). The system components and radiation and wind speed datasets are assumed to be fully deterministic. System costs involve investments, operation and maintenance as well as loss of load costs. Prices are all unfeigned and components are commercially available. Wind and radiation datasets are for northwest region (Ardebil, latitude: 38_170, longitude: 48_150, altitude: 1345 m) of Iran. From simulation results the proposed system is able to minimize the total annual cost of the system under study and reduce CO2 emission generated by diesel generator. Keywords: Renewable energy, Bacterial Foraging Algorithm, Optimization, Diesel generator, Emission
1. Introduction It is conjectured that two billion people in small villages in developing countries currently lack grid based electricity service. In very cases, grid extension is impractical because of dispersed populations, rugged terrain or both. So, small off grid stand alone renewable energy systems with diesel generators represent an important option for narrowing the electricity gap in rural parts of the developing world [1,2]. In many countries, power networks widely expanse for continuous and high power supply. However there still exist many isolated areas, which are not
connected to utility grid. Usually, diesel generators are utilized to supply power to those areas. Since the fuel price has been increasing in the last years, the generation cost has also increased. Moreover, the issue of global warming due to the combustion of fossil fuel could increase the temperature of the earth. Therefore, the application of renewable energy has become popular as an alternative and clean energy. Recently, the alternative energy sources such as wind and solar power are being used to reduce fuel Consumption for power generation. so, the output power generated by the renewable power sources always fluctuates depending on the weather conditions. At present, the capital cost of installing units for renewable energy system is still expensive compared with diesel generators. In order to generate the power from renewable energy continuously, there is need for backup systems such battery storage or diesel generator. Therefore, a paper methodology is required in order build Wind-PV-Diesel-Battery hybrid energy system. The block diagram of a stand-alone hybrid energy system is shown in Fig.1.To simplify the simulation, only the total capacity from diesel generator is to be used for simulation. The diesel generator is connected at AC bus so that it cans directly supplying the load demand without any converter unit.
Wind Turbines
Diesel Generator
AC/DC Converter
Bidirectional Inverter
Battrey bank
PV panels
DC/DC Converter
DC bus
Load AC bus
Figure1. The system configuration
In previous studies, the optimal sizing problem is solved for wind-fuel cell hybrid system [3], and for wind-solar-fuel cell hybrid system [4]. Furthermore the optimal
sizing of wind-solar-battery hybrid system is performed by means of genetic algorithm [5]. The hybrid Wind-PV-Diesel-Battery system has been studied extensively [6,7]. However, these refrence do not calculate the gas emission such as CO2 produced by diesel generators. Ref [7] have been introduced an applications of artificial neural network on the operation control of the PV/Battery/Diesel hybrid power generation system. Some of renewable energy software applications like HOMER have been developed in order to model the renewable energy supply with economical consideration. However, this software applications could not solve the optimal configuration of existing diesel generators system with PV and wind turbines already installed [8]. Some of the previous studies used monthly average weather condition such as wind speed and solar insolation for a year. However, it is not suitable and not accurate to design the optimal system due to the intermittent nature of renewable energy sources [9]. Hence, by configuration the detail simulation time would obtain the accurate result. Some of the researchers used the iteration techniques to determine the optimal number of renewable energy power. This paper proposes a Bacterial Foraging Algorithm based-optimal design of Wind-PV-Diesel-Battery hybrid energy system for stand-alon applicaion. Bacterial Foraging Algorithm (BFA) is utilized to minimize the ojective function in the study system. The casee study is taken from northwest region (Ardebil) of Iran. The objective function is to minimize the total annual cost. 2. DESCRIPTION OF THE HYBRID SYSTEM COMPONENTS On the design point of view, optimization of the size of a hybrid plant is very important, and leads to a good ratio between cost and performances. Before the system sizing, it's necessary to have enough information about each component of the system. Therefore they are presented in the following sections. Wind Generator Output power of wind turbine generator versus wind speed in this study is shown in Fig. 2. A wind turbine generator needs to consider the cut-in wind speed and the cutout wind speed. If wind speed exceeds the cut-in value, the wind turbine generator starts generating. If wind speed exceeds the rated wind speed, then it generates constant output, and if the wind speed exceeds the cut-out value, the wind turbine generator stops running to protect the generator [10,11]. The power of the wind turbine is described in terms of the wind speed by (1): A.
