Operation window constrained strategic energy management of microgrid with electric vehicle and distributed resources
2016; Institution of Engineering and Technology; Volume: 11; Issue: 3 Linguagem: Inglês
10.1049/iet-gtd.2016.0654
ISSN1751-8695
AutoresLokesh Kumar Panwar, Srikanth Reddy Konda, Ashu Verma, Bijaya Ketan Panigrahi, Rajesh Kumar,
Tópico(s)Electric Vehicles and Infrastructure
ResumoIET Generation, Transmission & DistributionVolume 11, Issue 3 p. 615-626 ArticleFree Access Operation window constrained strategic energy management of microgrid with electric vehicle and distributed resources Lokesh Kumar Panwar, Lokesh Kumar Panwar Center for Energy Studies, IIT Delhi, Delhi, IndiaSearch for more papers by this authorSrikanth Reddy Konda, Corresponding Author Srikanth Reddy Konda srikanthreddy@ee.iitd.ac.in Department of Electrical Engineering, IIT Delhi, Delhi, IndiaSearch for more papers by this authorAshu Verma, Ashu Verma Center for Energy Studies, IIT Delhi, Delhi, IndiaSearch for more papers by this authorBijaya Ketan Panigrahi, Bijaya Ketan Panigrahi Department of Electrical Engineering, IIT Delhi, Delhi, IndiaSearch for more papers by this authorRajesh Kumar, Rajesh Kumar Department of Electrical Engineering, MNIT, Jaipur, IndiaSearch for more papers by this author Lokesh Kumar Panwar, Lokesh Kumar Panwar Center for Energy Studies, IIT Delhi, Delhi, IndiaSearch for more papers by this authorSrikanth Reddy Konda, Corresponding Author Srikanth Reddy Konda srikanthreddy@ee.iitd.ac.in Department of Electrical Engineering, IIT Delhi, Delhi, IndiaSearch for more papers by this authorAshu Verma, Ashu Verma Center for Energy Studies, IIT Delhi, Delhi, IndiaSearch for more papers by this authorBijaya Ketan Panigrahi, Bijaya Ketan Panigrahi Department of Electrical Engineering, IIT Delhi, Delhi, IndiaSearch for more papers by this authorRajesh Kumar, Rajesh Kumar Department of Electrical Engineering, MNIT, Jaipur, IndiaSearch for more papers by this author First published: 01 February 2017 https://doi.org/10.1049/iet-gtd.2016.0654Citations: 17AboutSectionsPDF 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 In this study, an operation window constrained strategic energy management (OWCSEM) for microgrid operational scheduling is examined. The scheduling problem is formulated as cost minimisation/benefit maximisation of distribution resources. The scheduling operation of distributed resources is evaluated for various perspectives such as customer driven, micro grid operator (MGO) driven and MGO driven with utility constraints. The MGO driven scheduling operation considers the distribution network loss minimisation in addition to economic benefits of distributed resources. The loss estimation of distributed network/microgrid is performed using forward-backward sweep algorithm, considering the ill-natured radial network. Thereafter, the optimisation is performed using modified gradient-based search method, considering the linear nature of the problem. The simulation results are presented and discussed for the three scheduling scenarios proposed under operation window constraints. The customer driven scheduling of distributed resources resulted in highest profit among all the scheduling alternatives. On the other hand, the MGO driven scheduling horizon resulted in lower line loss cost component at a compromise of distribution resource benefit. The utility constrained MGO driven scheduling objective attained a compromise solution in terms of loss component and customer benefit. Thus, it is identified as best practice for overall welfare maximisation of microgrid/distribution network. Nomenclature line flow between i, j buses and associated power flow limit active and reactive power demands at time t upper and lower voltage tolerance limits for the i th bus power and energy statuses of EV during time slot t lower and upper state of change (SOC) levels of EV fleet upper bounds of power and energy of EV voltages of EV and URFC for time slot t power and energy statuses of URFC for time slot t lower and upper reactive power bounds of PCC (bus/node 1) lower and upper active power bounds of PCC (bus/node 1) Pi, Qi active, reactive powers of the i th bus Vi, Vj voltages at bus i and j, respectively Ij current flowing from bus j R, X, Z line resistance, reactance and impedance between buses i and j ηE, ηFC fuel cell efficiencies in electrolysis (E) and fuel cell (FC) modes, respectively line loss and differential price tariff at time slot t ηch, ηdis charging and discharging efficiencies of EV battery power status of the m th EV at the t th time slot power shift of the k th DR at the t th time slot minimum and maximum DR schedules for time slot t availability of the m th EV at the t th hour 1 Introduction Global warming and climate