Research On Fault Diagnosis And Maintenance Of Hydraulic Turbine Based On Digital Twin

The fault diagnosis of hydraulic turbine is polymorphic,fuzzy and uncertain.The traditional way of fault diagnosis depends on the work experience of maintenance personnel,which leads to slow feedback of fault diagnosis,low efficiency and passive maintenance that the accuracy can not be guaranteed.This paper introduces digital twin technology,builds lightweight digital twin model,combines digital twin with numerical simulation,uses Bayesian network model for turbine fault diagnosis,generates fault maintenance knowledge map,and builds digital twin turbine operation and maintenance management system based on ITwin.The main research results of this paper are as follows:(1)Build the digital twin model of Francis turbine.Through the physical entity of Francis turbine and the numerical simulation theory of fluid,a digital twin five-dimensional model is created,which integrates physical information,multi-dimensional virtual simulation,data driven,dynamic interaction and application services.From the three levels of physical data,virtual data and fused data,the consistency test of virtual and real interaction is carried out to realize the real feedback of virtual mapping.(2)Lightweight processing of digital twin model for hydraulic turbines.An improved QEM algorithm is introduced,emphasizing triangular mesh facet reduction based on boundary features and geometric details.Using LOD multi level of detail technology,a mesh model with the highest fitness is generated to meet different levels of detail requirements.Render and invoke multi level of detail instance models in Unity3 D,and ultimately achieve lightweight applications of the digital twin model for hydraulic turbines.(3)Combining digital twinning with numerical simulation,an implementation scheme for hydraulic characteristics analysis of turbine inner runner is constructed.Take the Francis turbine as an example.Through the interactive model of digital twin and finite element established in Fluent,the CFD numerical calculation of multi-condition and all-channel is completed.To simulate the velocity,pressure and flow pattern of the flow field in hydraulic turbine,a complete and feasible software interaction method of digital twinning and numerical simulation is constructed.(4)The fault diagnosis and maintenance of hydraulic turbine under the digital twin background are studied.Summarize the fault history data of hydraulic turbine,construct the fault tree model,and transform the fault tree model into Bayesian model.Using the backward reasoning diagnosis technology of Bayesian network,the fault diagnosis is embodied in the form of probability quantification,and the sensitivity analysis is carried out through the root node to improve the accuracy of fault diagnosis and location.The ontology and logical relationship of turbine fault diagnosis and maintenance are established by using prot protégéand Neo4 J,and the knowledge map of turbine fault maintenance is generated,and the digital twin fault operation and maintenance management system based on ITwin is built.Therefore,the research on hydraulic turbine fault diagnosis and maintenance under the background of digital twinning expands the application scope and depth of digital twinning technology in water conservancy industry.The development of digital twin has practical and important application value to promote the development of water conservancy digitalization and enhance the operation and maintenance management ability of hydropower stations.