Research On Flow Corrosion Prediction And Methods Of Condition Regulation Of Heat Exchange Equipment Based On Data Driven

With the increasing demand for oil and natural gas resources in our country and the serious shortage of domestic oil reserve resources,the utilization ratio of petrochemical refining enterprises to high-sulfur,high-acid and low-chlorine crude oil has been increasing year by year,causing the corrosion problem of oil refinery equipments to be aggravated.In addition,with the rise of artificial intelligence and the Internet,petrochemical enterprises have seen an overall increase in the level of informationization.The data collected in heterogeneous systems with complex formats and diversified information has been extending.How to apply the data analysis to the corrosion prediction of the refinery of petrochemical enterprises and realize the intelligent on-line supervision and alarm protection of the high-risk equipment of the oil refining corrosion is an important focus of the integration of the two cultures.Under this background,this thesis takes the mobile corrosion problem of the key refinery equipment in petrochemical enterprises as the research object,and conducts the research on the flow corrosion prediction and the condition supervision method,and mainly completes the following research work:(1)Aiming at the characteristics of non-linear correlation and large data dimension of oil refinery equipment operation status in petrochemical enterprises,an appropriate data-driven neural network(Random Vector Functional Link,RVFL)combined with Pearson's correlation coefficient is selected as the basic algorithm to improve a small-weight random weights neural network model proposed.The dataset test shows that the model converges faster and the calculation error is better.Compared with the feedback neural network(Back Propagation Neural Network,BPNN)model and the original RVFL model,it has better generalization ability and prediction accuracy.(2)Aiming at the problem of complicated storage format of heterogeneous database system and scattered monitoring data in heterogeneous database system of petrochemical enterprises,a flow corrosion monitoring platform based on B / S mode is put forward and the data integration of heterogeneous data systems such as DCS and LIMS is completed.The proposed mobile erosion prediction model analyzes a large amount of data and forms a unique database of mobile erosion parameters.Corresponding warning information and protective measures are given based on experience and standards,and the intelligent monitoring of the corrosion status of the refinery operation is achieved.(3)The designed flow corrosion monitoring platform is applied to the real-time supervision of Zhoushan Petrochemical hydrotreating unit.The results shows that the model can accurately and quickly predict the crystallization temperature of ammonium chloride with an error of 5.51% and an average calculation speed of 159 ms.The designed and developed supervision platform realizes on-line real-time supervision of the flow corrosion state,reduces the risk of corrosion of oil refining heat exchange equipment,and ensures that the equipment can operate stably for a long period of time.The innovation of this paper lies in establishing a data-driven model of flow corrosion prediction and embedding the model into the expert diagnosis and supervision platform designed so that a large amount of data information on hold in petrochemical enterprises can be effectively utilized,providing a key parameter for mobile corrosion Calculation method.In addition,the monitoring platform can provide functions such as real-time monitoring of oil refinery data,back-end process simulation,flow corrosion prediction and overrun alarm,which provide operating platforms and data support for intelligent corrosion prevention and control and risk assessment of refinery equipment.