change are major forces that are driving the use of sustainable practices such as cleaner electricity generation through renewable, distributed generation (DG) and greener ways of transportation through electric vehicles (EV) [1]. Thus, renewable energy technologies (RETs), DG and EVs provide cleaner practices of electricity and transportation sectors [2]. Grid wide deployment of RETs is achieved using economically efficient practices, fiscal policies and technology breakthroughs [3]. The successful deployment of EVs has been going through a lean pace due to immature storage technologies as well as lack of awareness amongst public community. However, curbing of emissions from conventional/fossil fuelled vehicles seems imperative due to their considerable share (13%) toward total global emissions [4]. Many fiscal policies were introduced to encourage consumer for shifting to EV from conventional vehicles to achieve comparable share of EVs on roads. For example, US had an ambitious target of bringing 1 million EVs on road by 2015 [5]. Similarly, there is an improvement in the public awareness in countries such as UK in which 80% of car owners expressed willingness in replacing conventional vehicle by EV and 75% of the same survey population agreed to do so by 2015 [6, 7]. The motivation for this is associated with various reasons of deploying EV such as environmental concerns, cost savings, social privilege etc. Therefore, effective management of charging and discharging operations in electric network is imperative where high penetration of EVs is expected in near future. The EV storage capability confronts various challenges in the operation of distribution system. The charging loads and supply capacities of EVs depend on aspects such as user behaviour and usage diversity in distribution system and have to be estimates for examining their effects on microgrid operational aspects [8]. Thus, the anticipated effects such as voltage regulation, capacity congestion can be effectively addressed using intelligent scheduling of EV fleets [9]. Also, the application of EV fleets is then extended to energy management based on stochastic model for energy loss minimisation through optimal charging of plugin hybrid electric vehicles (PHEVs) [10]. In the same application, PHEVs are used to compensate reactive power to improve energy economics of distribution network. The mobile storage devices/PHEVs were also used to alleviate intermittency of renewable energy technologies through dual objective, i.e. cost as well as loss minimisation [11]. The distributed energy storage devices such as fuel cell (FC) along with DG and mobile energy storage/EV are utilised intelligently for multi-objective optimisation, minimising losses yet maximising the economic value of distributed resources [12]. The uncertain nature in PHEV arrival and departure time zones, electric price signals are considered through fuzzy reasoning system to optimise microgrid operation while satisfying voltage and load constraints [13]. Therefore, coordinated energy management of PHEV and RET underlines the optimal operation of distributed network/microgrid. Similar to the distributed energy resources, the benefits of demand response (DR) entities range from reduced network congestion, improved sustainability, efficient energy utilisation etc. One of the principal objective/benefit of DR is to flatten system demand over the day [14]. The energy management of DR prosumers (customers acting as DR producers) attracted much attractions in recent years [15]. The operational management of DR is evaluated as utility-oriented practice in microgrid systems with AC/DC facilities [16]. The power flow management under multi-constrained environment is performed using DR preferred depth of control and contractual prioritisation of contractual demand by distribution network operator [17]. Recent research initiatives include the multi-seller and multi-buyer schemes in distribution grid using dual decomposition technique [18]. Allocation of real-time costs for fair payment of DR is also modelled for distribution network energy management [19]. The management schemes for DR may affect the network operating constraints and limiting/expanding the opportunities for charge–discharge operations of EV fleets. Therefore, examining the effect of DR operations on energy management of EVs in microgrid comprises a key issue in distribution network/microgrid. This paper discusses the window constrained charge–discharge operation of PHEVs under the presence of various distributed entities: namely, DG, DR and distributed storage. The various objectives that can be considered for PHEV coordination in distribution microgrid are developed. In addition, variation of security window and charge–discharge schedules under different scheduling objectives is also examined. Rest of this paper is organised as follows: system description with operating constraints of the distributed resources and network elements are introduced in Section 2. Thereafter, Section 3 details about the formulation of various scheduling objectives/scenarios for strategies energy management of microgrid. The solution methodology for proposed scheduling scenarios is elaborated in Section 4. The test system description and simulation results are presented and discussed in Section 5. The key outcomes of the study are summarised and concluded in Section 6. 