dc.description.abstract |
Pulsed electro-acoustic measurement system is one of the versatile systems for
measurement of charges accumulated in dielectric materials. The charge is measured as
a voltage signal observed at an amplifier output which gets input from an acoustic
sensor. Because of limited frequency bandwidth of the acoustic sensor and amplifier, all
experimentally acquired signals require processing technique like deconvolution to
estimate original signal (charge). Due to presence of noise in all experimental signals,
deconvolution turns out to be an optimization problem. The optimum deconvolution
involves estimation of optimum value of a deconvolution parameter, also known as
regularization parameter. This parameter controls the tradeoff between signal
distortion and noise minimization. In literature, several methods were suggested for
estimation of optimum value of regularization parameter. However, each method yields
a different estimate for the parameter. In this paper, the suitability of these
conventional methods is investigated for estimation of space charge. Also a simple, yet,
better estimation method, specific to space charge estimation, is presented. Interesting
results and discussion on efficacy of these methods against, simple spectral division
followed by Gaussian filtering, is also presented. |
en_US |