Noise Source Identification Applied in Electric Power Industry Using Microphone Arrays

Pengxiao Teng

doi:10.4236/eng.2013.51b028

Abstract

The noise source identification is an important issue in noise reduction and condition monitoring(CM) for machines in- site using microphone arrays. In this paper, we propose a new approach to optimize array configuration based on particles swarm optimization algorithm in order to improve noise source identification and condition monitoring performance. Two distinct optimized array configurations are designed under the certain conditions. Furthermore, an acoustic imaging equipment is developed to carry out experiments on transformer substation equipment and wind turbine generator, which demonstrate that the acoustic imaging system allows a high resolution in identifying main noise sources for noise reduction and abnormal noise sources for condition monitoring.

Highlights

As noise reduction and condition monitoring has gained in importance to modern industries, noise source identification has become the focus of a wide variety of research approaches in recent years
An acoustic imaging equipment is developed to carry out experiments on transformer substation equipment and wind turbine generator, which demonstrate that the acoustic imaging system allows a high resolution in identifying main noise sources for noise reduction and abnormal noise sources for condition monitoring
As is known that the noise source identification is highly related to the array beampattern, the beampattern formula is given for the designed array, and an modified particle swarm optimization (PSO) method is proposed to optimize array configuration both according to mainlobe width (MLW) and sidelobe level (SLL)

Summary

Introduction

As noise reduction and condition monitoring has gained in importance to modern industries, noise source identification has become the focus of a wide variety of research approaches in recent years. There has been considerable progress in single-channel acoustic CM in recent years[1,2], acoustic signals are, often adversely influenced by their measurement environment and by the range of different acoustic sources within a typical monitoring location. This can make it very difficult to extract useful information for condition monitoring purposes. Particle swarm optimization iteratively updates parameters to converge according to the best individual solution and the best swarm solution It is intuitive for array configuration optimization, and PSO method is much easier to implement.

Array Model

Planar Array Model

Modified Particle Swarm Optimization

Experimental results

Conclusions