Title: Optimal energy management strategy in microgrids with mixed energy resources and energy storage system
Abstract: IET Cyber-Physical Systems: Theory & ApplicationsVolume 5, Issue 1 p. 80-84 Research ArticleOpen Access Optimal energy management strategy in microgrids with mixed energy resources and energy storage system Yordanos Kassa Semero, Corresponding Author Yordanos Kassa Semero [email protected] Trina Solar Energy (Shanghai) Co. Ltd, 200000 Shanghai, People's Republic of ChinaSearch for more papers by this authorJianhua Zhang, Jianhua Zhang School of Electrical and Electronic Engineering, North China Electric Power University, 102206 Beijing, People's Republic of ChinaSearch for more papers by this authorDehua Zheng, Dehua Zheng Goldwind Science and Technology Co., Ltd, 100176 Beijing, People's Republic of ChinaSearch for more papers by this author Yordanos Kassa Semero, Corresponding Author Yordanos Kassa Semero [email protected] Trina Solar Energy (Shanghai) Co. Ltd, 200000 Shanghai, People's Republic of ChinaSearch for more papers by this authorJianhua Zhang, Jianhua Zhang School of Electrical and Electronic Engineering, North China Electric Power University, 102206 Beijing, People's Republic of ChinaSearch for more papers by this authorDehua Zheng, Dehua Zheng Goldwind Science and Technology Co., Ltd, 100176 Beijing, People's Republic of ChinaSearch for more papers by this author First published: 13 November 2019 https://doi.org/10.1049/iet-cps.2019.0035Citations: 13AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract The continued growth of distributed generation (DG) in the electrical grid has led to the expansion of microgrids. Microgrids contain distributed power generation units, energy storage devices, and controllable loads with the capability to operate in both grid-connected and island modes. The economic operation of a microgrid is achieved through an energy management system that optimally schedules DGs and storage devices and continuously balances supply and demand. In this study, a formulation of optimal unit commitment (UC) and dispatch scheduling of DGs in a grid-connected microgrid system is presented. Mixed-integer linear programming is used to implement the optimal UC and dispatch scheduling model. The objective is to minimise the overall operating cost of the system by optimally utilising an energy storage device and a combined heat and power (CHP) generation unit using load and renewable energy generation prediction. Operational constraints such as generation limits of DGs, battery charging/discharging limits and state-of-charge limits are to be satisfied during all intervals of operation. Simulation results indicate that the operational cost of the system is significantly reduced through optimal scheduling of an energy storage system and a CHP unit using the proposed strategy. Nomenclature t time generated power wind energy conversion efficiency air density A turbine blade area V wind speed maximum generation capacity of a PV installation G solar radiation reference irradiance certain irradiance value set as 150 W/m2 operational cost of a microturbine (MT) C operational cost of microgrid f cost function of a conventional generation unit output power of MT minimum output power limit of MT maximum output power limit of MT efficiency combined heat and power (CHP) efficiency electrical efficiency of MT power supplied by the utility utility electricity tariff output power of wind turbine output power of the PV system load demand charging rate of energy storage maximum charging rate of energy storage charging efficiency of energy storage storage capacity of an energy storage system (ESS) SOC state-of-charge (SOC) of ESS minimum SOC limit of ESS maximum SOC limit of ESS output power of the jth conventional unit at t starting cost of the jth conventional unit at t status (UC) of the jth conventional unit at t k number of time intervals in a simulation day 1 Introduction In the last few decades, major economic and environmental policy shifts have resulted in an increased number of distributed generation plants. This has brought major topological changes to the electric power systems and opened up space for the expansion of microgrids. A microgrid can be defined as a cluster of loads, distributed energy resources (DERs) and energy storage systems (ESSs) operated in coordination to reliably supply electricity, and can be connected to a host power system at the distribution level at a single point of connection, called the point of common coupling (PCC) [1]. The future power grid