A load forecasting method for steel industry based on ICEEMDAN

Abstract

In response to the challenge of low accuracy in load forecasting for the steel industry， a load forecasting approach is introduced based on ICEEMDAN （improved complete ensemble empirical mode decomposition with adaptive noise）. Firstly， ICEEMDAN is employed to decompose the electricity load in the steel industry into high and low-frequency modal components. Subsequently， the genetic algorithm （GA） is used to identify the primary influencing factors of the components. The PSO-BP （particle swarm optimization-backpropagation neural network） algorithm is then applied to construct prediction models for the high and low-frequency components. Finally， the forecasted results of each component are iteratively recombined to derive the ultimate load forecasting outcomes. Case analysis results indicate that this approach exhibits minimal prediction errors and higher precision.

Keywords

• impact load
• load forecasting
• iceemdan model
• ga
• pso-bp model