On Line Retrieval Of Time-varying Heat Flux And Temperature Field In Participating Media Based On Kalman Filtering

Participating media are widely used in the aeronautics and astronautics,energy power and high-tech fields,such as the thermal protection materials of the reentry vehicles,high temperature ceramic parts of the various engines,optical windows of the spacecraft,solar collectors and thermal insulation layers of the turbine engines are participating media.The time-dependent heat flux on the boundary and internal temperature fields need to be monitored in near real time even real time to e nsure the safe and efficient operation of the above equipments.However,direct contactmeasurement of the time-dependent heat flux on the boundary and internal temperature fieldsof the above equipments can be incredibly challenging even impossible due tosevere heat exchange environment.Hence,they need to be reconstructed by the indirectly means(inverse technique).At present,near real time even real time reconstructed of the time-dependent heat flux on the boundary and internal temperature fields are mainly focused on the pure heat conduction problems.The reconstruction of boundary time-varying heat flux and internal temperature distribution in participating media is mainly based on the off-line method of traditional gradient and random search algorithms.Therefore,the strong robust and precise calculation model for the near real time even real time retrieval of the time-dependent heat flux on the boundary and internal temperature distributi on in participating media is urgent to be built.Thisthesisfocusesonthe near real time even real time reconstruction of the time-dependent heat flux on the boundary and internal temperature fields in participating media.The standard Kalman filter(KF)technology for linear system,extended Kalman filter(EKF)technology for weak nonlinear system and unscented Kalman filter(UKF)technology for strong nonlinear system are introduced into study the problem.The main work is as follows:Firstly,the KF,EKF,and UKF are introduced in detail.Meanwhile,the KF coupled with the recursive least-square Estimator(KF-RELSE)is also introduced,which is widely applied in the inverse heat conduction problem.The time-dependent heat flux on the boundary and internal temperature fields are reconstructed by employing the KF and KF-RLSE,respectively.The validity and reliability of the KF technique are verified based on the experimental temperature information.When the light-thermal parameters of the participating media do not change with the temperature,the KF technique and KF-RLSE algorithm are employed to reconstruct the boundary time-dependent heat flux and internal temperature distribution in the participating media with homogeneous refractive index in real time based on the temperature signal on the boundary.The retrieval results display that the KF technique has better stability and accuracy in the inverse conduction coupled with conduction heat transfer problems.Based on this,the KF technique is applied to retrieve the time-dependent on the boundary and internal temperature fields in the graded index media.When the light-thermal parameters of the participating media is a function of the temperature,the KF technique and EKF algorithm are employed to reconstruc t the time-dependent heat flux and internal temperature distribution in the two-dimensional participating media in real time.The reconstructed results show that when the KF is utilized,the big error is existed between the reconstruction result and real one.The time-dependent heat flux,internal temperature field and phaseinterfacein the nonlinear conduction-radiation heat transfer are reconstructed by the EKF and UKF in real time.The estimated results show that the EKF can only solve weakly nonlinear problems effectively.Furthermore,the light-thermal parameters,time-dependent heat flux and temperature field are retrieved by the UKF technique simultaneously.The retrieval results show that at least measurement informationof two positions is needed for simultaneous real-time reconstruction of the above parameters.In order to improve the stability and accuracy of the reconstructed results,the smoothing technique is introduced,which uses the measurement information in a short period of time in the future.Based on the measurement information in the future,the UKF coupled with fixed interval smoothing technique is applied to reconstruct the time-dependent heat flux and internal temperature fields in the nonlinear conduction-radiation heat transfer.The reconstructed results show that when the fixed interval smoothing technique is introduced,t he time delay and stability of reconstruction boundary heat flux and accuracy of reconstruction temperature distribution are significantly improved.Meanwhile,the effect of the number of the future temperature information on reconstructed results.The estimated results show that only the measurement information near the reconstruction moment can effectively improve the reconstruction results...