Study On A Fuzzy Inference Method Based On Variable Universe For Solving The Two-dimensional Steady Inverse Heat Transfer Problem

The system inversion problem or inverse problem is to solve some structuralfeatures of the system or part of the input information based on the output informationof the system. In the recent decades, some inverse research topics have gained rapiddevelopment, especially in the energy and thermal physics fields. Inverse heat transferproblems (IHTP) is a typical inverse problem,which is based on the internal or surfacetemperature of the object studied to get the unknown parameters such as thermalparameters, the geometry, the boundary conditions and the source.In this paper, two-dimensional steady-state heat transfer inverse problem is solvedby distributed fuzzy inference method. The method based on variable region of fuzzyinference method is proposed for inverse problem solving process problems on thedomain selected. Specific work mainly includes the following three aspects:1）The direct problem is solved by FEM for two-dimensional steady-state heattransfer model rectangular plate. Boundary temperature is obtained by distributed fuzzyinference method, and the effect of the selection of the initial domain is analysised.2.）Combined with traditional control theory and stability problems in the presenceof the inversion process variable universe fuzzy theory and heat transfer systems, fuzzyinverse heat transfer systems thinking is proposed. The initial output of the domainselected cases, according to the error input domain of real-time adjustment of the valueof the input variable, effectively ensuring the inversion process against ill-posed.3.）For the heat transfer mathematical model of double kiln furnace wall, the ideaof variable domain fuzzy inference is used to inverting the surface temperaturedistribution of furnace wall inner. On second thoughts, the effects of various parameterson the inversion results are also discussed.In this paper, the research results show that using decentralized fuzzy inferencemethod can effectively inverting the boundary temperature distribution of thetwo-dimensional steady-state heat transfer. In fuzzy inference process, the selection ofinitial domain greatly influences the inversion effect, and the variable domain thoughteffectively optimizes this problem. Relative to the fixed theory fuzzy inference, thevariable region fuzzy controller can be dynamic adjusted depending on the input error inthe fuzzy inference process, and reduces the requirement for initial domain, simplifiesthe design of fuzzy controller. For the actual heat transfer system, the decentralized fuzzy inference method has good resistance to discomfort.