2 System description The single line diagram of test system considered in this paper with photovoltaic (PV) generation, unitised regenerative FC (URFC) storage, DR aggregator and EVs is presented in Fig. 1. The impedances of the lines shown correspond to base voltage of 13.5 kV and base power of 5 MVA. The solar generation and load demand under base case is shown in Fig. 2 and is obtained from [20]. The system load is equally distributed among all the nodes over the complete scheduling horizon. The network constraints considered during microgrid scheduling are voltage constraints, power balance constraint, transformer capacity constraint etc. Considering the ill nature of distributed networks, rather than Newton–Raphson approaches, simple and computationally efficient forward/backward seep algorithms are used for load flow analysis [21]. Therefore, the forward/backward sweep algorithms are used in conjunction with various constraints associated with DR, EV, URFC explained as follows. Fig. 1Open in figure viewerPowerPoint Microgrid line diagram with distributed energy resources and distributed energy storage Fig. 2Open in figure viewerPowerPoint Flowchart for OWCSEM 2.1 Constraints 2.1.1 Power line constraint The line flow across any two buses i and j of the network is bounded by maximum line carrying capacity of the line (1) 2.1.2 Load constraints The loads distributed amongst various buses in the network operate under the voltage limits, active and reactive power limits associated with that bus as are given as follows (2) 2.1.3 SPV plant constraints Similar to the conventional generator, the PV generator works in accordance with the equality and inequality limits (3) 2.1.4 EV battery constraints This paper considers two EV buses equipped with, lithium iron phosphate (LiFePO4) battery operated under charging and discharging modes from and to grid, respectively. The operation of battery pack at any time instant is constricted by SOC limits, bus voltage, energy and power limits are given as follows (4) The upper and lower limits of SOC are set to 1 and 0, respectively, considering the superior life cycle of LiFePO4. 2.1.5 URFC constraints The URFC in this paper operates under charging and discharging modes, which are designated as electrolysis mode and FC-mode, respectively (5) 2.1.6 Point of common coupling (PCC) constraints The constraints associated with the PCC between microgrid network and utility grid are given as follows (6) The load flow equations at the receiving end are given by (7) (8) The voltage of receiving end can be obtained by rearranging (7) and (8) and is given by (9) (10) (11) In (11), and can be deduced from forward/backward sweep algorithms for radial networks. 2.2 URFC system In the microgrid under study, URFC acts as a distributed energy storage element. The URFC operations under E-mode and FC-mode are represented by following equation/reaction (12) where ΔH denotes the enthalpy of chemical reaction (12) which is equal to ±237.2 kJ/mol. The operation of URFC in E-mode energises/recharges the FC with hydrogen. The amount of hydrogen generated/produced can be expressed as (13) where (14) In (14), represents the URFC system efficiency and is equal toηE, ηFC in E-Mode and FC-Mode, respectively. The highest range of efficiencies achieved in E-Mode and FC-Mode are 93 and 53%, respectively [22]. 2.3 EV system In this paper, distributed and mobile storage is considered through pure EV in the form of electric bus [23]. The LiFePO4 battery packs offer a long life of 3500 cycles at 80% capacity retention when discharged with a maximum of 100% depth of discharge [24]. Unlike the Li–ion battery, whose SOC operating window is constricted from 20 to 80% due to voltage stability issues, the LiFePO4 provides deep discharge provision (15) The operational efficiency of the EV battery-grid interfacing system is considered as 90% for both charging and discharging modes. The E-bus considered in this paper is used for transportation of employees inside the industrial sector/bus. Each E-bus operates for three trips per day with an average of 40 km per trip. The average mileage of the commercial E-bus technologies available