Real-time Inverse And Applications Of Heat Transfer Process Based On Fuzzy Adaptive Kalman Filtering

Inverse heat transfer problem (IHTP) is a typical inversion problem, which uses the partial temperature output information to inversely acquire some unknown characteristics of the heat transfer system, such as thermal physical parameters, boundary conditions, geometric conditions and input sources. The application of various types of inverse heat transfer problems is widely used in the fields of energy, materials, aerospace, biomedical, and nondestructive testing,Based on the system state space model, Kalman filtering (KF) is used to estimate the state parameters of the system recursively with the the output observation information, and the mean estimation error square is required to be minimum. In this work, the real-time inversion method of the unsteady heat conduction process based on Kalman filtering and its application are studied. The problem of the traditional Kalman filtering method is presented in the thesis. Based on the fuzzy adaptive Kalman filtering (FAKF) and the order reduction technique, a real-time temperature field reconstruction method is proposed. The main research works are as follows,①For the two-dimensional unsteady heat transfer system, the state vector, state transfer matrix, input matrix and observation matrix are obtained by using the energy balance method and the corresponding state space model is established.②When the covariance of process noise is seriously mismatched with the actual situation, the temperature field obtained by the traditional Kalman filtering method will be able to appear obvious deviation or even unable to obtain the stable reconstruction results. To solve this problem, an adaptive Kalman filtering algorithm based on fuzzy inference is proposed. The error covariance of the Kalman filter is changed in real time by fuzzy inference unit whose inputs are the observation error and error change rate of the system’s state estimation.③Another practical difficulty of using the state space model and the Kalman filtering method to reconstruct the system’s temperature state is that it needs to provide adequate temperature measurement information. This requirement cannot be realized in practical engineering. In this work, the Aggregation Approach is used to reduce the order of two-dimensional unsteady heat transfer system’s state space model. By that means, the temperature field reconstruction results can be acquired with less measurement points, which significantly declines the dependency of reconstruction results on the measurement number.④Based on the fuzzy adaptive Kalman filtering (FAKF) and the order reduction technique, a method for on-line monitoring the temperature field of boiler drum is established. Based on a small amount of temperature measurement points on the outer wall of the drum, the distribution of heat flux during the startup process of the drum is inversed, and the temperature field of the metal wall of the drum is reconstructed in real time. The efficiency of the above method is demonstrated by simulation experiments on the startup curve of a 600MW subcritical boiler.⑤The problem of real time inversion of the moving heat source is studied based on the fuzzy adaptive Kalman filtering (FAKF) method. Taking the instantaneous moving point heat source in laser processing as an example, the state space model of the moving heat source system is established. The source term of the heat transfer system is inversed by the FAKF method in real time. The effects of measurement error and the number of points and the position of the measurement point on the inversion results are discussed. The results show that the method established in this thesis has more widely applicability for inverse heat transfer problem with moving heat source.