Research On Fault Diagnosis System Facing Wind Turbinc Gearbox

Gearbox is an important component of wind turbine transmission system, and is one of the components which have higher failure rates. Wind turbine outage event caused by gearbox fault have occurred occasionally, the losses of direct and indirect are growing. Because of its small installation space, and is located high above the ground, once in the event of failure, maintenance is very difficult. So making quick and accurate diagnosis and handling correctly is of great practical significance and economic value. In this paper, wind turbine gearbox is taken as research object, the fault diagnosis method is studied.In the context of that artificial intelligence technology continues to develop and gradually penetrate into the field of machinery fault diagnosis, a wind turbine gearbox fault diagnosis method combining with based on fault tree analysis and based on GRNN is proposed.Based on the collation of wind turbine gearbox components and analysis of failure mechanism, the wind turbine gearbox fault tree is drawn; fault tree knowledge is stored in the database by tiered storage, the knowledge database is built; the reasoning process is designed by backward reasoning, the inference engine is established.the vibration signals under three typical fault conditions including normal, tooth wear, tooth fracture are analyzed in the time domain and frequency domain respectively, five characteristic parameters including clearance factor, kurtosis factor, peak value factor, impulse factor and power spectral entropy is extracted to be taken as the input of GRNN, and the network is trained to be a gearbox fault state recognition model based on GRNN.After tested through using reserved vibration signals, the diagnosis result of fault diagnosis method based on GRNN in good agreement with actual situation; by using the system function of searching and querying, maintenance staff can quickly find the cause of the fault and expert-level solutions from Fault diagnosis based on fault tree analysis.Experiments show that, the system can give an accurate and efficient identification on wind turbine gearbox fault state according to vibration signals; it can diagnose the known fault rapidly and give expert-level solutions. Therefore, the system is conducive to the accurate and rapid diagnosis and maintenance of wind turbine gearbox failures.