Parametric Identification of Unsteady Heat Conduction Processes Under Conditions of Bounded Uncertainty

Abstract

Paper presents the problem of parametric identification of processes of technological thermophysics based on the solution of inverse heat conduction problems under conditions of random disturbances. The solution of the boundary inverse heat conduction problems is sought, for which the method of parametric optimization of physically substantiated desired characteristics on compact sets of polynomial functions is used. The proposed approach is based on the alternance properties of optimal trajectories, optimizes the solution in a uniform estimation metric (using the minimax functional), and is the algorithmically accurate method. An increase in the intensity of disturbing factors causes difficulties in the direct application of alternance properties. An algorithmically accurate method has to be combined with additional algorithms that provide adequate results under conditions of bounded uncertainty of disturbances. As possible approaches, preliminary smoothing of information, control of an ensemble of trajectories satisfying the conditions of interval uncertainty, and the method of artificial intelligence are used. Using an exact analytical model, the combination of the minimax optimization method with algorithms operating under conditions of information uncertainty leads to regular solutions of the inverse heat conduction problem that are satisfactory inaccuracy.