PWG
0 3 V VC PR W VR VC PR
VW VC ,VW VF VC VW VR
(1)
VR VW VF
Where PWG and PR are wind turbine output power and rated power respectively. VW wind speed [m/s]; VC cut-in wind speed [m/s]; VF cut-out wind speed [m/s]; VR rated
or nominal wind speed [m/s]. Measured data at any height can be converted to installation height through exponent law:
Vw Vwmeasure (
hhub hmeasure
)
(2)
Where, α is the exponent law coefficient. This value varies from less than 0.10 for very flat land, water or ice to more than 0.25 for heavily forested landscapes. The one-seventh power law (0.14) is a good reference number for relatively flat surfaces such as open terrain of grasslands away from tall trees or buildings.
Figure 1. Power output characteristic versus wind speed.
PV cells output model The output power of each PV array, with respect to the solar radiation power, can be calculated through Eq. (3): B.
Ppv
G Ppv ,rated MPPT 1000
(3)
Where, G is perpendicular radiation at array’s surface (W/m2), Ppv,rated is rated power of each PV array at G= 1000W/m2, and ηMPPT is the efficiency of PV’s DC/DC converter and Maximum Power Point Tracking System (MPPT). PV systems are usually equipped with MPPT systems to maximize the power output , therefore it is reasonable to believe that the PV array working states stay around the maximum power point [12]. Using these systems, usually leads to about 30% increase in the average amount of the extracted energy from PV arrays and, as a result, it is economically reasonable to incorporate them into hybrid systems. Thus, in current study it is assumed that PV arrays are equipped with 95% efficient MPPT systems which provide a 48 V DC at DC bus side. It should be noted that, temperature effects are neglected here. The battery output model Since the output of PV cells and the turbine is a random behavior the State of Charge (SOC) of battery bank are constantly changing correspondingly in hybrid systems. When the total output power of the turbine and PV cells is greater than the load power, the battery is in the state of charging, and the charged quantity of the battery at the moment of (t) can be expressed through (4) [13]: C.
Pb t Pb (t 1).(1 ) [Pz (t ) Pl (t ) / inv ].bc
(4)
When the total output power of the turbine and PV cells is less than the load power, the battery is in the state of discharging, and the charged quantity of the battery at the moment of t can be expressed through (5):
Pb t Pb (t 1).(1 ) [Pl (t ) / inv Pz (t )]/ bf
(5)
Where: Pb (t), Battery charged quantity at time (t); Pb (t-1), Battery charged quantity at time (t-1); σ, Battery self-discharge rate per hour; Pz(t), The total output power of the turbine and PV cells in the time interval (t-1) -(t) ; Pl(t) ,The total load power in the time interval (t -1) - (t); ηinv, Inverter efficiency; ηbc, Battery charging efficiency; ηbf, Battery discharging efficiency; Diesel Generator In this simulation diesel generator is operated to meet the load demand. Hence, in order to calculate operational cost of the DG, the mathematical model D.
