Microsoft Word - 6-V8N4(2023)-AITI#11744(303-312).docx Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 English language proofreader: Chih-Wen Teng An Optimal Energy Control System for Campus Microgrid Using Crow Search Algorithm Considering Economic Dispatch Agim Tetuko, Subiyanto*, Muhammad Addin Malik Department of Electrical Engineering, Universitas Negeri Semarang, Central Java, Indonesia Received 10 March 2023; received in revised form 13 September 2023; accepted 14 September 2023 DOI: https://doi.org/10.46604/aiti.2023.11744 Abstract This article presents an optimal energy control system that considers economic dispatch (ED) for a campus microgrid to reduce its operating cost. A newly developed crow search algorithm (CSA) is used to enforce the ED in this work. To achieve this purpose, an optimal size of distributed energy resources (DERs) in the campus microgrid is assumed. CSA is used to optimize the energy control system and find the minimum operating cost of the campus microgrid. To indicate the effectiveness of CSA, several scenarios under various load demand conditions in grid- connected and stand-alone microgrid modes are investigated in this work. According to the findings, the suggested model is capable of sufficient power supply in all scenarios and reduces the operating costs more effectively than the reference delineated in the same case. The outcomes confirm that the suggested model’s performance is optimal for the energy control system of a campus microgrid. Keywords: optimal energy control system, economic dispatch, operating cost, campus microgrid, crow search algorithm 1. Introduction The worldwide renewable energy capacity is expected to increase rapidly over time [1]. This encourages researchers to develop microgrid systems capable of harnessing renewable energy's potential. Generally, the microgrid system can be operated in two modes [2-5]. It’s a stand-alone mode for inaccessible utilities like remote areas or isolated islands [6], and a grid-connected mode that is suitable for supplying residential, urban, commercial, and central areas, up to educational facilities [7]. Among these, educational campuses are particularly well-suited for the implementation and development of microgrid systems. The availability of reliable human resources and regular administration in this area can be classified as a prosumer area that is very suitable for the implementation and development of microgrid systems [8-9]. A wide range of renewable energy sources (RES), including solar energy, wind, water, etc., can be used by microgrids [10]. However, the use of these RES like photovoltaic (PV) depends on stochastic climatic conditions and time resulting in varying electricity generation in the microgrid system [11-12]. Therefore, energy storage systems (ESS) such as battery energy storage systems (BESS), have been combined in microgrid systems to maintain power continuity in microgrids [13]. A diesel generator (DG) that is independent of time and weather generator also has been combined in microgrid to handle more complicated system conditions, such as power outages or when the renewable energy generation and storage systems are no longer able to handle load demand [14]. Although preventive attempts have been made through the combination of ESS and independent generation in microgrids, the complexity of microgrid operation is still a crucial issue that needs to be * Corresponding author. E-mail address: subiyanto@mail.unnes.ac.id Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 304 considered to make microgrids optimal and reliable. In this regard, the operation of microgrid systems needs to consider electricity storage systems, load devices, and generation units, while ensuring optimal and reliable operation of microgrid networks to handle the uncertainties in microgrids and minimize the operating costs [15]. To overcome the economic dispatch (ED) problem, which is based on the optimal power search of each distributed energy resources (DERs) to reduce the operation cost, several matters such as power continuity and various operating constraints are considered [16]. Various methods have been developed by several researchers in this field. Mellouk et al. [17] used the genetic algorithm (GA) to minimize grid charges and peak hours of energy consumption. Ali [18] has also used particle swarm optimization (PSO) and differential evolution (DE) to optimize energy management for microgrids in grid-connected and stand-alone