BANDWIDTH EMPIRICAL MODE DECOMPOSITION AND ITS APPLICATION

Abstract

There are some methods to decompose a signal into different components such as: Fourier decomposition and wavelet decomposition. But they have limitations in some aspects. Recently, there is a new signal decomposition algorithm called the Empirical Mode Decomposition (EMD) Algorithm which provides a powerful tool for adaptive multiscale analysis of nonstationary signals. Recent works have demonstrated that EMD has remarkable effect in time series decomposition, but EMD also has several problems such as scale mixture and convergence property. This paper proposes two key points to design Bandwidth EMD to improve on the empirical mode decomposition algorithm. By analyzing the simulated and actual signals, it is confirmed that the Intrinsic Mode Functions (IMFs) obtained by the bandwidth criterion can approach the real components and reflect the intrinsic information of the analyzed signal. In this paper, we use Bandwidth EMD to decompose electricity consumption data into cycles and trend which help us recognize the structure rule of the electricity consumption series.