Reconstruction Of Temperature Field Of Coke Tower Based On Reduced Order Model

To ensure the operational safety of coke towers and to increase their service life,the rate of temperature change is usually limited,but it also means that the production cycle is extended,i.e.there is a contradiction between the life of the coke tower and its productivity.To resolve this contradiction,accurate and real-time monitoring of the temperature field at each process stage is required.However,the coke tower temperature field is difficult to measure directly: the harsh measurement environment inside the coke tower,during which processes such as raw coke and decoking are also involved,makes it impossible to arrange temperature or heat flow measurement devices inside the coke tower and on the inner walls;isolated and limited temperature measurement points cannot directly represent the temperature field.To address the above problems,this paper reconstructs the temperature field of coke tower by the heat transfer inverse problem solving method;considering the real-time requirements of coke tower temperature field monitoring,a fast reconstruction system of coke tower temperature field is constructed.The main research contents are as follows.(1)Mechanistic analysis of the heat transfer process in different process stages of the coke tower,the complex heat transfer conditions between different work substances and the tower wall in different process stages of the coke tower are uniformly considered as the work substance exerts a heat flow on the inner wall of the coke tower,and a unified physical and mathematical model of the heat transfer process is established;the finite element method is applied to solve the heat transfer equations of the coke tower,and numerical experiments are conducted to verify the effectiveness of the heat transfer The validity of the model and the finite element method is verified,and the transient data are provided for the next step of the reducedorder model.(2)To address the problem that the traditional finite element method is computationally expensive and the computation period is too long to meet the requirements of monitoring the temperature field of the coke tower when it is used to solve the positive problem of the reconstructed system,a reduced-order model of the coke tower with mixed thermal boundary conditions is constructed based on the temperature data obtained by the finite element method,and is compared with the finite element method through simulation experiments.The experiments show that the temperature calculation results of the reduced-order model are basically consistent with the finite element method,but can greatly reduce the calculation time and can be used for the calculation of the positive problem of the reconstructed system.(3)For the uncomfortable characteristics in the coke tower temperature field reconstruction system,this paper adopts a decentralized fuzzy inference algorithm to construct a coke tower temperature field reconstruction system to suppress the amplification of measurement errors in the reconstruction process.The difference between the measured temperature of the outer wall of the coke tower and the calculated temperature at the same position is used as input information,and the corresponding inference components are calculated by using the fuzzy inference unit;the spatial distribution sensitivity matrix constructed is used to weight and synthesize the inference components to obtain the compensation amount of the heat flow of the inner wall of the coke tower,so as to realize the identification by the heat flow of the inner wall and then obtain the whole temperature field of the coke tower.(4)Experimental validation: Firstly,numerical experiments are used to compare the accuracy and operational efficiency of the two methods when the POD reduced-order model and the finite element method are used as positive problems in the reconstruction system;the effects of the initial guess value,measurement error and the number of measurement points on the fast reconstruction results of the coke tower temperature field are investigated.The simulation results show that the reconstruction method proposed in this paper is more accurate,faster,less affected by the initial guess value and measurement error,and less sensitive to the change of the number of measurement points.(5)Engineering application: The above research was applied to a coke tower in a petrochemical enterprise: temperature measurement points were arranged on the metal outer wall of the coke tower,and the reconstruction method proposed in this paper was used to reconstruct the temperature field of the whole process of the operating coke tower.The reconstructed results better conform to the operation law,which can provide a basis for the safety monitoring and process optimization of the coke tower,and support the construction of the digital twin of the coke tower in the next step.