is considered in this paper, which is around 0.7648 km/kWh [23]. Therefore, the aggregated energy consumption over a trip is ∼50 kWh. 2.4 DR system The DR services can be intelligently used to alleviate the peak power requirement. This can be carried out by shifting the responsive loads from peak (high tariff) to off peak period (low tariff). In this system, active responsive loads are considered for DR operation by microgrid operator (MGO). The total energy consumption of responsive load over the day is constrained by stipulated energy demand. The maximum shiftable demand over any time step is constrained by (16) (17) The total shiftable demand over the day () should be less than or equal to the maximum shiftable DR over a day () and can be ensured using the following constraint (18) 3 Problem formulation The operational scheduling objective of the microgrid can have various perspectives. The objectives of various entities/stakeholders of microgrid are considered in this paper. The implications of differential tariff in microgrid structure motivate different distributed resources to respond and make informed decisions to reduce the overall daily cost of operation to reduce energy bills. 3.1 Scenario 1: customer benefit maximisation: customer driven The first scenario considers operational cost minimisation of microgrid entities such as DR, EV and URFC as given by (19) where Nev, NURFC and NDR are number of EV's, URFC's and DR aggregators in microgrid structure, respectively. represents the availability of the m th EV for the t th time slot as given by (20) (21) The grid to vehicle (G2V), vehicle to grid (V2G) and transit time energy contents of EV are given as follows (22) Similarly, the operating conditions and energy content of URFC at the t th time slot are given by (21) and (22), respectively (23) (24) It can be observed from (21) and (23) that for discharge events EV/URFC are subjected to negative costs which can be seen as revenue generation. It is assumed that, during discharge events, the price equivalent to differential tariff will be paid to respective discharging entities. 3.2 Scenario 2: customer benefit maximisation and loss minimisation: MGO driven The MGO driven scheduling considers the economically efficient operation of microgrid elements as well as total power loss of the microgrid structure. Thus, the total scheduling objective can be deduced as follows (25) where Nb represents the number of buses in the system and i, j denote bus indices. 3.3 Scenario 3: customer benefit maximisation, loss minimisation and load levelling: MGO driven In this scenario, the security constraints for efficient operation of system are considered through uniform loading accompanied by load levelling objective as follows (26) 4 Solution methodology The flowchart of the operational window constrained microgrid operation for various cases is explained in Fig. 2. The grid integration of EV, i.e. G2V and V2G operations is constrained by transit time of the EV. However, for remaining time slots excluding the transit times, EVs are not obligated to undergo G2V/V2G operation. Thus, depending on the scheduling objective and the tariff structure, particular EV maybe used to choose ideal for a particular time slot. The efficiencies of charging and discharging modes are considered for estimating the energy calculations during grid iterative time slots of EV. Since, the EVs considered in this network are E-bus fleet, no coordinated effort is considered across all the objectives. Therefore, the charge–discharge events of EV1 and EV2 are independent practices as does their transit times. The operation of DG, distributed storage in this paper accompanies optimisation problem of different objectives. However, the scheduling operation of various elements is constricted to operational constraints of different network elements such as distribution line capacity, transformer rating etc. The complexity associated with the scheduling objectives increases with the dimension of the problem which is 24 h rather than a single time slot. The forward–backward sweep algorithm is applied to calculate the microgrid losses considering the network topology. The voltage profile is assumed to be flat given the root/source node. The principal procedure followed in carrying out the forward/backward sweep algorithms is explained in brief as follows [21]: Calculation of node currents: To begin with, every node of the network is sorted by its distance from farthest node, which is designated as the n th node. For the k th iteration, nodal current of the i th node is given by (27) where Si denotes the power injection at the i th node, represents the voltage of the i th node for the iteration and Yi denotes the sum of shunt element admittances connected to node i. Backward