A noise reduction forecasting method of bus load based on sequence decomposition

Abstract

In the context of new power systems， the diverse distributed power sources and user-side behavior have introduced instability in bus loads， thus presenting a fresh challenge for short-term load forecasting. In response to this challenge， a noise reduction forecasting method for bus load based on sequence decomposition is proposed. The construction and decomposition rules of the variational mode decomposition （VMD） are applied to the sequence decomposition of bus load. The residual terms following the sequence decomposition are smoothed using locally weighted regression （LWR） to forecast the noise reduction of bus load. Based on the measured active power data of bus load in an area， a recurrent neural network （RNN） is constructed to forecast the bus load after noise reduction. The results indicate that this method effectively eliminates the noise of the bus load sequence， and the sequence tends to be smooth while retaining the characteristics of the original bus load sequence. Moreover， it yields excellent forecasting curves and accurate forecasting results.

Keywords

• bus load forecasting
• noise reduction of bus load
• vmd
• rnn
• lwr