Improving Load Forecasting of Electric Vehicle Charging Stations Through Missing Data Imputation

Abstract

As the penetration of electric vehicles (EVs) accelerates according to eco-friendly policies, the impact of electric vehicle charging demand on a power distribution network is becoming significant for reliable power system operation. In this regard, accurate power demand or load forecasting is of great help not only for unit commitment problem considering demand response but also for long-term power system operation and planning. In this paper, we present a forecasting model of EV charging station load based on long short-term memory (LSTM). Besides, to improve the forecasting accuracy, we devise an imputation method for handling missing values in EV charging data. For the verification of the forecasting model and our imputation approach, performance comparison with several imputation techniques is conducted. The experimental results show that our imputation approach achieves significant improvements in forecasting accuracy on data with a high missing rate. In particular, compared to a strategy without applying imputation, the proposed imputation method results in reduced forecasting errors of up to 9.8%.

Systematic Comparison of Imputation Models for Automatized Gap Filling on Electrical Load Data of Compressor Composites in the Industrial Sector

“…Kanda et al [4] inserted data from similar meters in the missing periods before implementing a probabilistic load forecast in the GEFCom2017 final match competition. Lee et al [5] introduced an imputation method using both univariate and multivariate imputation techniques combining spline interpolation and Expectation Maximization based on maximum likelihood estimation. Ryu et al [6] applied a denoising autoencoder for imputation on smart meter data with different missingness patterns, like random, block-wise, with different configurations and predefined missing scenarios.…”

Comparison of Missing Data Imputation Methods in Time Series Forecasting

“…Long Short-Term Memory (LSTM) is a deep learning model for learning sequence data that can be applied widely, such as in natural language processing [4], human action recognition [30], and time series forecasting [2]. In LSTM neural networks, three gating mechanisms are implemented, thereby providing advantages to gradient vanishing/exploding problems, which are significant disadvantages of machine learning models based on artificial neural networks.…”

“…Among the wide array of data classes, time series, which is a sequence of data arranged in chronological order (i.e., a typical type in data analysis), has many real-world applications in various domains, such as energy [1,2], climate [3], economics [4], business [5] and healthcare [6]. The use of a significant amount of time series data that can be obtained via sensors and computing devices that have evolved in recent years might enhance analysis and forecasting abilities for solving real-life problems.…”

Electric Taxi Charging Load Prediction Based on Trajectory Data and Reinforcement Learning—A Case Study of Shenzhen Municipality

“…The literature [10] calculated EV trip and charging characteristics to establish probabilistic models by collecting GPS and charging data from 15 EVs in Ireland and used the Monte Carlo method to predict the charging load of EVs. The literature [11] refined the EV GPS data by univariate and multivariate interpolation techniques and predicted the EV charging station loads based on long-and short-term memory neural network models.…”