modes. Nevertheless, renowned algorithms do have their restrictions. Examples of common issues include the slow convergence rate of GA, the instability of convergence in DE, and the tendency for both DE and PSO to easily fall into local optima, often requiring a substantial amount of time to converge [19]. One of the latest metaheuristic algorithms is the crow search algorithm (CSA) introduced by Askarzadeh [20], which adopts the memory-based nature of crows to hide and steal their food from other crows. CSA has only one equation and two tuning parameters, making it easy to implement while still being able to maintain the consistency and robustness of the algorithm by spending less computational time to achieve the best fitness value. Spea in [21] has used CSA to optimize the microgrid energy management with DERs consisting of PV/wind/DG systems in remote areas. Dey et al. [19] also used CSA to optimize it in microgrids with PV/wind turbine/DG configurations in grid-connected and stand-alone modes. The research results demonstrated that the algorithm excels in addressing energy management issues within the microgrid systems in terms of power continuity, economy, and emissions. However, most researchers often overlook specific areas of microgrid application, especially those with unique energy consumption patterns like campus areas. On the other hand, the optimal size of the DER implemented is also a crucial factor that must be considered. In this work, the optimal size of DER in a grid-connected microgrid has been implemented in the campus area referred to [14] with the same site study while still being adjusted to the conventional market size. Therefore, this article presents an energy control for optimizing the energy management system in the grid-connected microgrid of campus areas with the ED problem to find the optimal power of each DER and minimize the total operational cost using CSA. In the proposed model, operating costs and the optimal unit capacity of each DER are considered. Furthermore, day-ahead control energy has been thoroughly explored concerning several constraints, such as power generation, electricity price, and various load conditions. This paper is organized as follows: Section 2 presents the ED model of grid-connected campus microgrids, including the objective function, some constraints, and the case site profiles of the campus microgrid. CSA is the method used to solve the ED problem, and case studies are discussed in Section 3. The results of this study are presented in Section 4, and the conclusion is presented in Section 5. 2. ED Model of Grid-Connected Campus Microgrid The ED problem essentially aims to find the lowest generation cost by finding the optimal output power of each DER. Nevertheless, in the ED problem, many aspects can be considered, such as power trading, scheduling, and demand side management (DSM) [16]. ED is a complex problem with large dimensions and many constraints to be considered. In grid- connected microgrids, especially in the campus area, the aspects of power trading, scheduling, and unusual load usage patterns need to be considered. Therefore, in this work, ED will be analyzed on a grid-connected microgrid in the campus area with PV, BESS, and DG configurations as shown in Fig. 1. Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 305 Fig. 1 University campus microgrid structure 2.1. Objective function The main objective of ED in this work is to find the optimal power of each DER in a grid-connected campus microgrid to minimize the total operational cost using CSA. To achieve this purpose, the characteristics and optimal size of each DER must be considered. Furthermore, day-ahead control energy has been thoroughly explored regarding several constraints, such as power generation, electricity price, and various load conditions. In this case, a grid-connected microgrid that uses RES in the form of PV, ESS, and DG, is considered as follows [16]: ( ) ( ) ( ) ( ) ( ) 1 1 1 T n n t c buy buy sell sell t i i MinC F P t C t P t C t P t = = =    = + −  (1) where MinCt is the total cost function of the grid-connected microgrid, and Fc (t) is the total operational cost of all DER units consisting of DG, BESS, and RES units in the form of PV in the time interval t. Cbuy (t), Csell (t), and Pbuy (t), Psell (t) are the powers and prices of electricity purchased and sold at t. RES, such as PV is a source of clean energy that does not require fuel costs. Although there are still installation, maintenance, and