Load Forecasting for Data Centers Based on Time Series

Abstract

With the booming development of the data center industry around the world, the power load forecasting of the data center plays an important role. In this paper, based on wavelet analysis and time series-based power load forecasting methods, data center power load forecasting is carried out. This paper uses the monthly power load data of a data center to carry out medium and long-term load forecasting, and achieves a good forecasting effect. This paper also uses daily load data to carry out short-term load forecasting, and uses a combination of one-step forecasting and time series-based methods for forecasting to achieve better forecasting results.