Stochastic scheduling of a renewable-based microgrid in the presence of electric vehicles using modified harmony search algorithm with control policies

https://doi.org/10.1016/j.scs.2020.102183Get rights and content

Highlights

  • Proposing an environmental/economic model for optimal energy management.

  • Considering renewable energy resources in the presence of Electric Vehicles

  • Presenting Modified Harmony Search Algorithm for optimizing Micro-grid

  • Reporting superior solutions in Short-Term Scheduling and Micro-grid Energy Management

Abstract

Modern power systems are seeking for alternative energy sources will less emission to conventional fossil-fuel power plant to mitigate the global concerns on environmental issues. On one hand, renewable energies have turned to the first choice of system operators. On the other hand, the system should provide the required infrastructure to appropriately accommodate such energy sources. In this respect, microgrids (MG) would provide the needed conditions for integrating renewable energy sources (RESs). Thus, the optimal energy management of such systems in the presence of highly uncertain renewable power generation is of great importance. Accordingly, this paper provides a stochastic programming framework for the optimal scheduling of an MG equipped with RESs and plug-in electric vehicles (PEVs). The power sources considered include a wind energy system in the form of wind turbine (WT), a solar photovoltaic (PV) system, a fuel cell (FC), a microturbine (MT), besides a battery storage system (BSS). The mentioned problem is formulated as a single-objective optimization problem aimed at minimizing the total operating cost over the scheduling period. The MG is considered in the grid-connected mode where it can transact power with the upstream system. The uncertainty of the problem is due to the intermittent power output of the wind energy system and the PV unit, as well as uncertain behavior of the EV owners in charging/discharging their vehicles. The proposed stochastic optimization problem is the solved using an effective and efficient optimization algorithm named “modified harmony search (MHS) algorithm”. Finally, the simulation results are discussed and the superior performance of the suggested algorithm is verified through making a comprehensive comparison with some well-known methods.

Introduction

The conventional power plans with fossil fuels were first used in electric power systems due to their reliable power generation and also the considerable size which could meet the system load demand requirements. However, such power plants are associated with dramatically high emission which raised the global concern. Renewable energies are taken into consideration appropriate alternatives to replace the existing fossil-fuel power plants (Izadbakhsh, Gandomkar, Rezvani, & Ahmadi, 2015; Rezvani, Gandomkar, Izadbakhsh, & Ahmadi, 2015). One possible solution to mitigate the environmental emissions and also power losses in transmission lines is to make the load demand and power generation close to each other. In this regard, distributed energy resources (DERs) have emerged in distribution systems (Ghaedi, Dehnavi, & Fotoohabadi, 2016). DERs are comprised of various energy sources, either renewable or non-renewable. Some well-known DERs are fuel cell (FC), wind turbine (WT), photovoltaic (PV) panels, and microturbines (MTs) (Mortazavi, Shiri, Javadi, & Dehnavi, 2015). The existing distribution systems are suffering from the lack of sufficient and necessary infrastructure to satisfy the requirements of modern power systems. To this end, a novel concept has been formed in electric networks named microgrid (MG). An MG is a low-voltage (LV) distribution network comprising DERs, storage systems and also active loads capable of participating in the scheduling. Furthermore, an MG can operate in different modes, either connected to the upstream system through the point of common coupling (PCC) or disconnected from the grid. These two operation states are called grid-connected and islanded operation modes. MGs would provide the electricity at a relatively higher reliability and lower emission compared to conventional electric systems with centralized structure (Shao, Gholamalizadeh, Boghosian, Askarian, & Liu, 2019). On the other hand, it should be noted that adding such systems to the existing electric power networks more complicates the control and operation systems. Many studies have been dedicated to developing effective and efficient models and methods for the MG operation. Since passing the Clean Air Act Amendments in 1990, electric power systems operators have been forced to take into consideration the environmental issues in the optimal operation of electric power systems (Shao et al., 2018). In this respect, genetic algorithm (GA) has been used in Ref. (Shao and Vesel (2015)) for the problem of optimal determination of the capacity of an MG with renewable energies through a multi-objective optimization framework aimed at simultaneously maximizing the penetration level of renewable energies and minimizing environmental issues. An optimization framework based on the multi-cross learning-based chaotic differential evolution method has been suggested in Ref. (Hemmati, Amjady, and Ehsan (2014))) for the optimal operation of an islanded MG while investigating the techno-economic and environmental issues. MGs are capable of taking part in the energy and reserve markets. In this regard, a stochastic programming optimization framework has been suggested in (Shi, Luo, & Tu, 2014) for a MG aimed at participating in the energy and reserve markets, while taking into account the uncertainties arisen from the load demand forecast error and the output of renewable energy sources (RESs). An operation index has been introduced in Ref. (Zhang, Gari, and Hmurcik (2014))) with respect to the electrical energy price, as well as the emission level and quality of service, which each has been assigned a weighting factor, for the optimal operation of an islanded MG. An intelligent energy management system has been developed in Ref. (Motevasel and Seifi (2014)) aimed at minimizing the total operating cost and emission for interconnected MGs including wind energy besides other DERs. Using the same objectives, an intelligent system has been proposed in (Motevasel, Seifi, & Niknam, 2013) for the day-ahead operation of an MG with combined heat and power (CHP) generation. A hybrid multi-objective optimization method has been developed in (Chaouachi, Kamel, Andoulsi, & Nagasaka, 2012) based on the artificial intelligence approaches and linear programming (LP) for the energy management of an MG. An MG scheduling model has been presented in (Aghaei & Alizadeh, 2013), where there are electrical and heating load demands, a storage system, and demand response (DR) programs. A bi-objective optimization model using the teaching-learning based algorithm has been suggested in (Niknam, Azizipanah-Abarghooee, & Narimani, 2012) for the optimal operation of an MG. A bi-level framework has been presented in (Buayai, Ongsakul, & Mithulananthan, 2012) with a sensitivity analysis and using the non-dominated sorting genetic algorithm II (NSGA-II) for the optimal operation of an MG.