operation costs that can be calculated to determine the cost function of RES [21]. In this study, the investment cost of RES is not considered as described by: ( ) ( )1 1 c pv p pvN r F P O P r −         = + − + (2) where Fc (Ppv) is the operating cost of the PV, Ppv is the PV output power, r and N are the interest rate and lifetime of the unit in years, and Op is the ratio of operating and maintenance costs to installed unit power. In [22], the value of r is set to 0.09 and N = 20 years, the value of Op is 0.016 $/kW used in this study and some previous studies [21, 23-24]. The PV cost function in Eq. (2) can be replaced by: ( ) 0.12554647c pv pvF P P×= (3) 2.2. Constraint Total power production from each DER should be equal to the electricity load demand, as represented by: ( ) ( ) ( ) ( ) ( )g pvDG BESS loadP t P t P t P t P t+ + + = (4) ( ) ( ) ( )g but sellP t P t P t= − (5) Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 306 On the other hand, to ascertain system operation stability, each DER should have maximum and minimum limits according to, max min i iP P P≤ ≤ (6) In a grid-connected microgrid, it can be ensured that power buying and selling transactions to and from the grid as represented in: ( ) ( ) ( ) ( ) ( ) { } 0 0,1, , 0  ≥   = ∀  ≤   … p t Cbuy t Pg t if Pg C t t T Csell t Pg t if Pg (7) where �DG (t), PBESS (t), and PPV (t) are the output power of DG, BESS, and PV at t, while PLoad (t) denotes the power demand at t. P� ��� and P� ��� represent the minimum and maximum output power limits of unit i. (t) is the total electricity cost at t. A positive value for grid power means the system is buying power from the grid. Whereas, if the value is negative, it means the system is selling power to the grid. Restrictions and patterns of BESS usage are also important to be considered in the microgrid system’s operation. The output and input power capacities of the BESS, state of charge (SOC) limitations, and changes in the BESS are represented by the following equations [16]: ( ) ( ) ( ) ( ) ( ) ( ) max min max min, , , 0 ; <0 ; 0  ∈ − − − −   > → → = → BESS dis dis ch ch BESS charges BESS discharges BESS idle P t P P P P if P t BESS if P t BESS if P t BESS (8) ( ) ( )min 1 − − = ≤ ≤critical BESSBESS BESS cap e SOC t SOC t e (9) ( ) ( ) ( ) 1 BESS BESS BESS BESS cap P t SOC t SOC t e τ − + = + (10) where � �� and � ��� are the maximum and minimum BESS charges, respectively. Whereas ���� �� ��� ���� ��� is the maximum and minimum BESS discharges. SOCBESS-min (t) is the minimum SOC limit, SOCBESS (t) is the SOC of BESS, ecritical is the energy consumed during critical load, τ is the time period, and eBESS-cap is the capacity of BESS. 2.3. Case site profiles As has been explained in the previous section, the campus area is the most suitable place for the implementation and development of microgrid systems. The focus of this site study is the campus areas, specifically the university buildings in the Electrical Engineering Department of the Faculty of Engineering at Universitas Negeri Semarang. These buildings are E6, E8, and E11 as shown in Fig. 2. It’s located in Sekaran, Gunungpati, Semarang City, Indonesia, with coordinates of 7.05° south latitude, 110.40° east longitude, and an altitude of 187 meters above sea level [25]. As a result, the climate can be classified as tropical, featuring two distinct seasons: the dry season and the rainy season, which occur throughout the year. The daily load pattern in the site study has an unusual pattern where peak hours occur during the day, with the main load consisting of lighting, air conditioning, and campus electrical equipment such as computers, projectors, and electrical trainers set in the laboratory. The total daily load profile is 246.5 kWh and the annual average load is 68,204.1 kWh/year [14]. DG units have been combined in a microgrid system to handle more complicated system conditions, such as power outages or when the renewable energy generation and storage systems are no longer able to handle load demand. Therefore, in this work, the optimal capacity of DG was considered. The cost characteristics of DG units are described in Table 1. Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 307 Fig. 2 E6, E8, and E11 buildings location [25] Table 1 DG generation cost characteristics DG State Generation capacity (kW) Cost ($/kW) DG1 (28 kW) Low 7-14 0.185 Medium 15-21 0.16 High 22-28 0.156 DG2 (48 kW) Low 12-24 0.171 Medium 25-36 0.143 High 37-48 0.132 3. Crow Search Algorithm One of the latest metaheuristic algorithms for solving energy management problems, specifically the ED problem, is CSA. However, in its implementation, CSA also has several tuning parameters that need to be adjusted. In this section, an overview of CSA and its implementation in ED problems is addressed. 