Home // eKNOW 2020, The Twelfth International Conference on Information, Process, and Knowledge Management // View article
Authors:
Keh-Kok Yong
Woan-Lin Beh
Boon-Yaik Ooi
Wai-Kong Lee
Keywords: Electricity Consumption Forecasting; Simple Exponential Smoothing; Holt Linear Trend; Holt Linear Damped Trend; Holt-Winters; Auto Regressive Integrated Moving Average.
Abstract:
Electricity usage prediction is important for planning and facility expansion in the industrial sector. Accurate prediction can save the operation and maintenance costs, increased the reliability of power supply and delivery system for future development. This paper is to compare various Exponential Smoothing models (Simple Exponential Smoothing, Holt Linear Trend, Holt Linear Damped Trend and Holt-Winters) and ARIMA (Autoregressive Integrated Moving Average) model in an attempt to predict the daily electricity usage in the industry in the production of Hammer and Pellet with high accuracy. The data used is from 30 September 2019 to 06 November 2019, which consist of only 39 observations after excluding the non-production day. These models are precise and modelled well when the time series data is in a short period and in short period forecasting. Accuracy level of each model is measured by comparing the Root Mean Square Error (RMSE) of forecasting value with the actual value. Based on the comparison result, the best model with the smallest RMSE value is given by Holt Linear Trend for the electricity usage for the production of Hammer (in Mill 1 and Mill 2) and the production of Pellet (in Mill 1). In the data of the electricity usage for the production of Pellet (in Mill 2), the smallest RMSE value is given by Holt Linear Damped Trend. To improve the training and forecasting speed, we adopt Graphics Processing Unit (GPU) acceleration through RAPIDS Artificial Intelligence framework.
Pages: 91 to 96
Copyright: Copyright (c) IARIA, 2020
Publication date: March 22, 2020
Published in: conference
ISSN: 2308-4375
ISBN: 978-1-61208-765-8
Location: Valencia, Spain
Dates: from November 21, 2020 to November 25, 2020