In addition to the generation sector in the electric energy systems, the transportation systems also largely contribute to environmental emissions. In this regard, governments were looking for an appropriate alternative to replace conventional fossil-fuel vehicles. Electric vehicles (EVs) are the most suitable options to mitigate the stress in the transportation systems as they are associated with considerably less emission. Accordingly, EVs in the form of plug-in hybrid electric vehicles (PEVs) with the battery as the only source of energy on-board, plug-in hybrid electric vehicles (PHEVs) with battery and a combustion engine to provide the required energy and gridability option, and hybrid electric vehicles (HEVs) with a battery storage system and a combustion engine, without the option to connect to the grid, have largely penetrated into market. Among these three types of EVs, PEVs and PHEVs with the capability to connect to the electrical grid to charge/discharge are increasingly integrating with power systems. Thus, numerous research studied have been devoted to investigating the impact of such vehicles on electric power networks, particularly at the distribution level (Duan et al., 2017; Zhao, Liu, Liu, & Hao, 2019). Besides all good feature such EVs may bring the system, a significant penetration level of EVs would cause severe problems in power systems, especially regarding the economic and secure operation of power systems (Zhili, Boqiang, & Chunxu, 2019). Hence, it is required to provide a smart charging platform for EVs to mitigate the concerns on this issue. To this end, various algorithms have been so far presented for the effective charging/discharging behavior of electric vehicles (He, Venkatesh, & Guan, 2012). Re. (Sortomme & El-Sharkawi, 2010) designed a model using the specific regulation with price constraint and an arbitrary system for maximizing the profit of an EV aggregator equipped with a parking lot. An energy management system has been presented for an MG in (Honarmand, Zakariazadeh, & Jadid, 2014) while taking into consideration the intermittency of renewable power generation and reserve requirements of the system. In this respect, the parking lot makes a comparison between the energy price and the desired hourly price for charging/discharging prices delivered by the PEV owners. Consequently, the decision on the status of EVs is made. The impacts of market regulation on the players, particularly on PEV owners, have been investigated in Ref. (Shafie-khah, Moghaddam, Sheikh-El-Eslami, and Rahmani-Andebili (2012))). In this respect, the bilateral contract made between the PEV owners and aggregator has been also taken into consideration. PEVs and PHEVs are capable of connecting to the grid to charge, known as grid-to-vehicle (G2V) or discharge, which is known as vehicle-to-grid (V2G) capabilities. The V2G capability of PEVs has been discussed in Ref. (Tehrani, Shrestha, and Wang (2013))) using a price-based PEV charging/discharging aimed at maximizing the social welfare. Ref. (Soleymani, Ranjbar, and Shirani (2007))) studied the impacts of market restructuring on generation companies utilizing a game theory designed for participation of the players in the energy and reserve markets. Ref. (Fathabadi (2015)) has studied the role of EVs and RESs in electric networks. Ref. (Bahramara and Golpîra (2018)) developed a robust risk-constrained optimization framework for the operation of MGs equipped with PHEVs aimed at minimizing the total operating cost. Ref. (Tabatabaee, Mortazavi, and Niknam (2017))) has studied the optimal day-ahead scheduling of MGs in the presence of PHEVs and RESs using a single-objective optimization framework solved using the bat optimization algorithm (BOA). Ref. (Aluisio, Conserva, Dicorato, Forte, and Trovato (2017))) presented a non-linear programming framework for the day-ahead scheduling of MGs with PEVs taking into account the V2G capability.