sweep: Moving from farthest to root node, the nodal current balance equations are used to estimate branch currents in every iteration as follows (28) where is the current injected at node i. Forward sweep: Forward sweep starting form root node to farthest node is run to calculate nodal voltages. Through Kirchhoff's voltage law and using current (calculated in step 2) flowing from the i − 1th node to the i th node, i.e. the nodal voltages can be calculated as (29) where Zi −1, i is the line impedance between and i th nodes. Steps 1–3 are repeated until convergence criteria is satisfied. The voltage and nodal current estimation using forward–backward sweep algorithms is followed by loss estimation. The estimated losses are then supplied to appropriate objective fitness for multiplying with tariff and subsequently included in fitness function. The iterative procedure is started then to optimise the microgrid operation considering operation window constraints as explained by Fig. 2. The iterative process is repeated or terminated depending upon the solution quality and termination criteria as explained in (Figs. 3–5). Fig. 3Open in figure viewerPowerPoint Transformer operation limit check Fig. 4Open in figure viewerPowerPoint Voltage constraint check Fig. 5Open in figure viewerPowerPoint Modified gradient method 5 Simulation results and discussion The simulation trails are carried out using MATLAB platform on a 32 bit Pentium dual core 2.3 GHz processor working on Linux operating system. The charging window of the EV, URFC and shifting window of DR entities are constrained by active, reactive flows of the microgrid/distribution system. The samples of weekly active power demand, reactive power flows with 10% deviation are considered over a year. The maximum and minimum values of active and reactive loads of the network are shown in Figs. 6a and b, respectively. The average weekly values of maximum and minimum solar insolations are presented in Fig. 6c. The microgrid operation is examined for differential tariff scheme whose variation (assumed to be 10% on average) over a week is presented in Fig. 6d. The PV capacity considered in this paper comply with a maximum capacity of 0.5 MWA. The availability of EVs is defined as a function of EV transit time. The EV transits are assumed to be same for all the days of weeks considering the industrial microgrid structure. For EV1, 3rd, 11th and 19th hours of the day mark the start of transit times, whereas EV2 transit times start at 5th, 13th and 21st hours, respectively, travelling for the same distance and consuming same energy as that of EV1. The charge–discharge operations are considered to be occurring at an efficiency of 90%. Fig. 6Open in figure viewerPowerPoint Average weekly variation of a Active power demand limits b Reactive power demand limits c Solar insolation limits d Differential tariff limits The operational window is constrained for scheduling decisions of various microgrid entities. The same is deduced by considering the line flow limits of PCC, transformer operation limits, voltage tolerance limits etc. The operation window for DR, URFC and EVs is presented in Fig. 7. It can be observed that the operation window limits are different for different entities. Also, the limits depend on type of device and operation mode. For example, the limits for charge and discharge events of URFC and EV are considerably different. The charge operation of URFC and EV is more constricted compared with discharge limit due to the microgrid/system welfare aspects of storage acting as DG (discharge mode). The simulation results and discussion for three operational scenarios considered are presented and discussed in the following sections. Fig. 7Open in figure viewerPowerPoint Operation window constraints for a URFC b EV1 c EV2 d DR 5.1 Case I: microgrid operation strategy with UFRC In this scenario, all the microgrid entities such as DR, EV and URFC are scheduled with overall cost minimisation subjected to differential tariff. Therefore, the cost–benefits of the distributed resources of the microgrid would be maximum for this case. The total saving/benefit over 24 × 7 scheduling horizon is 109030.5 Indian rupees (INR) with ±13.11% variation. The average daily benefit for distributed resources (DR, EVs and URFC) is 15575.7 INR. The total benefit is the sum of DR benefit of 107606.37 INR for load shifting from peak to off peak hours, URFC benefit of 1135.7 INR for off peak charging and peak discharge operations and EV benefit of 288.44 INR for G2V and V2G operations of EV1 (net benefit of 106.28 INR) and EV2 (net benefit of 182.15 INR). The charge–discharge events of storage elements and load shift events of the DR in this scenario can be solely attributed to the differential tariff with peak and