could be described as an interconnection of several microgrid systems, where each microgrid could mainly consume or produce electric power [2]. Coordinated operation of distributed energy sources in the form of microgrids has the potential to increase system reliability and power quality due to the decentralisation of supply [3]. Microgrids can also provide economic advantages through small-scale renewable energy investments and implementation of appropriate control and management facilities [4]. DERs that constitute microgrids typically include wind turbines (WTs) and photovoltaic (PV) installations, where the primary energy sources are intermittent and fluctuating. This has led to trends to build microgrids equipped with sufficient energy storage facilities and spinning reserves. Apart from dealing with the fluctuations, ESSs play an important role in deciding optimal moments to purchase and sell electricity to the utility grid. From operational requirements point of view, the microgrid system is required to achieve a coordinated control and optimal management of energy resources while minimising operational and maintenance costs. Efficient operation of a microgrid system needs an energy management system (EMS) which plays a central role of controlling flow of electricity in the system by optimally setting amount of power exchanged between the microgrid and the utility grid, adjusting settings of dispatchable generation units and controllable loads depending on the current and predicted electricity tariff information, generation forecast, and load forecast in order to satisfy certain objectives and technical constraints [5]. In order to take full advantage of the economic and environmental benefits that come from optimal control of microgrids, efforts have been made to develop and apply various optimisation techniques and control strategies [6]. The stochastic programme-based strategy is formulated in [7] with the aim of minimising overall cost of electricity and natural gas usage in a building. Investigation of robust scheduling and unit commitment problems with penalty-based costs for uncertainty in generation and consumption is provided in [8]. Application of neural networks to solve the energy supply optimisation problem in power supply networks has been demonstrated in [9, 10]. The dynamic economic dispatch model for microgrids employing a dynamic programming algorithm is presented in [11]. Xu et al. [12] propose a day-ahead battery scheduling for microgrids based on ordinal optimisation theory. A microgrid optimisation strategy based on linear programming is presented in [13]. A quantum-inspired evolutionary algorithm is employed in [14] to optimise operation of a microgrid in terms of reduction of operational cost and emissions. This paper deals with the application of mixed-integer linear programming (MILP) to establish optimal UC and power dispatching model for microgrids with mixed resources and ESS. The strategy is aimed at achieving optimal UC plan and dispatch scheduling of DERs in different operational modes while meeting various operational constraints on the bases of estimated output power from renewable energy resources, expected load demand details, cost of electricity production from conventional energy generation systems, electricity tariff, and other necessary information. Two operating scenarios where a microturbine (MT) unit in a microgrid produces electricity alone and provides the production of combined usable heat and power are separately examined. A relatively simple and efficient resource scheduling scheme and EMS for microgrids are proposed. The schematic diagram of the microgrid EMS is shown in Fig. 1. The study is conducted considering a grid-connected industrial microgrid system located in Beijing as a case study. Goldwind smart microgrid system consisting of a 2.5 MW WT, PV installations with a pick capacity totaling 480 kWp, 130 kW MT unit, and an energy storage unit rated at 150 kW is used as the simulation test case. All operational costs and electricity prices used in the calculations are in Chinese Yuan (RMB, ¥). Figure 1Open in figure viewerPowerPoint Schematic diagram of microgrid EMS The remaining sections are organised as follows. Section 2 presents modelling of components of the case study microgrid. Section 3 describes the formulation of the optimisation problem. Section 4 presents and discusses the results of the study. Section 5 provides the conclusions drawn from the results. 