The application of iterative regularization deconvolution to solve an ill-posed problem is discussed. Firstly, the theoretical foundation of iterative regularization is introduced, and a new Fredholm integral equation of the first kind including space charge distribution is derived .The effect of additive Gaussian white noise for iterative deconvolution algorithm, and the effect of iterative convergence criterion for the numerical solution are studied, respectively. It is shown that the solution of ill-posed problems could be obtained with iterative deconvolution algorithm under lower level noise circumstance. It also indicates that the precision of numerical solution is influenced by the iterative convergence criterion. Finally,the original experimental data are dealt with and the real space charge distribution in dielectrics is obtained.

A numerical algorithm for space charge distribution in the planar dielectric sample by Pressure Wave Propagation (PWP)method is presented. The real space charge distribution is obtained from the deconvolution based on Tikhonov regularization method.Several technical treatments related to the deconvolution technique have been adopted,such as erasing the high frequency noise with wavelet packet filters,treating with integral equation using Tikhonov regularization method etc..The affect of noise for deconvolution algorithm is studied indetail by numerical experiments. It is shown that deconvolution algorithm for solving space charge distribution in dielectric can obtain very good results under low noise circumstance,however the numerical results are affected significantly under high noise circumstance,although they are still acceptable. There is evident smoothing effect for numerical solutions using regularization parameter α,but there is no significant effect for integral of numerical solution. Applying the deconvolution method to analyse experimental data,space charge distribution and electric field distribution are as well obtained successfully.