3.1. Overview of CSA CSA is one of the latest metaheuristic algorithms introduced by Askarzadeh in 2016 [20], which adopts the memory- based nature of crows to hide and steal their food from other crows. This algorithm adapts the intelligent behavior of crows in a flock to find the best food source based on the objective function. In the optimization perspective, the crow acts as the searcher, the environment serves as the search space, every food source hiding place is a feasible solution, the quality of the food source is the objective function (fitness), and the best food source in the environment is the problem’s global solution [20]. Furthermore, it can be assumed that we have a d-dimensional search space in an environment occupied by several crows, then the position of Crow i at iteration time in the search space can be expressed by xi,iter = (x1 i,iter, x2 i,iter, x3 i,iter,…, xd i,iter) [21]. Where (i = 1, 2, 3,…, N), (iter = 1, 2, 3,…, itermax); d is the number of decision variables, N is flock size, and itermax is the maximum number of iterations. The crow’s current position, represented by mi,iter is stored in the memory of crow as the current best position. To update the crow’s position, it can be assumed that at an iteration, there is a Crow j going to its hiding place mj,iter. Then Crow i decides to follow Crow j to its hiding place. In this case, two possible circumstances will change the crow’s position. Case i: Crow j did not realize that Crow i was following it. Therefore, Crow i will find out the hiding place of Crow j. Case ii: Crow j realizes that Crow i is following it. Therefore, Crow j will try to trick Crow i into protecting its hiding place by randomly moving to other locations in the search space. Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 308 According to the possibility of both cases above, the next position of Crow i can be expressed by: ( ), ,, , , , 1 + + × × − ≥ = i j iter j j iteri iter i iter i iter i iter X r fl m X r AP X a random position otherwise (11) where ri and rj are randomly distributed numbers between 0 and 1. fli,iter is the flight length of Crow i at iteration iter, and APj,iter represents the awareness probability of Crow j at iteration iter. Furthermore, the crow’s memory will be updated by: ( ) ( ) , , 1 , 1 ,, , 1 , + + + =   i iter i iter i iter i iter X f x is better than f m mi iter m otherwise (12) where f (...) represents the objective function value. In this algorithm, diversification and intensification are controlled by fl and AP. The value of fl will be directly proportionate to the similarity value of the crow position, and the AP value will be proportionate to the crow position diversity level. Therefore, setting a small value for fl (fl < 1) generates a local search close to Xi,iter, whereas setting a large value for fl (fl > 1) will cause a global search far from Xi,iter. On the other hand, increasing the value of AP decreases the probability of finding a solution around the current best location and the algorithm will tend to explore solutions in the global search space. Rather than decreasing the value of AP, the algorithm will tend to search around the location where the current best solution is found [15, 17, 22]. 3.2. Implementation of CSA to ED problem Fig. 3 Flowchart of CSA for solving ED Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 309 In the previous section, it was mentioned that the ED problem aims to minimize the total generation cost of all microgrid DER units while considering the power availability and optimal capacity selection of the installed DERs. In this problem, CSA plays a role in finding the most optimal combination of power output from each DER of the campus microgrid to minimize its generation cost while considering power availability. Before presenting the CSA, it must be remembered that to solve the ED problem, it is necessary to meet some equality and inequality constraints as introduced in Eqs. (1)-(10). On the other hand, memory-based algorithms such as CSA essentially have different optimization solutions for each running instance. The settling of the algorithm relies on the initial position and the random movement of the