2 Modelling of components 2.1 Wind energy conversion system (WECS) WECSs are used to convert kinetic energy carried by the wind into mechanical form and then into electrical energy. The amount of electrical power that can be obtained from WTs depends on the surrounding air density, wind speed, turbine blade area, and wind energy conversion efficiency (1) The generated power is denoted by , is the density of air, A is the turbine blade area, V denotes the wind speed and is the wind energy conversion efficiency, which can be described as the turbine power in proportion with wind power and is related to aerodynamic characteristics of the turbine blades [15]. From the specifications of the WT installed at the case study microgrid system, the cut-in and cut-out speeds of the turbine are 3 and 25 m/s, respectively, and the rated power is 2.5 MW at a rated wind speed of 11 m/s. In this study, a hybrid wind power generation prediction technique described in [16] is considered to predict the output power of the WT. 2.2 PV system PV systems generate electricity by converting sunlight into electrical energy. PV systems do not create air pollution or any kind of emissions while operating. No rotating masses are required to generate electricity using PV arrays; hence minimal maintenance is required. The output power of a PV installation is given below [17]: (2) where represents the maximum generation capacity of the system at standard test conditions (STC), G is the predicted solar radiation, is a certain radiation point set as 150 W/m2, is the reference irradiance, which is 1000 W/m2, and K is the temperature coefficient. In this study, is 480 kW. In this work, a short-term PV power generation forecasting technique presented in [18] is used to estimate the output power of the PV system in the case study microgrid system. 2.3 MT system MT generation systems are small electricity generation units that burn gaseous fuels to turn a turbine, which is mechanically connected to the shaft of an electrical generator. They operate with low vibration and noise with relatively fast response to load variation. They offer higher electrical efficiency compared to traditional gas and diesel generators in the same size range. They are also known for not being significant threats to the environment as the emissions from these devices are very low. The major strength of MTs is the combined heat and power (CHP) application, where the clean exhaust heat can be recovered and utilised. These characteristics make them an attractive distributed generation choice for microgrid applications. The MT system in the case study microgrid has two units of 65 kW generating capacity each. Therefore, the total capacity of the MT system is 130 kW. The electrical energy generation cost function of the MT, derived from its gas consumption profile based on current gas price in Beijing, is given below: (3) where and , respectively, stand for the energy cost in RMB/kWh and the power produced by the MT in kilowatts at the time interval . The electrical efficiency of the MT is 29%. The overall CHP efficiency of the MT can reach >80%. Therefore, the cost function defined in (3) is modified to reflect the change in efficiency when the MT supplies both electricity and heat, assuming 80% efficiency. It is conservatively assumed that the additional operational cost reduction benefit obtained when the MT provides both electricity and heat could be approximately represented by improvement in electricity production. The improvement in efficiency when the MT operates in cogeneration mode (usable heat and power) over electricity generation mode is given below: (4) where stands for efficiency, is the overall or CHP efficiency of the MT when producing both heat and electricity, and is the electrical efficiency of MT, which represents the efficiency of the device when it provides only electricity. It is assumed that the cost function of the MT in combined usable heat and power production will be reduced proportionally to the increment in overall efficiency. This assumption does not differentiate between the value of the electrical power output and heat energy production as we do not have reliable data of heat energy price in similar application context with electricity. Instead, we treated electrical power output and heat energy output as having the same value which allows them to be added for further analyses. In reality, however, it is to be noted that electricity is considered a more valuable form of energy because of its unique properties and flexibility in how it can be utilised. Furthermore, the usage of recycled heat varies depending on several