non-peak prices during day and night, respectively. The axis of energy for URFC, EV is advanced by an hour compared with that of power axis in Fig. 8. This is to observe the variation of energy increment/decrement during charge–discharge events. The total number of grid interactive hours of EV1 and EV2 are same and is equal to 46 h over total scheduling horizon. The power drawn and discharged to grid by EVs during G2V and V2G operations may not be same considering the energy usage by EV for transit purpose. It can be observed from Figs. 8b and c that the EV charge duration and power are comparatively higher than that of the discharge event. Therefore, energy fed to grid by EVs is comparatively lower than that of URFC. Thus, the cost–benefit of EV (288.44 INR) is considerably lower compared with URFC (1135.7 INR). Fig. 8Open in figure viewerPowerPoint Schedule for Scenario 1 a Power drawn/supplied by URFC b Charging status of EV1 c Charging status of EV2 d Power schedule of DR e Energy contained by URFC f Energy status of EV1 g Energy status of EV2 It can also be observed that the charging and discharging events of URFC are not same with power during charge events higher compared with that of discharge events. This can be attributed to the difference between the efficiencies of E-mode (charge) and FC-mode (discharge) of URFC with latter being lower [22]. However, the total power drawn by DR entity during a day is equal to zero which accounts for load retention in off peak hours and curtailments in peak price hours. Thus, the net profit of DR gained over a day is positive yet drawing the same amount of power. 5.2 Scenario 2: customer benefit maximisation and loss minimisation: MGO driven In this scenario, along with customer benefit maximisation, MGO scheduling objective, i.e. network loss minimisation is also considered. The total benefit of the distributed resources in this scenario is 76130.3 INR with ±6.23 variation caused due to load variation. Therefore, the average daily profit of distribution resources reduced to 10875.76 INR from 15575.7 INR/day. The total benefit in this scenario resulted from DR is a net profit. While, the net cost applicable for EV and URFC are positive. Therefore, the inclusion of system losses in the objective function resulted in reduced economic benefits to distributed resources. The charge–discharge events of the EV and URFC are more distributed over the day due to the loss reduction objective. The more uniform charge–discharge profile helped to alleviate stress on the system at a cost of economic benefits of the microgrid entities. The net cost applicable for EV1 and EV2 over the scheduling horizon is 3292.16 and 3079.63 INR, respectively. Similarly, the net cost experienced by URFC over the scheduling horizon is 225.1 INR. The loss component of total system cost reduced from 23561.33 INR (Scenario 1) to 23482.16 INR (Scenario 2). The total amount of losses remained the same as that of Scenario 1. This can be attributed to the fact that the total load of the system remains same over scheduling horizon. However, the reduction in loss component of the objective function is due to the combined effect of change in time slots of charge–discharge events and load shifting events and differential tariff. The power drawn by EVs and URFC over the total scheduling horizon is same for Scenario 1 and Scenario 2. Similarly, the amount of energy discharge from EV and URFC is constant over both scenarios. However, the hourly power schedules and time instants have changed from Scenario 1 to Scenario 2. It can be observed that, EV charge–discharge events are more evenly distributed in Scenario 2 (Figs. 9b and c) compared with Scenario 1 (Figs. 8b and c), where EV charge events are sparsely distributed over the day. The total number of charge–discharge events of EV1 and EV2 are 133 and 132, respectively. Compared with Scenario 1, the total number of grid interactive hours of EV1 and EV2 are increased by 87 and 86, respectively. This can be attributed to the effect of loss minimisation objective which inhibits bulk charging in the same slot. Similarly, the total number of charge–discharge event hours of URFC has increased to 154 (Scenario 2) from 44 (Scenario 1). Therefore, the inclusion of loss minimisation resulted in uniform charge–discharge schedule of EV and URFC at a cost of economic benefit. Unlike EV and URFC, DR experienced slight reduction of 182 MW over the total scheduling time horizon. However, the total number of response events of DR increased from 63 to 167 which is one time slot lesser than total scheduling time horizon. Fig. 9Open in figure viewerPowerPoint Schedule for Scenario 2 a Power drawn/supplied by URFC b Charging status of EV1 c Charging status of EV2 d Power schedule of DR e Energy contained