Short‐term load forecasting at electric vehicle charging sites using a multivariate multi‐step long short‐term memory: A case study from Finland

Abstract

This study assesses the performance of a multivariate multi-step charging load prediction approach based on the long short-term memory (LSTM) and commercial charging data. The major contribution of this study is to provide a comparison of load prediction between various types of charging sites. Real charging data from shopping centres, residential, public, and workplace charging sites are gathered. Altogether, the data consists of 50,504 charging events measured at 37 different charging sites in Finland between January 2019 and January 2020. A forecast of the aggregated charging load is performed in 15min resolution for each type of charging site. The second contribution of the work is the extended short-term forecast horizon. A multi-step prediction of either four (i.e., one hour) or 96 (i.e., 24 h) time steps is carried out, enabling a comparison of both horizons. The findings reveal that all charging sites exhibit distinct charging characteristics, which affects the forecasting accuracy and suggests a differentiated analysis of the different charging categories. Furthermore, the results indicate that the forecasting accuracy strongly correlates with the forecast horizon. The 4-time step prediction yields considerably superior results compared with the 96-time step forecast.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Ancillary services and electric vehicles: An overview from charging clusters and chargers technology perspectives

“…In the first, charging is complementary to other user needs, such as going to the supermarket, in the second charging drives the choice of the user needs. Furthermore, charging behaviors are reflected at different charging sites: (i) home [155,156] and public residential charging [157,158]; (ii) curbside [159][160][161] and semi-public charging [162,163]; (iii) workplace charging [164]; (iv) fleet charging [165]; (v) large semi-public charging [166]; (vi) fast(en route) charging [167,168]; (vii) special semi-public charging [169,170]; (viii) charging forecourts [171,172]; (ix) semi-private charging [173] and (x) charging hubs [174]. Although it is still quite early for the clusters to mature, the charging clusters derived from the review are in line with the clusters used in the Working Group 4 of the IEA GEF Global e-mobility program [175,176].…”

Modeling and simulation of charging characteristics of electric vehicle group under the mode of autonomous driving‐shared travel

“…Recognizing that it is difficult to use simple map partitioning or virtual network to reflect the complexity of real traffic network topology, many researchers introduce real traffic data to study. References [14][15][16] use real traffic network data for modeling and simulation and predict the spatial and temporal distribution of urban EV group charging demand. Based on the information of EV's real-time location, velocity and the state of charge(SOC) in the traffic network, reference [17] studies the problem of EV path planning and charging navigation.…”

Short‐term commercial load forecasting based on peak‐valley features with the TSA‐ELM model

“…In recent years, intelligent algorithms based on machine learning theories have emerged and are used in various fields. These include areas such as medicine, 12,13 crime tracking systems, 14 and power load forecasting 15,16 . Researchers prefer artificial neural networks (ANN) in dealing with STLF problems, 17,18 owing to their strong nonlinear learning ability and fault tolerance 19,20 .…”

“…These include areas such as medicine, 12,13 crime tracking systems, 14 and power load forecasting. 15,16 Researchers prefer artificial neural networks (ANN) in dealing with STLF problems, 17,18 owing to their strong nonlinear learning ability and fault tolerance. 19,20 However, many ANN-based gradient-based methods, such as backpropagation or other variants, have some limitations in the field of electric load forecasting.…”