Global Feedback-fuzzy Inference Algorithm And Its Application To The Inverse Heat Conduction Problem

The inverse heat conduction problem(IHCP)is the problem of solving the unknown variables in the heat transfer system according to the given information of measurement points and other conditions.IHCP is widely used in engineering,such as biology,textile,nondestructive testing,and so on.Nowadays,the methods to solve IHCP mainly include the gradient method and non-gradient method.The gradient method has a small amount of computation,but it does not have the global searching ability,so it may get a local optimal solution instead of the global optimal solution.Besides,the IHCP is an ill-posed problem,the error of measurement information will be magnified during the solution process,thus affecting the accuracy of the calculated result.Non-gradient methods include Genetic Algorithm,Particle Swarm Optimization,etc.,which have global search ability but a large amount of computation.The Fuzzy Inference(FI)algorithm has a small amount of computation,strong robustness and discomfort resistance,which can effectively reduce the influence of measurement error on the inversion results.The above characteristics are of great help to solve the IHCP.However,the conventional FI algorithm has a strong dependence on the qualitative knowledge of the heat tran sfer system,and it is not effective when applied to the IHCP whose heat transfer law is unknown or the heat transfer law is qualitatively described as a non-monotonic function.In this paper,the Global Feedback-Fuzzy Inference(GFFI)method is proposed to overcome the above disadvantages of conventional fuzzy inference.The main research contents are as follows:Firstly,the basic principles of fuzzy set theory and the fuzzy inference method are introduced.Then,the Feedback-Fuzzy Inference(FFI)method is proposed by combining fuzzy inference method with the feedback theory.Furthermore,GFFI method is proposed by combining the FFI method with the Simulated Annealing algorithm.Finally,the effectiveness of the algorithm in solving mathematical inverse problems is verified through the inverse problems of unipolar valued functions and multipolar valued functions.Secondly,the finite volume method is used to simulate the temperature field in the integrated controller of the electric vehicle,and the equivalent simplification of this model is carried out to reduce the amount of calculation.Then,taking the design problem of the air-cooled radiator of the integrated control ler as an example,the geometric boundary of the air-cooled radiator is solved from the perspective of the IHCP by using the GFFI method.The effects of the initial value and measurement error on the inversion results are discussed,and the effectiveness of GFFI method in solving the IHCP with unknown heat transfer law are verified.And we compare the GFFI method with the conventional FI method to prove its superiority.Finally,,we using the GFFI unit to build the decentralized fuzzy inference(DFI)system and solve the unknown geometric boundary of the two-dimensional plate heat transfer system.The influence of the initial value and measurement error on the calculation results is discussed.Then,comparing this DFI system with the DFI system composed of conventional FI units,the effectiveness of this method are verified when the heat transfer law of IHCP is qualitatively described as a monotone function.Then we use the DFI system based on GFFI algorithm to solve the geometric boundary of a two-dimensional water-cooling channel,and it is proved that this system can effectively solve the IHCP with multiple input and multiple output when the qualitative description of the heat transfer law is unknown.