Research On Condition Monitoring And Fault Analysis System Of Petrochemical Pump

Petrochemical pumps are widely used in petrochemical industry,and various faults will inevitably occur in the operation process.In order to accurately locate and quickly respond to faults,real-time monitoring of the operating conditions is one of the most effective measures.The traditional monitoring system based on the on-site central control room is easily limited by geographical conditions,with low monitoring efficiency,poor resource management capability and often encountering difficulties in system expansion.In order to solve the above problems,this paper designs a petrochemical pump analysis system,centering on life cycle monitoring and management,integrates resource management and state monitoring to solve the deficiencies of traditional monitoring.First of all,by analyzing the operating characteristics of petrochemical pumps and combining the functional requirements of the monitoring system,the physical quantities of the state parameters to be monitored are designed: voltage,current,inlet and outlet pressure,vibration,etc.According to different business requirements,an edge adaptive platform is designed to collect and preprocess state parameters and complete information interaction with the cloud.The state data collected in real time has the attribute of time label,so the system uses the TimeScaleDB to design and develop the cloud data center to improve the data storage and access rate.Secondly,it uses.NET framework to develop a visual cloud platform for state monitoring.The platform realizes all-weather and all-round monitoring of the operation status of petrochemical pumps and provides the management service functions of petrochemical pump resources,mainly including planning and purchasing management,inventory management,operation and maintenance management,historical data management,etc.It finally realizes the monitoring and management of life cycle of equipment.The whole system has the characteristics of low development and maintenance cost,good expansibility,simple and convenient human-computer interaction,etc.Finally,a two-level fault analysis framework is proposed,in which high-speedsampled vibration signals are preprocessed quickly at the edge layer and comprehensive analysis is realized in the cloud.The cooperative operation of the two can not only shorten the fault response time of the whole system but also ensure the high accuracy of diagnosis.According to the different characteristics of each layer architecture,different fault feature extraction methods are studied: Symbolic Aggregation Approximation(SAX),discretizing vibration signals and reducing dimension of data,is selected for edge layer to improve response speed because of its fast,efficient operation,and so on.Variational Mode Decomposition(VMD)can effectively suppress mode aliasing and has good fault feature extraction effect.It is more suitable for cloud to realize accurate fault location and comprehensive analysis.