by URFC f Energy status of EV1 g Energy status of EV2 5.3 Scenario 3: customer benefit maximisation, loss minimisation and load levelling: MGO driven and utility constrained In this scenario, the utility perspective of scheduling the distributed system/microgrid is considered through load levelling objective. The total benefit/profit of the distributed resources in this scenario is 88755.46 INR ±9.1%. The cost–benefits of the DR and URFC are 91281 and 670.89 INR, respectively, whereas the net cost of charge–discharge events for EV1 and EV2 are 1576.66 and 1619.8 INR, respectively. It can be observed that, benefits of DR, URFC and costs of EV1 and EV2 are improved compared with Scenario 2. However, the net benefit for all the distributed resources (EVs, DR and URFC) are comparatively lower than Scenario 1. The charge–discharge events of EV, URFC and load shift events of DR for scenario 3 are presented in Fig. 10. Fig. 10Open in figure viewerPowerPoint Schedule for Scenario 3 a Power drawn/supplied by URFC b Charging status of EV1 c Charging status of EV2 d Power schedule of DR e Energy contained by URFC f Energy status of EV1 g Energy status of EV2 The total number of grid interactive hours of EV1 and EV2 are 107 and 102, respectively. The same are lowered, respectively, by 19.5 and 22.72% compared with Scenario 2. Similarly, charge–discharge events of URFC have reduced from 154 to 99 compared with their counterparts in Scenario 3. Also, the amount of load shift is reduced from 7000 MW (Scenario 2) to 6990 MW (Scenario 3). Therefore, Scenario 3 acts as a buffer solution between Scenario 1 (highest consumer benefit) and Scenario 2 (least consumer benefit). The cost–benefit of three scenarios for total scheduling time horizon is given in Fig. 11. The values in the negative side denote payment made by MGO/utility to microgrid entities: namely, DR, EV and URFC for grid interaction via charge–discharge events or load shift events. It can be observed that the customer driven scheduling scenario attained highest profit compared with any other scenario. The positive values denote the amount paid by distributed resources to MGO/utility for power drawn from grid. The MGO driven objectives (Scenario 2 and Scenario 3 resulted in lowered payments both to and fro between distributed resources and grid. Fig. 11Open in figure viewerPowerPoint Cost–benefit distribution for different scheduling scenarios 6 Conclusions In this paper, a window constrained operational scheduling is proposed for microgrid with distributed resources DR, EV and URFC. The proposed approach is simulated for an 11 bus distribution/microgrid network under various constraints such as load balance, line flow limits, charge–discharge constraints of EV/URFC, load shifting limits of DR etc. The problem is formulated as cost–benefit maximisation objective for charge–discharge schedules and load shifting events as variables of optimisation process. In addition, various operating scenarios considering customer, MGO and utility perspectives. The optimisation problem along with load flow operation for line loss estimation is solved using modified classical gradient descent approach along with forward–backward sweep algorithm. The simulation results are presented for discussion, which reveal the maximum profit under customer driven scheduling scenario compared with other two scenarios. The lowest loss component of cost is resulted for the objective with consideration of customer benefit as well as MGO objectives. Also, MGO driven scheduling scenario witnessed highest number of charge–discharge events for EV and URFC along with highest shift in DR. The scenario with MGO driven objective with utility welfare constraints performed comparatively between customer driven schedule and MGO driven scheduling scenarios. Therefore, consideration of customer, MGO and utility objectives can maximise the overall welfare of distribution network/microgrid. 7 Acknowledgment This study was funded in part by the DST Sponsored Project “Control of grid interfaced solar photovoltaic energy with networked electric vehicles to enable vehicle a smart grid (project no: RP03200)”. 