NSGAII-Based Fuzzy PID Controller for Load Frequency Control of Multi-Microgrids

Transcription

1 64 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 NSGAII-Based Fuzzy PID Controller for Load Frequency Control of Multi-Microgrids H. Shayeghi *,1, H. A. Shayanfar 2, M. Esmaeili 1 1 College of Technical & Engineering, University of Mohaghegh Ardabili, Ardabil, Iran. 2 College of Technical & Engineering, South Tehran Branch, Islamic Azad University, Tehran, Iran hshayeghi@gmail.com, hashayanfar@gmail.com, m.esmaeili@uma.ac.ir Abstract- In this paper, a fuzzy-pid controller is proposed for load frequency control (LFC) of an islalated multimicrogrid. Parameters of the proposed fuzzy PID controller are optimally tuned using NSGA-II algorithm to minimize overshoot, settling time and integral square error over a widerange of load variations. Also, the performance of the proposed control strategy is compared with optimized classical PID controller using NSGA-II algorithm to illustrate its effectiveness for solution LFC task of an islolated multimicrogrid. The performances of the controllers are simulated using MATLAB/SIMULINK package. A comparison of controllers indicates the superiority of the proposed fuzzy controller over optimal PID controller for the same conditions. Index Terms-- Fuzzy PID controller, LFC, Multi-micro grid, NSGA-II algorithm. I. INTRODUCTION Microgrids (MGs) could provide a solution to many of the problems of the current power system as they can independently operate as islanded and locally supply power to load centers [1-2]. In general, providing power in these small-scale grids could reduce energy losses, reduce supply disruptions, improve efficiency, and support self-decision making for small entities [3]. MGs are also considered to be more appropriate for the application of distributed renewable resources [4]. Recently, interconnecting adjacent MGs through the tie lines and building a resilient multi-microgrid (MMG) energy system has attracted more attention among researchers [5]. Compounding the various kinds of power sources would impact the quality of power supply within the microgrid and cause high frequency fluctuation due to the sudden variations in load [6,7]. A detailed literature review of load frequency control may be found in [8]. PID control scheme has been considered as a reliable and applicable solution for the load frequency control (LFC), because of its ease of use and also its robustness [9-11]. A novel technique for tuning PID controllers for LFC problem solution in an autonomous hybrid microgrid using Biogeography Based Optimization (BBO) was reported in [12]. The fuzzy logic based Proportional-Integral- Derivative (FID) control has extensively received attentions in various power systems applications. The fuzzy logic PID controller in a closed loop control system Corresponding author, H. Shayeghi (hshayeghi@gmail.com) is a non-linearity between the inputs and outputs, which can be tuned easily to match the desired performance of the control system in a more heuristic method without delving into the mathematical description of the modeled nonlinearity [13, 14]. A Fuzzy logic control system based on extended Proportional Integral (PI) controller has been suggested in [15] and [16]. A new online intelligent approach by using a combination of the fuzzy logic and the particle swarm optimization (PSO) techniques for optimal tuning of the most popular existing proportionalintegral (PI) based frequency controllers in the ac MG systems are addressed in [17]. In this paper, a fuzzy-pid controller is proposed for load frequency control of an islanded multi-microgrid. Also, optimal tuning of controller parameters to achieve the good performance and dynamics response is necessary. For this reason, parameters of the proposed fuzzy PID controller are optimally tuned using NSGA-II algorithm to minimize overshoot, settling time and integral square error over a wide-range of load variations. Also, the performance of the proposed control strategy is compared with classical PID controller optimized by using NSGA-II algorithm to illustrate its effectiveness for solution LFC task of an isolated multi-microgrid. The study MMG consists with two MGs that are interconnected by tie-lines. The simulation results A indicates the superiority of the proposed fuzzy PID controller over optimal PID controller for a wide-range of load variations. II. SYSTEM DESCRIBTION Here, an isolated MMG system is considered as a case study [17]. The study MMG consists with two MGs, each made up of diesel generator (DEG), wind turbine generator (WP), photovoltaic (PV), fuel cell (FC), battery energy storage system (BESS), flywheel, energy storage system (FESS), and load. The MGs are interconnected by tie-lines. Schematic diagram of an MG and an MMG are shown in Figs. 1 and 1, respectively. The parameters value of the ac MG system is given in Table I. P MG1 K/s MG

