Research On Online Fault Diagnosis And Early Warning Of Heat Exchangers Based On Virtual Sensing

Heat exchanger is an important heat exchange equipment in petrochemical,nuclear power and other energy industries.With the improvement of industry demand and process complexity,the safety monitoring of heat exchanger status is faced with the lack of direct,real-time and online monitoring methods,so it is difficult to achieve accurate fault diagnosis and maintenance,while the common monitoring and analysis methods that directly apply process parameters such as temperature and pressure only indirectly reflect the operating condition of the heat exchanger,and the actual fault degree is usually obtained offline after shutdown.In this thesis,starting from the virtual sensing theory,a virtual sensing method based on mechanism and data fusion is proposed,and a heat exchanger fouling thickness estimation method based on virtual sensing and a heat exchanger leakage fault diagnosis method based on virtual sensing are established.A heat exchanger condition monitoring and health management system has been developed,and the specific research contents include:(1)A virtual sensing method based on mechanism and data fusion is proposed.In view of the problems that the mechanism-based knowledge-driven modeling method is difficult to accurately describe the existence process of complex systems with multi-disciplinary correlation,and the experimental cost is high,and the data-driven modeling method has poor generalization ability and heavy dependence on data quality and other problems,the idea of virtual sensing is introduced.A virtual sensing framework for heat exchangers based on mechanism and data fusion is constructed.Combined with HHM,the main failure modes of heat exchangers are identified.The performance indicators of heat exchangers are analyzed from the aspects of heat transfer,resistance and comprehensiveness,and a virtual sensing modeling and reasoning method is proposed.Experiments simulate fault data,and improve the monitoring and analysis capabilities of heat exchangers.(2)A fault diagnosis method for heat exchanger leakage based on virtual sensing is established.In view of the various leakage failure modes of shell-and-tube heat exchangers,the leakage monitoring method relying solely on the primary monitoring parameters of the heat exchanger has problems such as difficulty in distinguishing the leakage form and accurately evaluating the leakage degree in real time.Aiming at the leakage of the flange sealing surface of the shell-and-tube heat exchanger and the leakage of the medium,the correlation between different leakage forms and monitoring parameters and performance indicators is studied,and a virtual perception model of heat exchanger leakage based on heat balance calculation is established.The model is applied to the leakage fault diagnosis of shell-and-tube heat exchangers,and the effectiveness of the method is verified by the leakage simulation experiment analysis of the actual heat exchanger.The average error between the virtual sensing results and the actual leakage is 5.3%,and the leakage virtual sensing model the leakage rate that can monitor the internal leakage of the medium is above 1.80%,and the leakage rate of the flange sealing surface leakage is above 1.83%.(3)A fault diagnosis method for heat exchanger leakage based on virtual sensing is established.In view of the various leakage failure modes of shell-and-tube heat exchangers,the leakage monitoring method relying solely on the primary monitoring parameters of the heat exchanger has problems such as difficulty in distinguishing the leakage form and accurately evaluating the leakage degree in real time.Aiming at the flange leakage of the shell-and-tube heat exchanger and the internal leakage,the correlation between different leakage forms and monitoring parameters and performance indicators is studied,and a virtual sensing model of heat exchanger leakage based on heat balance calculation is established.The model is applied to the leakage fault diagnosis of shell-and-tube heat exchangers,and the effectiveness of the method is verified by the leakage simulation experiment analysis of the actual heat exchanger.The average error between the virtual sensing results and the actual leakage is 5.3%.The leakage rate that can monitor the internal leakage is above 1.80%,and the flange leakage rate is above 1.83%.(4)The heat exchanger condition monitoring and health management system is developed.Analyzed the relationship between input and output data of each algorithm module,constructed a condition monitoring and health management algorithm library,analyzed system architecture and functional requirements,designed and developed heat exchanger data acquisition,data analysis,performance evaluation,failure mode analysis and health The state management function module realizes the application of parameter configuration,algorithm operation and visual result display of heat exchanger state monitoring and health management related algorithms.