Study On On-line Visual Monitoring And Fault Diagnosis System Of Wind Turbine Blade

Wind power is a more mature power generation method in power generation technology of new energy.Conditions on large-scale wind turbine are extremely complex and abominable.Wind blades are key components of wind turbine to absorb wind energy,which cost more than 23%in the whole unit.Because of the blades suffered from the indefinite load during long time operation,which will lead to different degree of cracks or excessive deformation and ultimately fracture fault of blade.If it can’t be discovered in time,which will resulted in serious accidents to the unit.Through the analysis of the fracture mechanism and the fracture mode of wind power generation blades,combining actual working condition,dynamic position deviation method was proposed to detect the blade faults by the mark at the tip of each blades.To improve the accuracy of fault diagnosis,the adaptive threshold method and empirical formula were proposed,and the delay judgment method was used to eliminate the influence of non-fault factors.The experiment of monitoring the blade running state showed that the method has better reliability and robustness.The main research results of this subject include:(1)The structure of wind turbine and aerodynamic performance of the blade were investigated.At the same time,the fracture mode and fracture mechanism of wind blades were analyzed and researched.The development process of the equipment fault diagnosis and running status monitoring technology were summarized.(2)From the perspective of image processing and computer vision,an effective remote on-line monitoring method of wind turbine blade-position deviation method is proposed.It is based on the full consideration of the influence of working environment,failure modes,natural bending or pitch on the image extraction results.(3)In order to improve the accuracy of the above method,the stability of system monitoring and the efficiency of early warning,the adaptive threshold method,delay judgment method,process judgment method,fault type prediction method and the proportion coefficient compensation value for improving the accuracy of the prediction are proposed at this core.(4)A computer vision-based on-line wind turbine blade condition monitoring and fault diagnosis model system including wind turbine model,PLC-based control subsystem and camera-based vision monitoring subsystem and data communication subsystem using Ethernet and RS485 communication was developed.Fully simulate the actual environment,to verify the above theory,to prove its accuracy and effectiveness.(5)A remote monitoring system for wind turbine blades adapted to the outdoor complex environment was designed.The selection of key equipment such as camera and fill light was analyzed,and wind turbine blade failure data cloud computing model was preliminarily designed.Based on the research results,it was declared to be Tianjin science and technology support plan.