Research On Temperature Optimization Control Method Of Tubular Furnace

In modern oil refineries,multi-branch tube heating furnace is among the most commonly used and the largest energy-consuming equipment.With the technological progress and control theory development,the safety and economic efficiency of the tube furnace can be improved by control scheme optimization.In this article,the tube heating furnace is taken as the main object to model the heating furnace control system,optimize the outlet temperature control and design the control method of multi-branch temperature equilibrium.A large number of simulations and actual measurement data are utilized to verify the effectiveness of the modeling and control,thereof the specific research work and results are as follows:(1)This article proposes a set of cleaning methods to process the measured data of heating furnace,in order to improve the quality of data-driven modeling and optimal control,and suppress the adverse effects of outlier and random disturbance toward the model and control strategy.Aiming at the outlier in data,an improved isolated forest-based outlier detection algorithm and a multiple interpolation-based outlier repair are constructed.A data denoising method based on improved wavelet thresholding is presented to overcome large random errors in measured data.The effectiveness of the above outlier detection,remedy and data denoising methods is verified through the measured data and extensive simulation.(2)In the heating furnace outlet temperature control link,this article proposes an outlet temperature cascade control system based on adaptive particle swarm optimisation Fuzzy-PID to solve the problem of unstable crude oil outlet temperature of tube heating furnace.Then the heating furnace model is established by NARX modeling method to validate and simulate the outlet temperature control performance.The results show that compared with the conventional control scheme,the method in this article has milder overshoot,shorter response time and better anti-disturbance capability,which is suitable for controlling the heating furnace outlet temperature,and significantly reduces the fluctuation of the outlet temperature.(3)In the multi-branch temperature equilibrium control link of the heating furnace,this article raised a multi-branch temperature equilibrium control method based on an improved genetic simulated annealing algorithm-optimized multi-input multi-output PID neural network(IGSA-MPIDNN)to alleviate multi-branch temperature imbalance in tube heating furnace.Then the heating furnace model is established by a hybrid modeling method to verify the multi-branch temperature control effect.The results show that compared to the non-optimized MPIDNN controller and the conventional PID control scheme,the IGSA-MPIDNN controller is able to make the multi-branch outlet temperature reach the target temperature faster and with less fluctuation,thus improving the multi-branch temperature control performance.(4)In terms of control system implementation and application,this article completes the engineering of temperature optimization control software for tube heating furnaces.The QT software platform and C++language were used to develop a heating furnace temperature optimization control system software,including the functions of offline data cleaning and modeling,online optimization control,OPC communication,database storage,real-time data and sampling data display,etc.The article tests the function of the software system and achieves all expected functions.The innovations of this article include but not limited to:An improved isolated forest method is proposed,which has higher detection precision and speed while detecting outlier in data.An improved wavelet threshold denoising method is proposed,which has better denoising capability while processing data.An adaptive particle swarm optimization Fuzzy-PID controller is proposed,such that the problem of fixed parameters in traditional PID controller leading to the poor control effect of the tube heating furnace outlet temperature can be solved.A hybrid modeling method is proposed to solve the problem of poor accuracy of single NARX model and BP neural network modeling.An IGSA-MPIDNN based multi-branch temperature equilibrium control method is proposed solve the problem that the traditional control scheme is difficult to ensure the multi-branch temperature balance of the tube heating furnace.