A method to determine the sensor transfer function and its deconvolution from acoustic emission signals

Interests in monitoring of microfracture in brittle materials or any other mechanical process by acoustic emission (AE) are steadily increasing. AE signals carry information about their source transferred in the form of elastic waves. These can be monitored by piezoelectric sensors. The characteristic of the sensor, however, distorts this information severely, particularly in the frequency range near its resonance. Hence, deconvolution of the sensor characteristic is desirable, in order to remove the sensor's effects on the measured data. Due to the high bandwidth and dynamic range of the sensor the evaluation of its characteristic requires a special approach. In this paper, a method for the determination of the transfer function of AE sensors is presented. For this purpose, an appropriate strategy with two simple sensor configurations has been developed. The advantage of the present method is that a sharp electrical impulse is used as an input function instead of a undefined mechanical excitation. For the determination of the sensor transfer function, a linear set of equations have to be solved. Division in frequency domain yields no reasonable results. By applying optimal deconvolution in time domain, problems of instability can be avoided. This method is also used to deconvolve the transfer function of the sensor in AE signal of subcritical crack growth in alumina ceramics under load in a three point bending test.