Fuzzy Estimation For The Steady Conjugate Heat Transfer System

The inverse heat transfer problem(IHTP) is to use the temperature information of the heat transfer system to estimate the unknown characteristic parameters, such as boundary conditions, thermal physical parameters, geometry configurations, etc. The heat-exchange equipments involving fluid-solid conjugate heat transfer are widely used in the field of power station, refrigeration, petroleum refining, and chemical engineering. The inverse heat transfer method have provided a promising technical solution for the optimization designing of heat exchanger, the comprehensive evaluation of the heat exchanger's performance, the online monitoring, etc.Due to the inherently ill-posedness nature of the inverse heat transfer problem, the typical inverse methods such as conjugate gradient method(CGM), and genetic algorithm(GA) has a strong dependence on the number of temperature measuring point and measurement error in the process of solving inverse problem. The decentralized fuzzy reasoning method(DFIM) based on fuzzy theory is an uncertainty reasoning method which possesses the ability of good robustness and fault tolerance, and it is able to make effective decisions based on inaccurate and incomplete observation information.In this paper, the inverse problem of the steady conjugate heat transfer system based on DFIM is studied, and the main works include the following four parts:(1) For the two-dimensional steady conjugate heat problem, a BEM-FVM method for solving the direct problem of fluid-solid conjugate heat transfer system is presented, and the method of BEM and finite volume method method(FVM) are respectively used to solve the conduction in solid region and heat transfer in fluid region. Combining the temperature and heat flux condition of the interface of fluid and solid region, the temperature of boundary node of solid region and the temperature field of fluid region are determined.(2) Introduced the basic principle and procedure to solve the inverse problem by CGM and GA, and the unknown boundary geometry of adhesive layer of the pipe model is studied. Some numerical tests are presented to discuss the effect of different initial guesses a of unknown boundary shape, the number of measuring points and measurement errors on the boundary shape inversion, which has shown the main problem of two method in solving the inverse heat transfer problems.(3) An decentralized fuzzy inference method(DFIM) was built to solve the two-dimensional geometry problem involving conjugate heat transfer by combining with the existed problem of inverse problem and the character of the fuzzy reasoning method. The deviations between the calculated and measured temperatures are taken as the input parameters of fuzzy inference units, and obtain the corresponding inference components by fuzzy inference. Then a synthesizing weighted approach based on normal distribution is built to weight and synthesize the inference components, and gain the compensations of heat flux distribution to revise the current initial guesses of inverse parameters.(4) The proposed DFIM is applied to estimate the boundary geometry and heat flux distribution of the two-dimensional fluid-solid steady conjugate heat transfer system. And the validity of the DFIM to sovle the inverse problem involving conjugate heat transfer problem is proved by comparison with CGM and GA. In the numerical test, we discuss the effect of different initial guesses, the number of measuring points and measurement errors on inversion result and comparsion with CGM and GA are are conducted to prove the validity of the DFIM. The results of tests demonstrate that the DFIM can significantly reduce the dependence of the number of measuring points, enhance the anti-interference capability of the temperature measurement errors. It can be concluded that DFIM possesses a good anti-ill-posed character.