Research On Inverse Problem Of Unsteady Heat Transfer And Its Appliancation Based On Predictive Control

Using the part of temperature measurement information of heat transfer system to identificate unknown information of heat transfer system,such as the strength of heat source,thermal boundary condition,the thermal parameters and boundary shape,the method was called as the inverse heat transfer problem(Inverse Heat Transfer Problem,IHCP).IHCP and its applications are widely existed in many engineering fields.Dynamic matrix control(Dynamic MatrixControl DMC)inversion method using a dynamic step response of system as the predictive model,using the history information and future input to predict the future output,through rolling optimization instead of global optimization,to avoid the ill posedness caused by model error or environmental interference.In this paper,based on the DMC algorithm,In this paper,based on the DMC algorithm,a new method is developed for transient heat transfer inversion.The DMC inversion method uses the dynamic step response function as the prediction model,and realizes the inversion of the transient heat flux at the moment through rolling optimization.In addition,this algorithm establishes an optimization method based on the principle of regularization parameter error,significantly reducing the dependence of the inversion results for the future time step dependence and the sensitivity of inversion results on the error of temperature measurement,can use a small future time step to obtain better heat inversion results.Dynamic matrix control has been used to solve the one-dimensional and two-dimensional unsteady heat transfer system in a certain amount of the IHCP,and the advantage of the dynamic matrix control in the inverse problem of unsteady heat transfer system has been demonstrated.In this paper,the inverse problem of three-dimensional unsteady heat transfer system based on DMC is studied,the main research contents are as follows:(1)The basic content of the predictive control and basic principle of dynamic matrix control is briefly introduced;a model of three-dimensional unsteady heat transfer problem is established,and by the finite volume method and alternating direction implicit algorithm,the transient temperature field model was achieved.(2)The basic principle of solving the inverse problem of heat transfer by using the SFSM method is summarized.Based on the SFSM,the transient heat flux of the three-dimensional unsteady heat transfer system is inversed.Combined with the numerical experiments,the influence of the future time step and the measurement error on the inversion results is discussed,and the problems existing in SFSM for solving the inverse problem are summarized.(3)In terms of the problem of SFSM in solving inverse heat transfer problem,based on the characteristics of DMC,the DMC method for solving the inverse problem of three-dimensional unsteady heat transfer is established.In the method,the dynamic step response function is used as the prediction model,and it is not necessary to assume the specific function form of the heat flux to be inversed in the next period of time.On the other hand,based on the principle of residuals,a regularization parameter optimization method is proposed.Numerical experiments are carried out with different future time steps and measurement errors,and the results are compared with SFSM.The test results show that,compared with SFSM,DMC method obviously reduces the dependence of the inversion results for the future time step and the sensitivity of the inversion results on the error of temperature measurement,when the measurement error increases,the method can use a small step to obtain satisfactory results of heat flux inversion.(4)In terms of the proton exchange membrane fuel cell,a three-dimensional unsteady heat transfer model of the fuel cell was established,and using the DMC method,the heat flux distribution of inner wall of proton exchange membrane fuel cell electrode and membrane electrode interface was inversed,and the transient temperature field of the wall was reconstructed.Through numerical experiments,the influence of the inversion results for future time steps and measurement error was discussed,the experimental results show that with the larger measurement error,the DMC method can inverse ideal results through the smaller time step.The DMC method can be used to monitor the variation of the temperature at the interface between the inner wall of the end and the membrane electrode of proton membrane fuel cell.