Microphone Phased Array Machinery Fault Diagnostic System

A microphone phased array machinery fault diagnostic system is proposed. In this system, a microphone phased array is positioned such as to have a direct line of sight to the machine to be monitored. The microphone signals are recorded simultaneously and used to scan or beamform over a grid enclosing the machine to capture its acoustic signature. The acoustic image is then integrated to obtain a machine noise spectrum that is used to identify damage or faults by an automated expert system. The architecture of the proposed fault diagnostic system consists of two main separated physical subsystems. The first subsystem consists of the microphone phased array, data acquisition system, and a computer to generate the cross spectral matrix of the microphone signals. This component is a single physical unit mounted in the industrial plant. The cross spectral matrix is then transferred through a wireless link to the second subsystem which is a computer cluster to perform the beamforming, integration, and fault diagnostic of the equipment. These two subsystems are two distinct hardware setups interconnected through a wireless link, i.e. they are not physically connected.