PV-based EV charging station with vehicle-to-grid services for business premises

Abstract

The transportation sector is the second major contributor to the global greenhouse gas (GHG) emissions, next to electricity generation sector. GHG emissions are projected to grow in the future due to the over-reliance on fossil fuels for both electricity generation and transportation sectors. However, the growth can be stalled by gradually introducing electric vehicles (EVs) as they produce less GHG compared to their conventional counterparts. Despite that, their proliferation has been stunted by a completely new set of challenges in addition to a very high capital cost. These challenges include low driving ranges, scarcity of electric vehicle charging stations (EVCS), long charging time and inability to provide charging service readily. Nevertheless, some of these challenges can be addressed by installing EVCS at business premises (i.e., offices, universities, etc.) and augmenting EVCS with onsite PV and battery energy storage (BES). In that case, the impacts of EV charging on the grid will be less significant; charging services will be available readily, and GHG emissions growth can be reduced significantly. Moreover, EVCS, under suitable conditions, would be able to provide vehicle-to-grid (V2G) services, which makes EVCS more attractive. Despite such prospective opportunities, offhand planning and operational planning of EVCS would deem them unachievable. In addition, the inherent intermittency of PV, and EV and grids loads will pose considerable hindrance on availing those benefits.Hence, this research focusses on an EVCS with onsite PV and BES, in the perspective of planning and operational planning including V2G services. The cornerstone of the planning and operational planning is an appropriate EV load model that captures the uncertainties of EV charging. In addition, the EV load model reflects the grid voltage and EV battery’s state of charge (SOC)-dependency of EV charging. Moreover, it accounts for the diversity of the EV population embodied by market shares, battery sizes, charging levels, and charging voltages, currents, and power factors. Furthermore, it is scalable to facilitate seamless assimilation of a large EV population. Therefore, a new EV load model is first proposed with MATLAB/Simulink validation reflecting the factors above along with the recommendations by the standard BS EN 61851-23:2014.This EV model is then incorporated into the planning activities, which include four chronological exercises. Firstly, the impact (i.e., grid voltage, current, losses, etc.) of the EV charging on the grid at different locations, namely home, office, public charging, is investigated considering the uncertainties. Secondly, the optimal location of the PV and BES based EVCS is divulged regarding the reduction of the impact and costs involved while enhancing the charging quality of service (QoS). Thirdly, a suitability analysis is performed for the obtained optimal location to find the most suitable combination of the grid upgrading, PV deployment, and BES deployment to satisfy the QoS, cost, and impact thresholds simultaneously. Finally, the optimal PV and BES sizes are obtained using sequential quadratic programming (SQP).The operational planning includes two tasks namely day-ahead scheduling of the EV load and probabilistic charging of the EV population regarding the uncertain PV output and grid load. The day-ahead scheduling method is developed incorporating the uncertainties of the combined SOCs of the existing, arriving, and departing EV populations alongside the diversity of the EV population mentioned above. Conversely, the probabilistic charging algorithm is devised considering the previously ignored correlation coefficients among the intermittent EV load, PV output, and grid loads. Moreover, BES is also included as a backup source in the proposed algorithm. To reduce the execution time of the tasks a new SOC-based charging strategy in conjunction with the new probabilistic impact indices are suggested.Finally; the real-time operation of the EVCS accomplishes two undertakings, e.g., 1) real-time charging of the EV population regarding the measured PV output, and 2) availing V2G services. The first exercise takes into account the present PV output measurement as well as the predicted PV output to dispatch the EV load with the aim to maximize energy harvest from PV while minimizing the impact on the grid. In contrast, due to the longer than required parking hour, the EV population do not need to be charged continuously. They can remain idle for a portion of time, which is defined as the laxity. Therefore, depending on the degree of laxity, they can provide V2G services, provided that they be willing to do so. Thus, a model is first developed to predict such V2G potential in time series. Then, a charging/discharging algorithm is proposed that can simultaneously maintain the QoS threshold and maximize V2G services at the minimal cost.The efficacies of the algorithms above are tested on the University of Queensland (UQ) parking lots coupled with the UQ electric grid and IEEE 37 bus test system with practical PV output and grid load data collected from UQ Solar and Australian Energy Market Operator. Favourable results confirm that this research will enormously help the uptake of the PV and BES based EVCS in the future. Moreover, the new EV load model for a large EV population encompassing the SOC and grid voltage-dependency of the EV load and the new centralized impact indices for the grid voltages and currents will not only expedite the proliferation of EVCS by facilitating more accurate, rapid planning and operational planning but also will form a new realm of future research. Furthermore, a more elaborate algorithm for obtaining the V2G services will ensure the satisfaction the EV owners by ensuring minimum QoS and reimbursing incentives, while earning the EVCS owners additional revenue.