population to find optimal solutions in the search space. Therefore, assessing memory-based algorithms in a single run is not an appropriate comparison. To assess the robustness of the algorithm, multiple test runs are required. The algorithm is considered reliable when it consistently produces results across all runs [26]. The tuning parameters directly affect the final result of the algorithm. Prior studies [21, 26-28] have established and demonstrated the optimal values of fl and AP, which are adopted in this work. Table 2 provides details of CSA tuning parameters. Finally, the detailed steps of CSA implementation for the ED problem are described in the flowchart presented in Fig. 3. Table 2 Details of CSA tuning parameters itermax Flock size fl AP 200 40 2 0.1 4. Results and Discussion In this section, CSA has been used to optimize the ED problem on the grid-connected campus microgrid that uses PV as a renewable energy source and is supported by BESS and DG. To provide a comprehensive investigation of the ED problem in this work, the microgrid was tested under various load demands in grid-connected and stand-alone modes to prove the feasibility of the microgrid system for solving the various possible conditions, as detailed in Table 3. Table 3 Detailed profile of scenarios Scenarios Grid Load demand 1 On < PV generation 2 On > PV generation 3 Off < BESS capacity 4 Off > BESS capacity 5 Off < DG capacity 6 Off > DG capacity 7 Off > DG + BESS capacity As discussed in the previous section, the algorithm tuning parameters greatly affect the final result of the algorithm. The best values of the fl and AP parameters in CSA have also been determined as 2 and 0.1 [17, 22-24]. Nevertheless, population- based algorithms such as CSA also depend on the flock size parameter. Therefore, in this section, various flock size values are compared to find the best flock size parameter value for the CSA to find the minimum operating cost with the lowest error value in the 24-hour simulation period. The error value is intended as the value of the equality constraints of generation power and demand. Table 3 shows the results of system testing with a variety of different flock size values. It can be seen that the flock size value also affects the performance of the algorithm to overcome the ED problems. Based on Table 4, it can be concluded that the best flock size parameter is 40. It is proven that the system with this parameter value gets the minimum value both in the error value and the total operating cost for the 24-hour simulation period. Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 310 Table 4 CSA performance under various flock size values Flock size Error Std. Div Total cost Best Mean Worst 10 0 0.0822397 0.84183 0.177009574 98.198854 20 0 0.0441478 0.635726 0.129478714 95.801322 30 0 0.0346462 0.320607 0.073146489 95.635962 40 0 0.0107046 0.056599 0.015878668 94.498001 50 0 0.0306957 0.640118 0.127186635 94.835849 60 0 0.0273400 0.478794 0.095874636 95.141849 To prove the effectiveness of the CSA algorithm in handling the ED problem, the hourly operation of the microgrid during the 24-hour simulation period is shown in Fig. 4. Based on the figure, it can be seen that during the peak load after 08:00, the proposed system can meet the load demand, even able to charge the BESS and sell the excess power to the grid. On the other hand, when the PV system is unable to meet the load demand, the system combines BESS, grid, and DG to meet the load demand while still keeping the objective of finding the lowest cost and considering the equality constraints of generation and demand. Fig. 4 The hourly output energy pattern of campus microgrid using CSA considering ED Table 5 The total CSA results for the ED problem in each scenario for several hours Scenarios Load (kW) PV (kW) BESS charge (kW) BESS discharge (kW) DG1 (kW) DG2 (kW) Grid purchase (kW) Grid sell (kW) Cost ($) 1 238.230 390.931 24.243 0 0 0 0 128.457 43.019871 2 267.514 161.881 0 27.720 0 26.278 51.635 0 29.295997 3 275.43 265.213 0 10.217 0 0 - - 33.500997 4 319.914 334.133 21.219 0 7 0 - - 44.517511 5 336.470 355.329 18.859 0 0 0 - - 45.741956 6 401.850 279.964 2.796 16.795 71.532 36.355 - - 41.904463 7 544.477 402.278 10.958 10.983 105.185 36.987 - - 70.292881 Finally, the total CSA results for the ED problem in each scenario for several hours are described in Table 5. It should be noted that the scenarios were tested under various load demand conditions on different days for several