8,760
FC C f
F( t )
(6)
t 1
Where; F(t): Hourly fuel consumption(liter/hour), based on load characteristic of the diesel generator and is calculated as follow
F ( t ) ( 0.246 PDG ( t ) 0.08415 PR ) (7) Where; PR Rated power of diesel generator,(KW) PDG(t) Power generated by diesel generator,(KW)
Cf Fuel cost per liter,(US$/1) From equation above the rated power and the power generation influence the fuel consumption of DG. Therefore, the DG should not be operated under its minimum point. Usually, manufacturer of the DG give the recommendation to set the minimum diesel operation. On the other hand, the maximum efficiency of DG corresponds to the rated power of the DG. The operation of DG has to be operated between the rated powers and specified minimum value [7]. PDGmin PDG ( t ) PDGmax
(8)
3. PROBLEM FORMULATION The aim of this study is to achieve a stand- alone hybrid generation system, which should be appropriately designed in terms of economic, and environmental measures subject to physical and operational constraints and strategies. a) System cost There are many ways to calculate the economic viability of distribution generation and energy efficiency projects. The capital and replacement costs, the operation and maintenance costs must be combined in some manner so that a comparison may be made with the costs of not doing the project. In this project we don’t need fuel cost because of not using fuel. We choose Net Present Cost (NPC) for calculation of system cost. 1) Net Present Cost The Net Present Cost (NPC) of each component is defined as (9) [13]: NPC = N×(Capital cost + Replacement cost × K +Operation maintenance cost ×
(9)
1 ) CRF(ir , R)
Where, N may be number (unit), R is the useful lifetime of the project (here, 20 years). ir is the real interest rate (here, 6%) which is a function of nominal interest rate (irnominal) and annual inflation rate (fr), defined by [15]: ir
irno min al fr 1 fr
(10)
Also, CRF and K are capital recovery factor [12] and single payment present worth [12], respectively, which are defined as follows:
CRF (ir , R)
ir (1 ir ) R (1 ir ) R 1
(11)
yi
1 n Li n 1 (1 ir )
K i (ir , Li , yi )
(12)
Where, L and y are useful lifetime and number of replacements of the component during useful lifetime of the project, respectively. Number of replacements of each component is a simple function of useful lifetimes of the component and the project, it can be calculated by:
R yi 1 if R is dividable to Li Li
(13)
R yi if R is not dividable to Li Li
(14)
2) Objective function The objective function is the sum of all net present costs. NPC = NPCwg + NPCpv + NPCbat + NPCinv +NPCDG
(15)
Constraints 1) Power balance constraint For any period t, the total power supply from the hybrid generation system must supply the total demand PLOAD with a certain reliability criterion. This relation can be represented by : PPV +PWG +PBAT PLOAD (16) 2) The constraints of the number of turbines, PV cells and batteries; N WG , N PV , N BAT 0
(17)
3) The constraints of the capacity of batteries;
Pbmin Pb Pbmax
(18)
Where, Pbmax means the maximum allowable capacity of batteries, which is generally set to rated battery capacity, Pbmin means the minimum allowable battery capacity, which is determined by the maximum depth of discharging DOD , that calculated by (19):
Pbmin = (1-DOD).Pbmax
(19)
4. Bacterial Foraging Algorithm Natural selection tends to eliminate animals with poor foraging strategies and favor the propagation of genes of those animals that have successful foraging strategies since they are more likely to enjoy reproductive success. After many generations, poor foraging strategies are either eliminated or shaped into good ones. The E. coli bacteria that are present in our intestines also undergo a foraging strategy. The control system of these bacteria that dictates how foraging should proceed can be
subdivided into four sections namely Chemotaxis, Swarming, Reproduction and Elimination and Dispersal [16-18]. a) Chemotaxis This process is achieved through swimming and tumbling via Flagella. Depending upon the rotation of Flagella in each bacterium, it decides whether it should move in a predefined direction (swimming) or altogether in different directions (tumbling), in the entire lifetime. To represent a tumble, a unit length random direction, sayψ, is generated; this will be used to define the direction of movement after a tumble in a tumble, the position of the the bacterium is updated as:
i ( j 1, k ,l ) i ( j , k ,l ) c
(20)
b) Swarming E-coli bacterium has a specific sensing, actuation and decision-making mechanism. As each bacterium moves, it releases attractant to signal other bacteria to swarm towards it. Meanwhile, each bacterium releases repellent to warn other bacteria to keep a safe distance from it. BFA simulates this social behavior by representing the combined cell-to-cell attraction and repelling effect as:
s
jcc ( i ( j , k ,l ), ( j , k ,l )) jcct ( i ) t 1
d attract exp( attract ( mi mt )2 ) t 1 m 1 p s d repellant exp( repellant ( mi mt )2 ) t 1 m 1 p
s
(21)
dattract , ωattract , hrepellant , ωrepellant are different coefficients that are to be chosen properly.
