Many studies about heat transfer characterization of single phase fixed bed matrix regenerators are devoted to the finding of experimental correlations. Despite several deep investigations, the emerged correlations are not well established, indeed the high complexity of the processes involved, the shape of the solid-fluid interface, the complexity of the geometry of the solid matrix, make accurate experimental data difficult to obtain. The aim of the present work pursuit a double objective: (i) to develop and propose an inverse method to identify h, the fluid-matrix heat transfer coefficient, by means of transient simulated experiments, and (ii) to investigate the sensitivity of the h reconstruction process to the variation of the control input parameters and material properties, in order to find the optimal value of the experimental control variables that allows the identification of this unknown coefficient to be performed with "minimum variance". The reconstruction technique is applied to numerical experiments and it is based on the simulated measurements of oscillating temperatures of the fluid at the inlet and outlet of the regenerator. The identification of h is performed by means of an inverse search technique, driven by the difference between simulated measurements and calculated temperature time histories at the regenerator outlet. At first, experiments in different operating conditions are simulated in order to investigate the ability of the algorithm to identify the correct value of 11 and its uncertainty. Then a parametric study is performed and the optimal control frequency of the known (imposed) oscillating temperature signal at the inlet is found as a function of the mass flow rate, the geometry and other operating and thermophysical characteristics of the system. (C) 2012 Elsevier Ltd. All rights reserved.