The present study discussed the optimal day-ahead scheduling of MGs equipped with PEVs and DERs, including different DERs, such as wind energy and solar energy, an MT, as well as FC. PEVs have been considered in this paper due to their potential and also, their large penetration into market. Applying the impact of such vehicles to the scheduling model of Mgs would make the current mode very complex. Thus, an efficient and effective optimization method is needed to solve the problem. The proposed problem is formulated as a single-objective optimization algorithm aimed at minimizing the total operating cost of the system. The performance of the suggested method and its superiority over other techniques has been shown by simulating the model on a case study in the grid-connected mode and comparing the results derived to those obtained by other methods. Overall, the main contributions of this paper can be briefly stated as follows:

  • Presenting an optimal day-ahead scheduling model for an MG with different types of RESs as well as PEVs.

  • Proposing an effective and efficient optimization algorithm named modified harmony search (MHS) algorithm for solving the developed problem.

  • Taking into account the cost of energy not served, and energy transaction with the main grid as well as V2G and G2V capabilities of PEVs in the proposed model.

The remainder of the paper has been organized as follows: Section 2 presents the model of different assets available in the system. The uncertainty characterization method, named “point estimate method (PEM)” has been described in section 3, while the MHS algorithm is presented in section 4. Section 5 includes the results obtained from solving the problem using the MHS algorithm, while the relevant conclusions have been represented in the last section.

Section snippets

System component characteristics

The proposed problem is formulated as a single-objective stochastic optimization problem with different types of DERs, such as WT, PV, MT, and FC. The detailed models of the assets have been presented in Ref. (Izadbakhsh et al. (2015); Rezvani, Gandomkar et al. (2015)). A conceptual representation of an MG with different DERs and PEVs has been given in Fig. 1.

Problem formulation

The conceptual configuration of an MG is shown in Fig. 2. As it has been previously stated, the proposed problem, is formulated as a single-objective optimization problem aimed at minimizing the total operating cost. The objective function and the related constraints are represented in this section (Tabatabaee, Mortazavi, and Niknam (2017)).MinCost=t=1T(k=1NDGCDG,ktPDG,kt+CGridtPGridt+CENSt×i=1NCusLαitUit+(n=1NνUνtCνtPνt+Costdegt))

The total operating cost of the MG shown in (5) is

Harmony search algorithm

The harmony search (HS) algorithm has been first used in 2001 by inspired by the musicians to enhance the harmony (Geem, Kim, & Loganathan, 2001; Tabatabaee et al., 2016). This method is categorized as evolutionary algorithms and its basic idea is based on finding the harmony, i.e. the optimal solution, using the values (notes) generated by the musicians (control variables). This method is considerably incomplex compared to other algorithms with limited parameters to be set and potential

Simulation results

The proposed scheduling framework is simulated on a 12.66 kV test system with 32 buses having power transaction with the main grid through a transformer to assess its performance. The single-line diagram of the system has been demonstrated in Fig. 1, comprising two WTs, two MTs, one FC as well as one PV and two fleets of gridable EVs. Buses 10 and 14 have accommodated the two WTs, while buses 19 and 25 are the locations of the PV and FC units, respectively. The MT units have been installed at

Conclusion

This paper proposed a stochastic programming framework for the day-ahead scheduling of a microgrid (MG) operating in the grid-connected mode. The Mg was equipped with different generation technologies such as Wind turbines (WTs), photovoltaic (PV) unit, microturbines (MTs), and a fuel cell (FC). Besides, two fleets of plug-in electric vehicles (PEVs) were taken into consideration and their related uncertainties were addressed using the unscented transform. In this regard, three different case

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

This work is supported by the National Natural Science Foundation of China (Nos. 61702075), the Young Elite Scientists Sponsorship Program by CAST (No. 2018QNRC001).

This research was funded by the Deanship of Scientific Research at Princess Nourah bint Abdulrahman University through the Fast-track Research Funding Program.

References (38)

Cited by (72)

View all citing articles on Scopus
View full text