factors. Therefore, it would be reasonable to apply a correction factor to the resulting CHP cost function to reflect this. In this study, we assumed at least 60% usage of recycled heat is achievable. The cost function of the MT when it operates in heat and electricity generation mode with 80% overall CHP efficiency and 60% recycled heat energy usage is therefore approximately expressed as (5) The MT should satisfy the following constraint on generation capacity limits during any interval of operation. It is to be noted that the primary function of the MT is electricity generation. Therefore, heat production constraints are assumed to be indirectly taken care of by the power production constraint equations (6) 2.4 Energy storage The ESS in the case study microgrid system is vanadium redox battery (VRB) with a capacity of 800 kWh, and charging and discharging power rating of 150 kW. The state-of-charge (SOC) of the VRB at the end of the tth interval (start of the ) interval is given below: (7) where is the charging efficiency, is the charging power at the tth interval and is the energy storage capacity of the battery. VRB-based energy storage technologies are known to have high deep discharging capability and long cycle life compared to other battery technologies [19, 20]. Therefore, we have not considered a limitation on the number of daily charging/discharging cycles as a constraint in this work. To avoid deep discharge of the battery, the VRB must satisfy the following constraints on the rate of charge/discharge limits and storage capacity limits during any interval of the day (8) (9) 2.5 Utility grid Surplus power produced within the microgrid will be either sold to the main grid or stored in the battery based on the decision of the EMS. Likewise, the energy storage will discharge and/or the utility grid will inject electric power into the microgrid when there is more demand than the power produced in the microgrid (10) 3 Problem formulation For ease of implementation and fast computation, the MILP approach is adopted to solve the optimisation problem for the UC and resource scheduling task in this study. Given vectors , , and , matrices , and , and corresponding vectors and , and a set of indices intcon, MILP finds a vector that minimises the objective function . (11) The aim is to develop a 24 h DER unit commitment and dispatch scheduling scheme for the microgrid to minimise operational costs while meeting the local electricity demand. Therefore, the objective is to achieve the lowest cost of generation and minimise the cost of energy purchased from the utility. The cost minimisation objective function C in a microgrid consists of the cost of energy purchased from the utility as well as the cost of producing energy within the microgrid using conventional generation units. The optimisation problem can be expressed as (12) (13) (14) (15) (16) in (12) is an integer variable. This variable represents the status or commitment schedule of the dispatchable generation units in the microgrid. In our case, it is used to indicate the UC plan of the MT unit at any time interval. is 1 if the MT unit is running and 0 otherwise. Equation (13) describes the constraint related to minimum and maximum generation limits of the MT unit in the microgrid. Battery charging/discharging rate limits and allowable battery SOC range are handled by constraint (14) and (15), respectively, while (16) takes care of the power balance criterion. The outputs of the microgrid unit commitment and dispatch optimisation model are The status (commitment) schedule of the MT unit. Dispatch schedule of the MT unit, which is the only conventional/dispatchable power generation unit in the case study microgrid system. Schedule of hourly power transaction between the microgrid and the main grid for 24 h. Energy storage (VRB) charging and discharging schedules for 24 h. 4 Results and discussion Twenty-four hours ahead resource scheduling is considered in this study.Predicted hourly average load demand and the power generation expected from the2.5 MW WT and 480 kWp PV systems are shown in Fig. 2. The time interval is 1 h and, therefore, there are 24 intervals in aday. When the simulation starts, the SOC of the ESS is 20%. The hourly price of gridelectricity is shown in Fig. 3. Figure 2Open in figure viewerPowerPoint Load and generation prediction Figure 3Open in figure viewerPowerPoint Hourly price of electricity Simulations were performed for two separate cases. In the first case, it is assumed that the MT produces electricity only. In the second case, the MT