Research of Extracting Characteristic Parameters from Acoustic Emission Signals Based on Neural Network

摘要

The extraction of characteristic parameters from acoustic emission is the key point for pattern recognition, according to which testing objects can be analyzed and diagnosed accurately.In this paper, a method of extracting characteristic parameters from AE signals based on BP neural network is proposed, and some experiments are carried out by breaking a pencil lead and a gentle knock on the aluminum plate using the AE detector SAM OS made by Physical Acoustics Corp.