C)Reproduction In BFA, a fixed total number of reproduction steps, Nre is given. Only the first half of populations survive in each reproduce step a surviving bacterium splits into two identical ones, which occupy the same position in the environment as the one in previous step. Thus the population of bacteria keeps constant in each chemotactic step. After Nc chemotactic steps, the fitness values for the it bacterium in the chemotaxis loop are accumulated and calculated by: i health
j
Nc 1
j ( j , k ,l ) j 1
i
(22)
start
Y
Initialization of variables
Compute value of Fitness Function for Each Bacteria J(I,j)
Elimination and dispersal loop counter
J(I,j)<J(I,j+1) i=i+1
L>Ned
Yes
Yes
No
No Reproduction loop counter K=k+1
No
Swiming No Ns> Swiming
No K>Nre
Tumble Yes
No X
Chemotactic loop counter J=j+1 No
I>S Yes X
j>Nc No Y Figure 2. Bacterial Foraging Algorithm flowchart
The smaller the ih jealth is, the healthier the bacterium is. To simulate the reproduction character in nature and to accelerate the swarming speed, all the bacteria are sorted according to their health values in an ascending order and each of the first Sr ( Sr = S / 2 ) , for convenience S is assumed to be a positive even integer) bacteria splits into two bacteria. The characters including location and step length of the mother bacterium are reproduced to the children bacteria. Through this selection process the remaining Sr unhealthier bacteria are eliminated and discarded. To simplify the algorithm, the number of the bacteria keeps constant in the whole process. d)Elimination and Dispersal For the purpose of improving the global search ability, elimination-dispersal event is defined after Nre steps of reproduction. The bacteria are eliminated and dispersed to
random positions in the optimization domain according to the probability Ped . This elimination-dispersal event helps the bacterium avoid being trapped into local optima. The number of the event is denoted as Ned. The flowchart of the algorithm is shown in Fig 3.
5.
Results and Discussions An optimum configuration data, load performed in this simulation. Meteorological data, load profile and the other factors were taken from northwest region (Ardebil) of Iran. Table 1 shows data of the diesel generator and tables 2 until 5 show the data of renewable energy units. The installation cost has been included in the capital cost of the devices and maintenance cost of each unit per year. The maintenance cost is expressed as a fraction of the component cost. The available data consist of hourly averages of wind speed and solar radiation in one of the northwestern provinces of Iran, i.e. Ardebil (latitude: 38_170, longitude: 48_150, altitude: 1345 m). The chosen load profile is the IEEE household consumptions with a peak of 1 kW. For the sake of simplicity, we have considered the weekly mean in input data in our simulation. The data are the wind velocity, solar radiation and the demand in every one hour in a day. So, an average of the input data in each hour is calculated during a week. In a year, we have 1248 (52 × 24) data about the wind velocity and demand. These data are shown in Figs. 4–7. The simulation time is done in one-hour time step until 8760 hour in one year. The project lifetime in this study is 20 years. Batteries need to be replaced throughout the project lifetime .The fuel price and O&M for DG shows in table 1. The simulation result is shown in Tables 5 and 6. From simulation results by Bacterial Foraging Algorithm it can be seen that the optimal design of hybrid system consist of one wind turbine, 14 of PV panels , 7 batteries storage, one diesel generator. From Table 7 it can be seen that the annual cost of hybrid Wind-PVDiesel-Battery is equal to 110329.28 US$ .The annual cost of DG only is 113671.08 US$. Also convergence curves of the Bacterial Foraging algorithm , are depicted in Fig. 8. Therefore, from economical of view the hybrid Wind- PV-Diesel-Battery is suitable to be implemented in Ardebil. The proposed configuration also compared with existing system only used DG. Table 7 shows, the operation of Wind-PVDiesel-Battery system could decrease the total annual cost of the system. The operation of diesel only consumes a lot of fuel so that the operation cost is expensive that shows in table 6. Moreover, the oil prices always increase year by year. Hence, it is cost effective to make combination between renewable energy with DG. In addition, the operation of DG will generate CO 2 emission higher than the proposed system.