8 References 1Farhangi, H.: ‘The path of the smart grid’, IEEE Power Energy Mag., 2010, 8, (1), pp. 18– 28 (doi: 10.1109/MPE.2009.934876) 2‘ Multi-Year Program Plan 2011–2015 Vehicle Technologies Program’, Energy efficiency and Renewable Energy, U.S Department of Energy. Available at http://www.energy.gov/eere/vehicles/downloads/vehicle-technologiesofficemulti-year- program-plan-2011-2015 3‘ Guide to purchasing green power renewable electricity, renewable energy certificates, and on-site renewable generation’. Available at http://www.epa.gov/greenpower/documents/purchasingguideforweb.pdf 4Metz, B., Davidson, O.R., Bosch, P.R. et al. (EDs): ‘ Climate change 2007: mitigation of climate change. Contribution of working group III to the fourth assessment report of the intergovernmental panel on climate change (IPCC)’ ( Cambridge University Press, Cambridge, UK and New York, NY, USA, 2007) 5One Million Electric Vehicles By 2015, Status Report, Department of Energy, US, February 2011. Available at https://www1.eere.energy.gov/vehiclesandfuels/pdfs/1_million_electric_vehicles_rpt.pdf 6Investing in ultralow emission vehicles in the UK, 2015 to 2020. Available at https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/307019/ulev-2015-2020.pdf 7Public Attitudes to Electric Vehicles, statistical release, Department of Transport, 19th June 2014. Available at https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/321157/electric-vehicles-2014.pdf 8Shaaban, M.F., Atwa, Y.M., El-Saadany, E.F.: ‘PEVs modeling and impacts mitigation in distribution networks’, IEEE Trans. Power Syst., 2013, 28, (2), pp. 1122– 1131 (doi: 10.1109/TPWRS.2012.2212467) 9Azzouz, M.A., Shaaban, M.F., El-Saadany, E.F.: ‘Real-time optimal voltage regulation for distribution networks incorporating high penetration of PEVs’, IEEE Trans. Power Syst., 2015, 30, (6), pp. 3234– 3245 (doi: 10.1109/TPWRS.2014.2385834) 10Nafisi, H., Agah, S.M.M., Askarian Abyaneh, H. et al.: ‘Two-stage optimization method for energy loss minimization in microgrid based on smart power management scheme of PHEVs’, IEEE Trans. Smart Grid, 2015, PP, (99), pp. 1– 9 11Su, W., Wang, J., Roh, J.: ‘Stochastic energy scheduling in microgrids with intermittent renewable energy resources’, IEEE Trans. Smart Grid, 2014, 5, (4), pp. 1876– 1883 (doi: 10.1109/TSG.2013.2280645) 12Panwar, L.K., Srikanth Reddy, K., Kumar, R. et al.: ‘Strategic energy management (SEM) in a micro grid with modern grid interactive electric vehicle’, Energy Convers. Manage., 2015, 106, pp. 41– 52 (doi: 10.1016/j.enconman.2015.09.019) 13Masoum, A.S., Deilami, S., Abu-Siada, A. et al.: ‘Fuzzy approach for online coordination of plug-in electric vehicle charging in smart grid’, IEEE Trans. Sustain. Energy, 2015, 6, (3), pp. 1112– 1121 (doi: 10.1109/TSTE.2014.2327640) 14Safdarian, A., Fotuhi-Firuzabad, M., Lehtonen, M.: ‘Benefits of demand response on operation of distribution networks: a case study’, IEEE Syst. J., 2016, 10, (1), pp. 189– 197 (doi: 10.1109/JSYST.2013.2297792) 15Brusco, G., Burgio, A., Menniti, D. et al.: ‘Energy management system for an energy district with demand response availability’, IEEE Trans. Smart Grid, 2014, 5, (5), pp. 2385– 2393 (doi: 10.1109/TSG.2014.2318894) 16Kinhekar, N., Padhy, N.P., Li, F. et al.: ‘Utility oriented demand side management using smart AC and micro DC grid cooperative’, IEEE Trans. Power Syst., 2016, 31, (2), pp. 1151– 1160 (doi: 10.1109/TPWRS.2015.2409894) 17Luo, T., Dolan, M., Davidson, E. et al.: ‘Assessment of a new constraint satisfaction problem based active demand control approach to address distribution network constraints’, IET Gener. Transm. Distrib., 2015, 9, (15), pp. 2363– 2373 (doi: 10.1049/iet-gtd.2014.1133) 18Deng, R., Yang, Z., Hou, F. et al.: ‘Distributed real-time demand response in multiseller–multibuyer smart distribution grid’, IEEE Trans. Power Syst., 2015, 30, (5), pp. 2364– 2374 (doi: 10.1109/TPWRS.2014.2359457) 19Negash, A.I., Haring, T.W., Kirschen, D.S.: ‘Allocating the cost of demand response compensation in wholesale energy markets’, IEEE Trans. Power Syst., 2015, 30, (3), pp. 1528– 1535 (doi: 10.1109/TPWRS.2014.2349516) 20Gaurav, B., Pratap, S.B., Rajesh, K.: ‘Strategic optimisation of microgrid by evolving a unitised regenerative fuel cell system operational criterion’, Int. J. Sustain. Energy, 2014. Available at http://www.dx.doi.org/10.1080/14786451.2014.941367 21Eminoglu, U., Hocaoglu, M.H.: ‘Distribution systems forward/backward sweep based power flow algorithms: a review and comparison study’, Electr. Power Compon. Syst., 2009, 37, (1), pp. 91– 110 (doi: 10.1080/15325000802322046) 22Trovão, J.P., Henggeler Antunes, C.: ‘A comparative analysis of metaheuristic methods for power management of a dual energy storage system for electric vehicles’, Energy Convers. Manage., 2015, 95, pp. 281– 296 (doi: 10.1016/j.enconman.2015.02.030) 23BYD Electric Bus. Available at http://www.byd.com/na/auto/download/BYD%20Motors%20Inc%20%20DBE%20Program140122.pdf 24BYD Fe Battery for HEV/EV, 8 February 2009. Available at http://www.produktinfo.conrad.com/datenblaetter/50000-274999/251704-in-01-en-BYD_LITHIUM_FE_BLOCK_12V_10_AH_B_BMS.pdf Citing Literature Volume11, Issue3Special Issue: Distributed & Autonomous Dispatch and Control for Active Distribution Networks/Microgrids Potential Scheme to Realise Plug & Play of DERFebruary 2017Pages 615-626 FiguresReferencesRelatedInformation
Referência(s)