Smart coordination of predictive load balancing for residential electric vehicles based on EMD-Bayesian optimised LSTM


IET RENEWABLE POWER GENERATION, vol.16, no.15, pp.3216-3232, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 16 Issue: 15
  • Publication Date: 2022
  • Doi Number: 10.1049/rpg2.12572
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Applied Science & Technology Source, Aquatic Science & Fisheries Abstracts (ASFA), Communication Abstracts, Compendex, Computer & Applied Sciences, Greenfile, INSPEC, Metadex, Civil Engineering Abstracts
  • Page Numbers: pp.3216-3232
  • Gazi University Affiliated: Yes


The charging load forecasting of residential Electric Vehicles help grid operators make informed decisions in terms of scheduling and managing demand response. The residence can include integrated residential appliances with multi-state and high-frequency features. For this reason, it is difficult to estimate the total load of residence accurately. To overcome this problem, this paper proposes a hybrid forecasting model using the empirical mode decomposition and Bayesian optimised Long Short-Term Memory for load balancing based on residential electricity meter data. The residential electricity meter data includes three datasets as Electric Vehicle, heat pump and photovoltaic system. To decompose of the data characteristics, the empirical mode decomposition method performs to the original data. Then, the Bayesian optimised Long Short-Term Memory is applied to forecast for each sub-component of the data sequentially. The main features of the proposed model include a significant improvement in prediction accuracy and capture the local maximums. The advantage of the proposed method over existing methods are also verified over with experiments of data-driven on the IEEE 33 busbar test system. The result of simulation forecasting model indicates that predict closely the busbar outflow power, voltage drop, transformer loading states and power losses to compare with actual load model.