supplies both heat and electricity and, hence, serves as a CHP generation unit. For simulation case 1, where there is no heat load demand, the optimal dispatch schedule of dispatchable units and overall power transactions in the studied grid-connected microgrid system for 24 h period is shown in Fig. 4. It shows the optimal usage of the energy storage, MT and utility supply for the simulation day. Positive values of grid power indicate the utility grid supplies power to the microgrid, while negative values indicate power export to the utility grid. Concerning the VRB, positive values of power correspond to discharging periods, whereas negative values correspond to charging periods. The SOC of the VRB at different time intervals of the simulation period for case 1 is shown in Fig. 5. It shows the amount of energy stored in the VRB at different intervals as a percentage of the nominal storage capacity. The MT remains OFF throughout the simulation period in case 1, suggesting its relatively low efficiency when it is used only for electricity production. It can be observed that the microgrid imports power from the utility grid during most of the intervals of the simulation day. Without making use of the VRB, the cost of operation in case 1 during the 24 h period is ¥2651.07. With usage of the VRB and the optimal scheduling scheme, the operational cost for the same period is reduced to ¥1911.5. Figure 4Open in figure viewerPowerPoint Resource scheduling in case 1 Figure 5Open in figure viewerPowerPoint SOC of VRB in case 1 The optimal dispatch schedule summary of the conventional generation unit (MT), utility grid and energy storage for case 2 is presented in Fig. 6. The SOC of the VRB at different time intervals is similarly shown in Fig. 7. It can be observed that the MT comes into operation (status '1') during the intervals where the electricity price of the utility grid gets higher. It remains OFF (status '0') when grid electricity price is low, suggesting that it is more economical to use grid electricity instead of producing using the MT by burning fuel. As in the previous case, the microgrid imports power from the grid during most of the intervals. With the optimal scheduling of the VRB and the CHP, the cost of operation for the given renewable energy generation and electricity consumption profiles is further reduced to ¥1426.8. The status (commitment schedule) of the MT unit at different time intervals for the two simulation cases during the 24 h simulation period is presented in Tables 1 and 2. Visual comparison of Fig. 3 and the values in Tables 1 and 2 indicates the MT operates during the intervals high grid electricity price to reduce electricity bills and remains off when grid prices go low to reduce fuel consumption. The simulation results also show that the energy storage charges during the intervals where electricity price is low, and discharges during periods of high price. Excess energy generated within the microgrid during low price periods could be used to charge the VRB and sold to the grid during high price periods. Therefore, optimal scheduling of the energy storage and MT enables the microgrid to minimise operational costs or maximise profit over the designated planning period of one day. Figure 6Open in figure viewerPowerPoint Resource scheduling in case 2 Figure 7Open in figure viewerPowerPoint SOC of VRB in case 2 Table 1. UC plan of MT for the first half of the simulation day Interval 1 2 3 4 5 6 7 8 9 10 11 12 case 1 0 0 0 0 0 0 0 0 0 0 0 0 case 2 0 0 0 0 0 0 0 0 1 1 1 1 Table 2. MT UC plan for the second half of the simulation day Interval 13 14 15 16 17 18 19 20 21 22 23 24 case 1 0 0 0 0 0 0 0 0 0 0 0 0 case 2 1 1 1 1 1 1 1 1 1 1 1 1 In Fig. 8, comparison of hourly operating cost of the microgrid with and without scheduling of the ESS for simulation case 1 is depicted. It can be observed that the overall operating cost for the day is improved through proper planning of charging and discharging intervals of the energy storage. The negative value of operating cost in a particular interval indicates that excess power is being exported to the utility grid during that interval, thereby resulting in revenue from energy sales. The corresponding curves for simulation case 2 are depicted in Fig. 9. Optimal scheduling of the MT and energy storage units in the microgrid results in an even more enhanced daily operating cost reduction. Figure 8Open in figure viewerPowerPoint Hourly operational cost of MG in simulation case 1 Figure 9Open in figure viewerPowerPoint Hourly operational cost of MG in