hours in each scenario, resulting in different amounts of PV generations and load demand. From the table, it can be seen that the proposed system could handle different load demand conditions in both grid-connected and off-grid modes in all scenarios. In grid-connected mode, especially in Scenario 1, the system is even able to sell excess generation back to the grid. In addition, in the worst-case scenario (Scenario 7) when load demand exceeds the capacity of BESS and DG, the system can meet the load demand and Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 311 charge BESS. Fig. 5 presents the comparison of the total operational cost using the proposed method with the conventional method that optimizes the usage pattern of RES as the main source [29]. The results obtained show that the implemented CSA can reduce operating costs by 0.677% with a generation cost of $94.498001. Fig. 5 Cost comparison between energy control system using CSA and conventional method considering optimized use of RES [29] 5. Conclusions This paper presents an optimal energy control system using CSA that considers ED on a university campus microgrid using a DER configuration consisting of PV, BESS, and DG. The optimal capacity of the DER is assumed to support optimal generation within the microgrid system. The overall control energy is optimized by CSA to obtain the least operating cost while still considering the load demand. The optimal flock size parameter that can obtain a minimum value, considering both error and total operating costs, during a 24-hour CSA simulation period, is 40. Furthermore, the proposed system was tested in seven different scenarios under various load demands. In all scenarios, CSA proves sufficient to meet the load demand. In addition, the proposed method has been compared with conventional methods considering the optimized use of RES, and it has been proven that the proposed method can reduce operating costs better. The results of this study validate that the energy control model with CSA, considering ED, offers optimal performance for the operation of the university campus microgrid. Acknowledgments This research is partially supported by UNNES Electrical Engineering Students Research Group (UEESRG), Universitas Negeri Semarang, and also sponsored by the Lembaga Penelitian dan Pengabdian Masyarakat (LP2M) Universitas Negeri Semarang under grant no. 42.22.4/UN37/PPK.4.5/2020 and previous grant research funding. Conflicts of Interest The authors declare no conflict of interest. References [1] D. Nagpal and D. Hawila, Renewable Energy Market Analysis: Southeast Asia, Abu Dhabi: IRENA, 2018. [2] P. Niranjan, R. K. Singh, and N. K. Choudhary, “Comparative Study of Relay Coordination in a Microgrid with the Determination of Common Optimal Settings Based on Different Objective Functions,” International Journal of Engineering and Technology Innovation, vol. 12, no. 3, pp. 260-273, January 2022. [3] M. Ganjian-Aboukheili, M. Shahabi, Q. Shafiee, and J. M. Guerrero, “Seamless Transition of Microgrids Operation from Grid-Connected to Islanded Mode,” IEEE Transactions on Smart Grid, vol. 11, no. 3, pp. 2106-2114, May 2020. [4] S. Beheshtaein, R. Cuzner, M. Savaghebi, and J. M. Guerrero, “Review on Microgrids Protection,” IET Generation, Transmission & Distribution, vol. 13, no. 6, pp. 743-759, March 2019. Advances in Technology Innovation, vol. 8, no. 4, 2023, pp. 303-312 312 [5] E. E. Pompodakis, G. C. Kryonidis, and M. C. Alexiadis, “A Comprehensive Load Flow Approach for Grid-Connected and Islanded AC Microgrids,” IEEE Transactions on Power Systems, vol. 35, no. 2, pp. 1143-1155, March 2020. [6] A. Irigoyen Tineo, “Energy Management System Benchmarking for a Remote Microgrid,” Bachelor thesis, Energy Engineering, Universidad Carlos III de Madrid, Getafe, 2017. (In Español) [7] H. Dagdougui, L. Dessaint, G. Gagnon, and K. Al-Haddad, “Modeling and Optimal Operation of a University Campus Microgrid,” IEEE Power and Energy Society General Meeting, pp. 1-5, July 2016. [8] H. Abd ul Muqeet, H. M. Munir, A. Ahmad, I. A. Sajjad, G. J, Jiang, and H. X. Chen, “Optimal Operation of the Campus Microgrid considering the Resource Uncertainty and Demand Response Schemes,” Mathematical Problems in Engineering, vol. 2021, article no. 5569701, 2021. [9] F. Iqbal and A. S. Siddiqui, “Optimal Configuration Analysis for a Campus Microgrid—A Case Study,” Protection and Control of Modern Power Systems, vol. 2, article no. 23, June 2017. [10] I. Padrón, D. Avila, G. N. Marichal, and