Figure 3. ourly IEEE household
Figure 4. Solar radiation in the Ardebil city.
consumptions load profile.
Insulation (W/M2)
1000 800 600 400 200 0 0
5
10 15 Time (Hour of Day)
Figure 5. Wind speed in a year in the Ardebil city.
20
25
Figure 6. Solar radiation in 24 hour
6
3
x 10
Total cost (US$)
2.5 2
1.5 1
0.5 0 0
50
100
150
200
250
Number of iteration Figure 7. Convergence of optimization of algorithme. TABLE 1. DIESEL GENERATOR SPECIFICATIONS (WILSON P75P1 )
Power rating (KW)
Fuel consumpt (L/hr)
Fuel price (US$/Liter)
Capital cost (US$)
Maintenance cost (US$/year)
Lifetime (hour)
2
17.5
0.66
5600
5% of capital cost
15000
TABLE 2. WIND GENERATOR SPECIFICATIONS
Power rating (W)
Cut-in wind speed m/s
Rated wind speed m/s
Cut-out wind speed m/s
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
1000
2.5
11
24
3.000
3% of capital cost
20
TABLE 3. PV MODULES SPECIFICATIONS
Power rating (W)
Open circuit voltage (v)
Short circuit current (A)
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
110
21
7.22
4.86
1% of capital cost
20
TABLE 4. BATTERY BANK SPECIFICATIONS
Nominal capacity (A h)
Voltage (v)
230
12
DOD (%)
Battery charging and discharging efficiency
σ (%)
85%
0.02
80
Capital cost (US$/W)
Maintenance cost (US$/year) Lifetime (year)
0.171
1% of capital cost
4
DC/AC INVERTER SPECIFICATIONS
TABLE 5.
Power rating (W)
Efficiency (%)
Capital cost (US$/W)
Maintenance cost (US$/year)
Lifetime (year)
80
0.707
1% of capital cost
10
1500
TABLE 6.
OPTIMAL SIZING RESULTS FOR DIESEL GENERATOR-ONLY POWER SOURCE Diesel generator capacity (KW)
Total cost (US$)
CO2 emission(Ton /year)
2
113671.08
1398
TABLE 7. HYBRID SYSTEM OPTIMAL SIZING RESULTS
Optimization technique
Populatio n size
Iteration number
Time (s)
BFA
30
150
563. 5
Optimum values of parameters NWG NPV NBAT DG 1
14
7
1
Total cost (US$)
110329.28
CO2 emission(To n/year) 754.92
Conclusion
This paper presents an effective approach for optimal design, energy management, modeling and analysis of stand-alone hybrid Wind-PV-Diesel-Battery energy systems. The main purpose for designing hybrid energy systems is reliable supply of the load, under varying weather conditions, with minimum cost. In this paper, a stand-alone hybrid system is designed for a 20-year period of operation. The 20-year round total system cost is equal to the sum of the respective components capital and maintenance costs. The optimal design, energy management and cost function minimization is implemented using Bacterial Foraging algorithm. The simulation results of algorithm verify that the hybrid Wind-PV-Diesel-Battery systems result in lower system cost compared to diesel generators are used. The total cost of the optimized system showed that the system can deliver energy in a stand-alone installation with an acceptable cost. Hence, the hybrid proposed system is better from economical point of view and environmental effect. References [1] CD. Barley, "Optimal dispatch strategy in remote hybrid power systems". Solar Energy; (1996). 