Research On Early Composite Fault Information Extraction Technology Of Wind Turbine Gearbox

With the rapid development of wind turbine industry, the reliability requirements for wind turbines become higher and higher, and the fault diagnosis technology has been the tertiary industry in this industry. Strengthening the condition monitoring of the wind turbine gearbox, which is the key components of the wind turbine equipment, is of great significance.In the wind turbine gearbox, the occurrence frequency of planetary gear transmission fault is relatively high. Its fault diagnosis is more difficult than traditional fixed-axis gearbox because of its complicated construction. That how to extract the fault information of the planetary gear transmission effectively is an important research subject. For this purpose, a method based on resonance-based sparse signal decomposition(RSSD) was proposed to extract the early composite fault information of the planet carrier bearing and the planetary gear in the wind turbine gearbox, and obtain the good result.We selected the concrete construction which was primary planet transmission with secondary parallel shaft gear transmission for the research object, and determined that the fault types were planet carrier bearing outer races fault and planetary gear local fault. The transmission paths of the planet carrier bearing outer race fault and planetary gear local fault based on the specific construction were analyzed, and two degrees of freedom vibration model which is closer to the above transmission paths were proposed. We studied the time-domain waveform character of the actually measured planetary gear signal with broken teeth fault. This article also studied frequency spectrum characteristic of the planet carrier bearing outer race fault and planetary gear local fault, and analyzed the main components contained in the gearbox vibration signal.Combined with the time-domain waveform character of the early composite fault, this article proposed a method based on RSSD for early composite fault information extraction of the wind turbine gearbox. The effect on the result of fault information extraction of all the main parameters was explained through the detailed analysis of the parameters in RSSD. Considering the insufficient of the choice of the quality factors and the weight coefficients in RSSD, the quality factors and the weight coefficients of the high- and low-resonance components were be optimized by genetic algorithm that has the excellent global optimization performance. The optimized quality factors that considered the difference of the resonance of the composite faults, obtained the best matching effect with the impact response signal of the planet carrier bearing outer race fault and planetary gear local fault. The optimized weight coefficients could reduce the loss of fault information and avoided the coupling of the two resonance components effectively. Finally, considering the actual changed rotate speed, we utilized the envelope analysis or the order envelope analysis to show the result of the early composite fault information extraction.We used the wind turbine gearbox fault simulation setup of Tsinghua University to verify the method proposed in this article, and indicated the effectiveness of the proposed method with the gearbox early composite fau lts information extraction. Compared with the analysis result of the traditional RSSD, the method, the RSSD based on parameter optimization, proposed in this paper could avoid the misdiagnosis which would appear in the early composite faults information extraction of traditional RSSD, retaining more faults information in the signal processing and reduce the coupling between the high- and low-resonance components.