A Study On The On-line Prediction Of Fracture Appearance Transition Temperature Of 2.25Cr-1Mo Steel

In petrochemical and petroleum refining industries, hydrogenation technology is widely used, such as hydrosulfuring, hydrofining and hydrocracking technology and so on. Hydrogenation reactors are key facilities in hydrogenation units. Therefore, they are very important in engineering and academia to assure the safe operation of reactors and predict their remaining life during long-term service. 2.25Cr-1Mo steel has been widely used to construct the vessels of hydrogenation reactors. During the long-term service, the steel could degrade due to temper embrittlement and the degree of the degradation could increase with the extension of the operation of the hydrogenation reactors. So they are very significant in engineering and academia to investigate the temper embrittlement of the 2.25Cr-1Mo steel used in the hydrogenation reactors and predict the Fracture Appearance Transition Temperature (FATT) of 2.25Cr-1Mo steel on the basis of emergent technology. This thesis focuses on the online prediction of the FATT of 2.25Cr-1Mo steel used in hot-wall hydrogenation reactors based on the J2EE architecture and artificial neural network. At first, the model of BP network was built and the component was implemented to predict the FATT of 2.25Cr-1Mo steel with Java programming. Based on Struts and Hibernate, the web application for the online prediction of the FATT of the steel was developed with the J2EE architecture.Combined the web application with the component the result of the FATT prediction is given to users on the browser. In order to validate the predictive methodology, an engineering application example was illustrated. Good agreement is obtained when the predictive value was compared with the test data in open literature. The system is used on Internet so the user can input some operation parameters, such as the operation temperature and the base metal, and heat treatment rules, such as operation time and long service temperature to get FATT of the hot-wall hydrogenation reactors. Then the remaining life can be predicted. The system makes it possible to maintain hydrogenation reactors from plan-centered equipment management mode to prediction-centered mode.