J. A. Rodríguez, “Assessment of Hybrid Renewable Energy Systems to Supplied Energy to Autonomous Desalination Systems in Two Islands of the Canary Archipelago,” Renewable and Sustainable Energy Reviews, vol. 101, pp. 221-230, March 2019. [11] M. A. Hossain, H. R. Pota, S. Squartini, and A. F. Abdou, “Modified PSO Algorithm for Real-Time Energy Management in Grid-Connected Microgrids,” Renewable Energy, vol. 136, pp. 746-757, June 2019. [12] B. Meryem, N. Ahmed, and F. Ahmed, “Simulation and Implementation of a Modified ANFIS MPPT Technique,” Advances in Technology Innovation, vol. 5, no. 4, pp. 230-247, October 2020. [13] K. Ullah, G. Hafeez, I. Khan, S. Jan, and N. Javaid, “A Multi-Objective Energy Optimization in Smart Grid with High Penetration of Renewable Energy Sources,” Applied Energy, vol. 299, article no. 117104, October 2021. [14] S. Subiyanto, M. A. Prakasa, P. Wicaksono, and M. A. Hapsari, “Intelligence Technique Based Design and Assessment of Photovoltaic-Battery-Diesel for Distributed Generation System in Campus Area,” International Review on Modelling and Simulations, vol. 13, no. 1, pp. 63-73, 2020. [15] E. Grover-Silva, M. Heleno, S. Mashayekh, G. Cardoso, R. Girard, and G. Kariniotakis, “A Stochastic Optimal Power Flow for Scheduling Flexible Resources in Microgrids Operation,” Applied Energy, vol. 229, pp. 201-208, November 2018. [16] M. Abou Houran, W. Chen, M. Zhu, and L. Dai, “Economic Dispatch of Grid-Connected Microgrid for Smart Building considering the Impact of Air Temperature,” IEEE Access, vol. 7, pp. 70332-70342, 2019. [17] L. Mellouk, M. Boulmalf, A. Aaroud, K. Zine-Dine, and D. Benhaddou, “Genetic Algorithm to Solve Demand Side Management and Economic Dispatch Problem,” Procedia Computer Science, vol. 130, pp. 611-618, 2018. [18] M. Y. Ali, “Energy Management System of a Microgrid With Distributed Generation,” Master thesis, Department Electrical and Computer Engineering, University of Ontario Institute of Technology, Canada, 2019. [19] B. Dey, B. Bhattacharyya, A. Srivastava, and K. Shivam, “Solving Energy Management of Renewable Integrated Microgrid Systems Using Crow Search Algorithm,” Soft Computing, vol. 24, no. 14, pp. 10433-10454, July 2020. [20] A. Askarzadeh, “A Novel Metaheuristic Method for Solving Constrained Engineering Optimization Problems: Crow Search Algorithm,” Computers & Structures, vol. 169, pp. 1-12, June 2016. [21] S. R. Spea, “Combined Economic Emission Dispatch Solution of an Isolated Renewable Integrated Micro-Grid Using Crow Search Algorithm,” 21st International Middle East Power Systems Conference, pp. 47-52, December 2019. [22] N. Augustine, S. Suresh, P. Moghe, and K. Sheikh, “Economic Dispatch for a Microgrid Considering Renewable Energy Cost Functions,” IEEE PES Innovative Smart Grid Technologies, pp. 1-7, January 2012. [23] I. N. Trivedi, P. Jangir, M. Bhoye, and N. Jangir, “An Economic Load Dispatch and Multiple Environmental Dispatch Problem Solution with Microgrids Using Interior Search Algorithm,” Neural Computing and Applications, vol. 30, no. 7, pp. 2173-2189, October 2018. [24] B. Dey, S. K. Roy, and B. Bhattacharyya, “Solving Multi-Objective Economic Emission Dispatch of a Renewable Integrated Microgrid Using Latest Bio-Inspired Algorithms,” Engineering Science and Technology, an International Journal, vol. 22, no. 1, pp. 55-66, February 2019. [25] Google Maps, “Engineering Faculty of UNNES,” https://goo.gl/maps/1UGaiFzLxUbcScop7, October 24, 2022. [26] F. Mohammadi and H. Abdi, “A Modified Crow Search Algorithm (MCSA) for Solving Economic Load Dispatch Problem,” Applied Soft Computing, vol. 71, pp. 51-65, October 2018. [27] A. F. Sheta, “Solving the Economic Load Dispatch Problem Using Crow Search Algorithm,” Proceedings of the 8th International Multi-Conference on Complexity, Informatics and Cybernetics, pp. 95-100, March 2017. [28] R. Habachi, A. Touil, A. Boulal, A. Charkaoui, and A. Echchatbi, “Resolution of Economic Dispatch Problem of the Moroccan Network Using Crow Search Algorithm,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 13, no. 1, pp. 347-353, January 2019. [29] M. A. Prakasa and S. Subiyanto, “Optimal Cost and Feasible Design for Grid-Connected Microgrid on Campus Area Using the Robust-Intelligence Method,” Clean Energy, vol. 6, no. 1, pp. 59-76, February 2022. Copyright© by the authors. Licensee TAETI, Taiwan. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC) license (https://creativecommons.org/licenses/by-nc/4.0/).