Niamh O'Connell

Prediction and optimization methods for electric vehicle charging schedules in the EDISON project

Smart charging, where the charging of an electric vehicle battery is delayed or advanced in time based on energy costs, grid capacity or renewable contents, has a great potential for increasing the value of the electric vehicle to the owner, the grid and society as a whole. The Danish EDISON project has been launched to investigate various areas relevant to electric vehicle integration. As part of EDISON an electric vehicle aggregator has been developed to demonstrate smart charging of electric vehicles. The emphasis of this paper is the mathematical methods on which the EDISON aggregator is based. This includes an analysis of the problem of EV driving prediction and charging optimization, a description of the mathematical models implemented and an evaluation of the accuracy of such models. Finally, additional optimization considerations as well as possible future extensions will be explored. This paper hopes to contribute to the field of EV integration by coupling optimized EV charging coordination with the EV utilization predictions on which the former heavily relies.

Electric Vehicle (EV) charging management with dynamic distribution system tariff

Electric Vehicles (EV) presents a unique opportunity for large-scale flexible demand, particularly when subject to intelligent charging. A smart charging algorithm is proposed here, with the dual objectives of minimizing charging costs and preventing grid congestion. EVs are charged according to individual user requirements while respecting the constraints of the local distribution grid. A day-ahead dynamic distribution system tariff (DT) scheme is proposed to avoid congestion on the local distribution system from the day-ahead planning perspective. Locational marginal pricing is used to determine the dynamic distribution system tariff based on predicted day-ahead spot prices and predicted charging behaviors. Case studies were carried out using distribution grids from the Danish island of Bornholm and the case studies demonstrate the efficacy of the proposed EV charging schedule algorithm

Economic Dispatch of Demand Response Balancing Through Asymmetric Block Offers

This paper proposes a method of describing the load shifting ability of flexible electrical loads in a manner suitable for existing power system dispatch frameworks. The concept of an asymmetric block offer for flexible loads is introduced. This offer structure describes the ability of a flexible load to provide a response to the power system and the subsequent need to recover. The conventional system dispatch algorithm is altered to facilitate the dispatch of demand response units alongside generating units using the proposed offer structure. The value of demand response is assessed through case studies that dispatch flexible supermarket refrigeration loads for the provision of regulating power. The demand resource is described by a set of asymmetric blocks, and a set of four blocks offers is shown to offer cost savings for the procurement of regulating power in excess of 20%. For comparative purposes, the cost savings achievable with a fully observable and controllable demand response resource are evaluated, using a time series model of the refrigeration loads. The fully modeled resource offers greater savings; however the difference is small and potentially insufficient to justify the investment required to fully model and control individual flexible loads.

Efficient determination of distribution tariffs for the prevention of congestion from EV Charging

A dual objective electric vehicle (EV) charging schedule optimisation is proposed here whereby both consumer driving requirements and grid constraints are respected. A day-ahead dynamic tariff (DT) for distribution systems is proposed as a price signal to EV fleet operators (FO) bidding into the day-ahead market. The DT acts to disperse charging at congested periods and locations, thereby preventing congestion on a day-ahead basis. The magnitude of the DT is determined from a simulated locational marginal prices (LMPs), and the time extent of the DT is determined from analysis of the system loading curve prior to the application of the DT. Case studies were performed using a sample distribution network modelled on a network from the Danish island of Bornholm. A variety of price profiles were used to illustrate the efficacy of this approach. The case study results show that this approach is highly efficient at grid congestion prevention, and the precise level of congestion that can be alleviated is dependent on the price profile of the optimisation period in question.