2 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 65 III. CONTROL STRATEGIES In general, the structure of a PID controller can be described by: de(t) (1) u(t) =K e(t) +K e(t)dt + K dt Where, e(t) is the error, u(t) the controller output, and,, and are the proportional, integral and derivative gains, respectively. The effectiveness of PID control depends on values of, and gains. It is shown that the appropriate selection of PID controller parameters results in satisfactory performance during system upsets. Therefore, the optimal tuning of a PID gains is required to get the desired level of robust performance. Here, the gains of PID controller is optimized by using NSGA-II algorithm. However, it is shown that the PID controller have poor response for LFC task of MMG frequency control when the operating conditions are changed due to stochastic nature of daily load variation of. In order to overcome to this drawbacks, In this paper, the PID frequency controller is replaced by a fuzzy logic controller (FLC). The fuzzy logic PID controller in a closed loop control system is a nonlinearity between the inputs and outputs, which can be tuned easily to match the desired performance of the control system in a more heuristic method without delving into the mathematical description of the modeled nonlinearity [13, 14]. In the FLC, the reference frequency f is compared with the actual frequency to obtain the frequency error e(t). The inputs of the controller are the error (e) and the rate of change of the error e), while the outputs are the controller gain,,, and. Fig. 2 shows the proposed FPID model: Fig. Parameter 1. Schematic values diagram of the of block multi-microgrid diagram (Fig. system, 1) are given in Table I. studied microgrid Table I The parameters value of the ac MG system Parameter Kfess Tfess (s) Kbess Tbess (s) Kdeg Tdeg (s) Tgi (s) Kwtg Twtg (s) Ka Value Paremeter Kae Tae (s) Kfc Tfc (s) Kpv Tpv (s) Tt (s) R (Hz/pu) D (Hz/pu) 2H (pu s) Value Fig 2. The proposed FPID model In designing the FID controller, the important procedure is determination of the scale factors (,, and ), membership functions and control rules. In general, they are determined by the trial and error and designer s experiences but here they set as optimization parameters. Thus, optimal tuning of controller parameters to achieve the good performance and dynamics response is necessary. For this reason, parameters of the proposed fuzzy PID controller are optimally tuned using NSGA-II algorithm to minimize overshoot, settling time and integral square error over a wide-range of load variations. Every linguistic variable has seven values defined as NL (negative large), NM (negative medium), NS (negative small), R (zero), PS (positive small), PM (positive medium) and PL (positive large), respectively. The control laws of the FLC are represented by a set of chosen IF THEN rules. The designed fuzzy rules used in this work are given in Table II. f ' f Table II The designed fuzzy rules IV. SIMULATION RESULT Optimal PID controller and optimal fuzzy PID controller parameters are tuned using NSGA-II algorithm [18] to minimize Overshoot Percentage (M ), Settling Time (t ), and Integral of Square Error (ISE) as the optimization objective, and the robustness in frequency domain as the constraints. So, a multi-objective optimization cost function can be defined as follows:

3 66 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 The NSGA-II algorithm is chosen as the optimizer in the research reported in this paper Because of its low computational requirements, and its parameter less sharing scheme [18]. The population of NSGA-II is taken as 5 individuals and evolutionary cycle has stopping criterion of 1 generations. The flowchart of the NSGA- II method is shown in Fig Fig. 3. Flowchart of NSGA-II algorithm After using NSGA-II algorithm the parameters of optimal PID parameters and optimal fuzzy PID control are determined and listed in Table III. Table III Optimal PID parameters Optimal PID Optimal FID K 1 K 1 K.3 K.22 K.83 K.78 For comparing the Optimal PID and Optimal fuzzy PID several simulation tests are carried out and the performances of the proposed control methods are evaluated. For the sake of comparison in a sever condition, performance of the optimal fuzzy PID and the PID controllers are examined with applying pu step load disturbance. The closed-loop frequency responses are shown in Fig. 4 in MG Fig. 4. MMG frequency response; Solid-blue (OPID), Dashed-red (OFPID), 1 2 In this case, the proposed optimal control method provides a much better performance. To illustrate the dynamic response of the MMG, the closed-loop system is examined in the face of a multiple step load disturbance which is plotted in Fig. 5. The MMG frequency response using the Optimal PID and Optimal fuzzy PID controller in the face of multiple step load disturbance is shown in Fig. 5(b ) whereas, and P L are frequency deviation, and load disturbance pattern, respectively; which their values are given in pu. The suffix 1 and 2 are used for MG1 and MG2 quantities, respectively. As shown, the proposed fuzzy-pid controller regulates the system frequency following disturbance quite better than the optimal PID controllers. The fuzzy-pid controller has suitable performance in terms of settling-time and minimizing of frequency deviations. A better performance of the proposed intelligent control methodology is clearly visible from system frequency response following first step increase in the load disturbance, Fig. 4. As in the real world system, each microgrid subjects to different kinds uncertainties and disturbances because of plant parameter variations and system modeling errors due to some approximations in model linearization and un-modeled dynamics. As this may degrade the closedloop system performance, and the other word, seriously one of the main advantages of the intelligent control methods is robustness against environmental and dynamical changes. Thus, in the third case, for showing the fuzzy-pid controller, the main system parameters, in the frequency response model D (damping coefficient), H