58(4–6):165–79. [2] J. Bryne , B. Shen, W. Wallace, "The economics of sustainable energy for rural development" a study of renewable energy in China. Energy Policy; (1998). 26(1):45–54. [3] Z. Samaras, D. Zafeiris , “Optimization of a wind power fuel cell hybrid system in an autonomous electrical network environment,” International Journal of Renewable Energy, (2007). Vol. 32, pp. 57-79. [4] C. Wang, and M.H. Nehrir, “Power Management of a Stand-Alone Wind/Photovoltaic/Fuel Cell Energy System,” IEEE Trans. on Energy Conversion, (2008). Vol. 23, No.3. [5] D. Kolokotsa et al, “Methodology for optimal sizing of stand-alone photovoltaic/windgenerator systems using genetic algorithms,” International Journal of Solar Energy, (2006).Vol. 80(9), pp. 1072-1088. [6] T. Senjyu, D. Hayashi, A. Yona, N. Urasaki, and T. Funabashi: 2007 " optimal configuration or power generating in isolated island with renewable energy", Renewable Energy, Vol.32, pp.1917-1933 [7] R. Dufo and J.L. Bernal, "Design and control strategies of PV-Diesel system using genetic algorithms, Solar Energy,( 2004). Vol. 79, pp33-46 [8] https://analysis.nrel.gov/homer/ [9] B. Nelson, M. H. Nehrir, and C. Wang, " unit sizing and cost analysis of standalone hybrid wind/pv/feul cell power generation systems", Renewable Energy,( 2006). Vol.31, pp.1641-1656 [10] Y. Xingjia, L. Yingming, B. Jieqiu, and Xing Zuoxia, “Research and Simulation of Direct Drive Wind Turbine VSCF Characteristic” Proc. Int Conf. on Automation and Logistics Qingdao, China, September (2008). PP.1683-1687.
[11] M. S. Alam, and D. W. Gao, “Modeling and Analysis of a Wind/PV/Fuel Cell Hybrid Power System in HOMER” Proc. Int. Conf. on Industrial Electronics and Applications, (2007). PP.1594-1599. [12] H. Yang, W. Zhou, Z. Fang, “ Optimal sizing method for stand-alone hybrid solar–wind system with LPSP technology by using genetic algorithm ,” Sol Energy; (2008). 82:354–67. [13] S.M. Hakimi , S.M. Tafreshi , A. K. Kaviani. August, "Unit sizing of a standalone hybrid power system using particle swarm optimization (PSO). In: Proceeding of the international conference on automation and logistics, Jinan, China. (2007). [14] S.M. Hakimi, S.M. Tafreshi, Aug "Unit sizing of a stand-alone hybrid power system using particle swarm optimization (PSO)". (2007). IEEE ICAL; 18– 21:3107–12. [15] A.H. Shahirinia, S.M. Tafreshi, A.H. Gastaj, A.R. Moghaddomjoo, November 2005 “Optimal sizing of hybrid power system using genetic algorithm”, In: International conference on future power systems (2005); 16–18. [16] K. M. Passino, Jun. " Biomimicry of bacterial foraging for distributed optimization and control",. IEEE Control Syst. Mag., (2002). vol. 22, no. 3, pp.52.67. [17] W. Morgantown, "Emission rates for new DG technologies,. The Regulatory Assistance Project, [Online]". Available,http://www.raponline.org/ProjDocs/DREmsRul/Collfi le/DGEmissionspdf. May(2001). [18] H. Noormohamadi, S. Niknafs, A. Gharaveisi , "Automatic Generation Control of a Multi-Area Power System Using Bacterial Foraging Algorithm, International Review of Electrical Engineering (IREE), May (2010). vol. 5 n. 3, pp. 1183 . 1188.
