Research On Performance Analysis Method Of Fan Lubricating Oil Based On Neural Network

Green manufacturing and environmental protection are the only way to sustainable development.Economic development relies on the burning of minerals to generate more and more power.With the rise of various new energy industries,wind power technology is maturing and gradually becoming the dominant position of new energy.As one of the key components of wind turbines,the uearbox not only plays a decisive role in the overall performance of the wind turbine,but also the service life directly affects the service life of the wind turbine.Usually,the normal operation of the gearbox is closely related to the lubrication condition in the tank.Therefore,it is particularly important to predict the oil performance of the gearbox and evaluate it in real time.In this paper,the neural network is combined with the gray system and the fuzzy system to realize the prediction of the lubricating oil parameters and the comprehensive evaluation of the performance.The specific work is as follows:1)Inductive screening of the measured physical and chemical indicators of lubricating oil samples.According to the importance of each index in the real-time conditions of the fan gearbox,seven indicators are selected among the many indicators to focus on analysis and calculation.Firstly,the data is cleaned,and the selected indicator data set is processed by the enhancement factor to reduce the randomness of the data and reduce the error;then the weight of each indicator is determined by the analytic hierarchy process,and the deterioration degree of the index is calculated by the normalization method.2)Preliminary results of lubricating oil performance are obtained by grey relational degree analysis of the processed data sets:grey GM(1,1)prediction model and BP neural network prediction model are established to predict and evaluate lubricating oil parameters separately;and parallel grey BP neural network prediction model is established to realize parallelization of lubricating oil arameters.Combined mixed forecasting and comprehensive index evaluation.The following conclusions can be drawn:first,the model can predict the quantitative index of lubricating oil performance changing with time line,and predict other parameters with greater correlation from some known parameters;second,the results show that the parallel grey BP neural network can predict lubricating oil parameters accurately and accurately.Number change can significantly reduce the error of single model prediction.3)According to the calculated deterioration degree of each index and the weight of each parameter relative to health level,a fuzzy neural network based lubricant health assessment model is established.The following conclusions can be drawn:Firstly,the analysis of the deterioration degree of the lubricating oil index and the weight of the health level can accurately obtain the real-time deterioration degree and health condition of the oil product.Second,the established fuzzy neural network prediction model can predict the oil.The trend of the comprehensive deterioration of the product.