4 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 67 (inertia constant), (turbine time constant), (generator time constant) and are significantly changed. The purpose of this scenario is to test the robustness of the proposed controller against uncertainties and random large load disturbances. The MMG frequency response using the optimal FPID and PID controllers, in the case of a step load disturbance ( = ) is applied to MG1 and MG2, when uncertainty parameters of the systems as mentioned in the above is varied -2 % and +2% from nominal values is shown in Figs. 6 and 7, respectively. Using the proposed method, the frequency deviation of the two MGs are quickly driven back to zero and have very small settling time and overshoot Delta PL(pu) Fig. 6. Deviation of frequency in the face of a step load disturbance for 2 % uncertainty state; Solid-blue (OPID), Dashed-red (OFPID), Fig. 5. Multiple step load disturbances. MMG frequency response; Solid-blue (OPID), Dashed-red (OFPID). Here, to illustrate the effectiveness of the proposed controller quantitatively, the performance indices such as undershoot, overshoot and settling time are calculated for three operating conditions when the uncertainty system parameters is varied from nominal values, and results are given in Table IV Fig. 7. Deviation of frequency in the face of a step load disturbance for + 2 % uncertainty state; Solid-blue (OPID), Dashed-red (OFPID), 1 2

5 68 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 Uncertainty Normal -2% +2% Parameter OPID OFPID OPID OFPID OPID OFPID Table IV Comparison of overshoot and undershoot and setting time UnderShoot(pu) OverShoot(pu) Setting UnderShoot(pu) OverShoot(pu) Setting V. CONCLUSIONS In this paper, a fuzzy PID type controller has been investigated for the load frequency control of isolated multimicrogrid which ideally, is used in practical industries. The salient advantage of this controller is its high insensitivity and flexibility to large load changes and plant parameter variations even in the presence of system physical nonlinearities. Parameters of the proposed FPID controller is optimized using NSGA-II algorithm In order to enhance its performance and minimize overshoot, settling time. Simulation results on a multi-microgrid show that the proposed method can guarantee the robust performance for a wide range of plant parameter changes even in the presence of system nonlinearities such as time delay, generation rate constraint. In addition, the system performances such as undershoots, settling time, maximum oscillations with the proposed strategies are better than the conventional optimized PID controllers. The proposed method does not require an accurate model of the LFC problem and has simple structure. Thus, its implementation is fairly easy and can be useful for the real world multi-microgrid system. References [1] L. Che, M. Khodayar, M. Shahidehpour., "Only connect: Microgrids for distribution system restoration," Power and Energy Magazine, IEEE, vol. 12, pp. 7-81, 214. [2] H. Shayeghi, A. Gasemi, mprovement of Frequency Fluctuations in Microgrids Using an Optimized Fuzzy P-PID Controller by Modified Multi Objective Gravitational Search Algorithm. Iranian Journal of Electrical and Electronic Engineering, Vol. 12, pp , 216. [3] M. H. Amini, B. Nabi, M.-R. Haghifam., "Load management using multi-agent systems in smart distribution network," in IEEE PES General Meeting, Vancouver, BC, Canada, 213. [4] I. R. E. Series, "Microgrids and active distribution networks," The Institution of Engineering and Technology, 29. [5] A. H. Chowdhury and M. Asaduz-aman, "Load frequency control of multi-microgrid using energy storage system," International Conference on Electrical and Computer Engineering, 214, pp [6] P. Arul, V. K. Ramachandaramurthy, R. Rajkumar, "Control strategies for a hybrid renewable energy system: A review," Renewable and Sustainable Energy Reviews, vol. 42, pp , 215. [7] B. Mohanty, S. Panda, P. Hota, "Controller parameters tuning of differential evolution algorithm and its application to load frequency control of multi-source power system," International Journal of Electrical Power & Energy Systems, vol. 54, pp , 214. [8] S.K. Pandey, S.R. Mohanty, N. Kishor, "A literature survey on load frequency control for conventional and distribution generation power systems," Renewable and Sustainable Energy Reviews, vol. 25, pp , 213. [9] R. H. Kumar and S. Ushakumari, "Biogeography based tuning of