Estimation of the moving heat source intensity using the multiple model adaptive inverse method

Abstract

This study proposes a multiple model adaptive inverse (MMAI) method to estimate the heat source intensity of nonlinear heat transfer system with a moving heat source. In the whole motion space of heat source, the nonlinear heat transfer system is divided into several linearized subspaces, and a corresponding linear prediction sub-model for the temperature at the measurement point is established for each subspace. Based on the linear prediction sub-models, the components of the heat source intensity compensation corresponding to each linearized subspace are obtained through rolling optimization. Then, according to the instantaneous matching degree between each prediction sub-model and the actual heat transfer system, the components are weighted and synthesized to gain the compensation of the guessed value. Finally, the estimation of instantaneous moving heat source intensity can be realized. Numerical experiments are performed to study the effects of the number of prediction sub-models, the moving velocity of heat source, the temperature measurement errors and the number of future time steps on the inversion results. Comparison with the existing adaptive sequential function specification method is also conducted, and it shows the validity of the inverse method developed in this paper.