Research On Temperature Prediction Of Microwave Heating System Based On Combined Model

As a new type of heating technology,microwave heating is widely used in metallurgy,drying and other fields.Unlike traditional heating methods,microwave heating has huge advantages such as body heating,high efficiency and energy saving.Due to the unique characteristics of microwave heating,the temperature of the heated medium in the microwave field is difficult to predict,temperature gradients,"cold spots" and "hot spots" are prone to occur,leading to material burnout and even safety accidents.In the microwave field,the temperature of the medium are difficult to predict because of many factors,such as the distribution of the electromagnetic field,the presence of noise of the measurement device,the thermal conductivity of the medium and the characteristics of the wave absorption change with temperature.At present,the prediction model of medium temperature during microwave heating is mainly through mechanism modeling and modeling methods based on experimental data.In the establishment of the mechanism model,the relevant physical parameters of the medium are often simplified and some physical phenomena are ignored in order to facilitate the solution,which can only reflect the approximate temperature rise characteristics of the medium in the microwave field.The data-driven modeling method,although making up for the deficiencies of the mechanism model parameter simplification and conditional assumptions,but due to the noise contained in the data itself,the established model cannot accurately reflect the dynamic characteristics of the system.Based on the idea of model fusion,this paper first takes the heat and moisture transfer as an example to study the phenomenon of latent heat of evaporation that is easily overlooked when modeling the mechanism of microwave heating system,and the resulting changes in the absorption characteristics of the medium to illustrate the Influence of simplification of conditions to mechanism model.Then,the mechanism model is combined with the experimental data model to establish a combined temperature prediction model.The specific research is as follows:(1)Taking porous media titanium concentrate as the research object,the finite element mechanism model of microwave heating system was established.Calculating the moisture content change of titanium concentrate during microwave heating and the latent heat term taken away by the evaporation of water.At the same time,the change of the dielectric coefficient ?' and the dielectric loss coefficient ?" of the titanium concentrate during the heating process is calculated using the mixed medium theory.The effect of moisture and heat transfer on the final temperature distribution of the medium is analyzed.(2)Based on the data collected in the experimental environment,a particle swarm optimization support vector machine regression model was established,and the particle swarm algorithm was used to optimize the support vector machine penalty coefficient C and kernel parameter g,to predict the temperature of titanium concentrate in the microwave heating system.(3)The PSO-SVR model is used to compensate the estimation error of the mechanism model to establish a combined model of the parallel structure.According to the change of operating point,the combined model is updated through the move-bank data selection method.Finally,the prediction effects of the three modeling methods are compared.The experimental results show that the combined model has higher prediction accuracy and has a better advantage in the temperature prediction of the microwave heating system.