The purpose of the paper is to investigate an algorithm for recovery of the continuous relaxation time spectrum of viscoelastic materials from time-measurements of linear relaxation modulus. A new identification scheme based on the least-squares approximation of relaxation modulus by finite serious of modified Bessel functions is proposed. The analysis of the scheme properties, in particular the convergence and model error analysis, as well as the numerical studies conducted indicates that the scheme proposed is an accuracy and easy implementable tool for recovery of the relaxation time spectra. The identification algorithm can be successfully applied to study the mechanical properties of highly hydrated plant materials subjected to various kinds of loading as well as different water solutions used in the food industry.

77-91