PID controllers for Load Frequency Control in microgrid," International Conference on Circuit, Power and Computing Technologies, 214, pp [1] M. N. Anwar and S. Pan, "A new PID load frequency controller design method in frequency domain through direct synthesis approach," International Journal of Electrical Power & Energy Systems, vol. 67, pp , 215. [11] E. J. Oliveira, L. M. Honório, A. H. Anzai, L. W. Oliveira, E. B. Costa, "Optimal transient droop compensator and PID tuning for load frequency control in hydro power systems," International Journal of Electrical Power & Energy Systems, vol. 68, pp , 215. [12] R. H. Kumar and S. Ushakumari, "Biogeography based tuning of PID controllers for Load Frequency Control in microgrid," International Conference on Circuit, Power and Computing Technologies, 214, pp [13] H. Shayeghi, H.A. Shayanfar, A. Jalili, Multi stage fuzzy load frequency control using PSO, Energy Conversion and Management, vol. 49, pp , 28. [14] G. C. Sekhar, R. K. Sahu, S. Panda, "Load frequency control with fuzzy-pid controller under restructured environment," International Conference on Control, Instrumentation, Communication and Computational Technologies, 214, pp [15] H. Shayeghi, H.A. Shayanfar, A. Jalili, Load Frequency Control Strategies: A State- of-the-art Survey for the Researcher, Energy Conversion and Management, vol. 49, pp , 29. [16] C. F. Juang, C. F. Lu, Power system load frequency control by evolutionary fuzzy PI controller, in Proc. of IEEE International Conf. on Fuzzy Systems, Vol. 2, pp , Budapest, Hungary, July 24. [17] H. Bevrani, F. Habibi, P. Babahajyani, M. Watanabe, Y. Mitani,, "Intelligent frequency control in an AC microgrid: Online PSO-based fuzzy tuning approach," IEEE Transactions on Smart Grid, vol. 3, pp , 212. [18] H. Shayeghi and Y. Hashemi, "Application of fuzzy decision-making based on INSGA-II to designing PV wind hybrid system," Engineering Applications of Artificial Intelligence, vol. 45, pp. 1-17, 215. BIOGRAPHY Hossein Shayeghi received the B.S. and M.S.E. degrees in Electrical and Control Engineering in 1996 and 1998, respectively. He re ceived his Ph.D. degree in Electrical Engineering from Iran University of Science and Technology, Tehran, Iran in 26. Currently, he is a full Professor in Technical Engineering Department of University of Mohaghegh Ardabili, Ardabil, Iran. His research

6 Int'l Conf. on Advances on Applied Cognitive Computing ACC'17 69 interests are in the application of robust control, artificial intelligence and heuristic optimization methods to power system control design, operation and planning and power system restructuring. He has authored and co-authored of 6 books in Electrical Engineering area all in Farsi, one book and four book chapters in international publishers and more than 35 papers in international journals and conference proceedings. Also, he collaborates with several international journals as reviewer boards and works as editorial committee of three international journals. He has served on several other committees and panels in governmental, industrial, and technical conferences. He was selected as distinguished researcher of the University of Mohaghegh Ardabili several times. In 27 and 21 he was also elected as distinguished researcher in engineering field in Ardabil province of Iran. Furthermore, he has been included in the Thomson Reuters list of the top one percent of most-cited technical Engineering scientists in 215 and 216, respectively. Also, he is a member of Iranian Association of Electrical and Electronic Engineers (IAEEE) and senior member of IEEE. Heidar Ali Shayanfar received the B.S. and M.S.E. degrees in electrical engineering in 1973 and 1979, respectively. He received the Ph.D. degree in electrical engineering from Michigan State University, East Lansing, MI, USA, in Currently, he is a Full Professor with the Department of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran. His research interests are in the application of artificial intelligence to power system control design, dynamic load modeling, power system observability studies, voltage collapse, congestion management in a restructured power system, reliability improvement in distribution systems, smart grids and reactive pricing in deregulated power systems. He has published more than 52 technical papers in the international journals and conferences proceedings. Dr. Shayanfar is a member of the Iranian Association of Electrical and Electronic Engineers.