The Study On Modeling And Control Of COT Random Distribution System In The Decoking Process Of Ethylene Cracking Furnace

The production capacity of ethylene is a measure of the level of petrochemical industry.Therefore,the production equipment of ethylene has also become the core part of the refining system.At present,the vast majority of ethylene in the world is produced in cracking furnaces.Due to the phenomenon of coking production raw materials,in the radiant section furnace tube,cracking raw materials and steam undergo dehydrogenation and carbon chain scission reactions.With the occurrence of side reactions,coke deposits on the inner wall of the furnace tube in the radiant section,resulting in a decrease in the production capacity of the ethylene plant.,Threatening production safety,so when the coke on the inner wall of the furnace tube accumulates to a certain extent,it must be burnt.In the coking process of the ethylene cracking furnace,the common method is the steam-air coking method,that is,high-temperature steam is passed into the cracking furnace.During this process,the COT temperature at the outlet of the ethylene cracking furnace gradually rises.At this time,the COT is controlled.Most importantly,only by controlling the distribution of COT can the decoking process be carried out smoothly and orderly.The decoking of ethylene cracking furnace is an unstable process.During the heating process of decoking,COT distribution is sometimes more dispersed,and some furnace tubes with too high temperature will reduce the service life of the cracking furnace tube wall and cause the furnace tube The wall burns through,causing irreversible damage.At this time,the decoking must be suspended,and the temperature will continue to scorch after adjusting the COT,which will increase the decoking time and delay production.Too low temperature will lead to too long scorch time and cause unnecessary economic losses.The traditional control method is to control the average value of 60 COT,and the effect of COT distribution control is not ideal.In response to the above problems,this paper conducts in-depth research on the COT random distribution system during the decoking process of the ethylene cracking furnace,models it and designs a controller,and proposes a random distribution modeling method based on working condition identification,according to the COT temperature rise rate Speed is used as an index to identify the working conditions,judge the current decoking conditions,and establish a PDF model of the probability density function of the COT of the ethylene cracking furnace during the decoking process.The controller is designed using the MVC3 method and verified by industrial data.The feasibility of building the model and controller,and actually applying it to the ethylene cracking furnace of the olefin department of a refining company of Sinopec.The research uses B-spline neural network to convert the system output PDF into corresponding weights,and analyzes the relationship between weights and system inputs,treats weights as system state variables,and uses WaveARX neural network to identify the relationship between system inputs and state variables relationship.Through decorrelation of wavelet frame basis functions,the ability to approximate the model is improved.Finally,a state feedback controller is designed,and industrial data is used to verify its feasibility.This article provides a new modeling and control method for the COT distribution control of the coking process of the ethylene cracking furnace,and provides a guarantee for the stable progress of the coking process of the ethylene cracking furnace.Finally,the development of advanced control software for automatic decoking and its application in actual industrial production equipment are introduced.