simulation case 2 5 Conclusions The integration of more and more DERs, especially renewables, has resulted in the expansion of microgrid systems. The efficient operation of microgrids is realised through optimal unit commitment and dispatch scheduling of energy storage and other dispatchable units in the system. Stored energy is useful to backup dispatchable energy resources for balancing the intermittent nature of renewable energy generation units to achieve optimal operation. This paper implements MILP based one day ahead unit commitment planning and optimal dispatch scheduling model for an industrial microgrid system. Simulation results suggest that the operational cost of the microgrid can be effectively reduced through optimal scheduling of an ESS and a CHP unit using the proposed strategy. It takes load and generation prediction to optimally schedule the energy storage and CHP units. The developed strategy can be further improved to consider minimisation of emissions and integrate cost of battery charge/discharge cycles into the optimisation model. 6 References 1Olivares, D.E., Mehrizi-Sani, A., Etemadi, A.H. et al.: 'Trends in microgrid control', IEEE Trans. Smart Grid, 2014, 5, pp. 1905– 1919 2Celli, G., Pilo, F., Pisano, G. et al.: ' Meshed vs. radial MV distribution network in presence of large amount of DG'. IEEE PES Power Systems Conf. and Exposition, New York, USA, 2014 3Marnay, C., Rubio, F., Siddiqui, A.: ' Shape of the microgrid'. IEEE PES Winter Meeting, Columbus, Ohio, USA, 2001 4Zhang, Z., Li, G., Zhou, M.: ' Application of microgrid in distributed generation together with the benefit research'. IEEE PES General Meeting, Minnesota, USA, 2010 5Katiraei, F., Iravani, R., Hatziargyriou, N. et al.: 'Microgrids management', IEEE Power Energy Mag., 2009, 6, pp. 54– 65 6Parisio, A., Rikos, E., Glielmo, L.: 'A model predictive control approach to microgrid operation optimization', IEEE Trans. Control Syst. Technol., 2014, 22, pp. 1813– 1827 7Guan, X, Xu, Z, Jia, Q.S..: 'Energy-efficient buildings facilitated by microgrid', IEEE Trans. Smart Grid, 2010, 1, pp. 243– 252 8Bertsimas, D., Litvinov, E., Sun, X. et al.: 'Adaptive robust optimization for the security constrained unit commitment problem', IEEE Trans. Power Syst., 2013, 28, pp. 52– 63 9Kimura, K.: ' Neural networks approach for wind-solar energy system with complex networks'. IEEE 10th Int. Conf. on Power Electronics and Drive Systems (PEDS), Kitakyushu, Japan, 2013 10Aquino, R., Carvalho, M.: ' Recurrent neural networks solving a real large scale mid-term scheduling for power plants'. Int. Joint Conf. on Neural Networks (IJCNN), Barcelona, Spain, 2010 11Liu, X.P., Ding, M., Han, J.H. et al.: 'Dynamic economic dispatch for microgrids including battery energy storage', J. Energy Power Eng., 2011, 5, pp. 461– 465 12Xu, Q.Y., Zhang, N., Kang, C.Q. et al.: ' Day-ahead battery scheduling in microgrid considering wind power uncertainty using ordinal optimization'. North American Power Symp. (NAPS), Pullman, USA, 2014 13Rigo-Mariani, R., Sareni, B., Roboam, X.: ' A fast optimization strategy for power dispatching in a microgrid with storage'. 39th Annual Conf. of the IEEE Industrial Electronics Society, IECON, Vienna, Austria, 2013 14Juybari, E.Z., Hosseini, S.M.: 'Optimization of microgrid using quantum inspired evolutionary algorithm', I. J. Intell. Syst. Appl., 2014, 6, pp. 47– 53 15Murdoch, A., Barton, R.S., Winkelman, J.R. et al.: 'Control design and performance analysis of a 6 MW wind turbine generator', IEEE Trans. Power Appar. Syst., 1983, 102, pp. 1340– 1347 16Kassa, Y., Zhang, J.H., Zheng, D.H. et al.: ' A GA–BP hybrid algorithm based ANN model for wind power prediction'. Proc. of the 4th IEEE Int. Conf. on Smart Energy Grid Engineering, Oshawa, Canada, 2016 17Liang, R.H., Liao, J.H.: 'A fuzzy-optimization approach for generation scheduling with wind and solar energy systems', IEEE Trans. Power Syst., 2007, 22, pp. 1665– 1674 18Semero, Y.K., Zhang, J.H., Zheng, D.H.: 'PV power forecasting using an integrated GA-PSO-ANFIS approach and Gaussian process regression based feature selection strategy', CSEE J. Power Energy Syst., 2018, 4, (2), pp. 210– 218 19Aktar, P., Li, Y., Bao, J. et al.: 'Optimal charging of vanadium redox flow battery with time-varying input power', Batteries, 2019, 5, (20), pp. 1– 16 20Wei, Z., Bhattarai, A., Zou, C. et al.: 'Real-time monitoring of capacity loss for vanadium redox flow battery', J. Power Sources, 2018, 390, pp. 261– 269 Citing Literature Volume5, Issue1March 2020Pages 80-84 FiguresReferencesRelatedInformation