Optimal design and evaluation criteria for acoustic emission pulse signature analysis

Successful pulse recording and evaluation is strongly dependent on the instrumentation system selected and the method of analyzing the pulse signature. The paper studies system design, signal analysis techniques, and interdependences with a view toward defining optimal approaches to pulse signal analysis. For this purpose, the instrumentation system is modeled, and analytical pulses, representative of the types of acoustic emissions to be distinguished are passed through the system. Particular attention is given to comparing frequency spectrum analysis and deconvolution referred to as time domain reconstruction of the pulse or pulse train. The possibility of optimal transducer-filter system parameters is investigated. Deconvolution of a pulse is shown to be a superior approach for transient pulse analysis. Reshaping of a transducer output back to the original input pulse is possible and gives an accurate representation of the generating pulse in the time domain. Any definable transducer and filter system can be used for measurement of pulses by means of the deconvolution method. Selection of design variables for general usage is discussed.