Study Of Acoustic Emission Signal Evaluation Technology Based On Multi-agent System

With the development of economic, industries and technology, people pay more attentionto the safety problems. So various industries have put forward higher standard for thedevelopment. As an extensive applied non-destructive method, Acoustic Emission(AE)techniques can be used to inspect equipments on line. Sensor array attached on the structurewill receive signals from different sources, the damage state can be judged through post datatreatment and signal analysis.After that, the potential danger sources can be identified.MAS consists of several agents, which are coordinating each other. Also, each of themcan service for the others and finish an assigned job by working together. MAS decomposes alarge and complex system into some small systems, which can communicate with each otherand easy to be managed. At present, MAS technology is widely applied to intelligent robot,traffic signal lamp control and pattern recognition fields. In this paper, MAS is applied on AEsignal processing and evaluation. based on that, a new signal analysis method is advanced toincrease the accuracy of AE signal processing and evaluation. The main content is listed asthe follows:1. According to the principle of functionalization of the MAS, the easily damage part ofthe tested structure is predicted before AE testing. Then, inspectors can confirm specificsensor array to examine its important parts.2. Each element of MAS plays its role and has different importance. Based on theacoustic emission location data, the distances from the acoustic sources to each sensor areachieved. After that,these distances are used as the key parameter to put forward a newalgorithm for the weighting factor calculation of each sensor in the evaluating system.3. Based on the frequency attenuation theory of acoustic emission wave in the rock, acomputational method suitable for particular AE arrays is established to backward deducethe peak frequency of the acoustic sources.4. Based on the theory of waveform similarity, Putting forward the similarity analysismethod ofAE signals and the theory to identify the sound source categories.5. Applying AE techniques to monitor the tensile process of nodular cast iron sample,the noise signals are filtered based on similarity analysis method of AE signal. The AEfeatures in different tensile stages are identified. The result shows that the analysis methodadvanced in the paper is helpful on the evaluation of material acoustic emission characteristics. 6.After multi-sensor array AE experiments on pressure vessel, the calculation method ofbackward deducing the peak frequency of the sound source and similarity analysis method areapplied to analyze the testing data, the result shows that they can help to identify sourcescategory and improve the evaluation.