Single Channel Blind Source Separation for Gas Regulators’ Acoustic Signal Using Eemd-Fastica

Abstract

The gas pressure regulator (GPR) is an important equipment in the gas pipeline network and fault detection of the gas pressure regulator is essential. Recently, an acoustic-emission-based method has been proposed for gas pressure regulator fault detection, while the single-channel signal of the regulator is mixed-signal due to the vibration of other components like pipes and pilot. In this paper, a single channel blind source separation (SCBSS) algorithm based on ensemble empirical mode decomposition (EEMD) and fast independent component analysis (FastICA) is introduced for gas regulators’ acoustic emission signal. The algorithm can solve the problem of mixed-signal and recover the source signal of GPR. The degree of the signals’ frequency center (FC) change when fault occurs evaluates the performance of algorithm. The experimental result show that it has been improved 53.06%, from 294Hz to 450Hz.