Research On Key Technologies Of Predicting The Timing Of Active Remanufacturing And Recycling Of Machine Tools In Service

As an important part of green manufacturing,remanufacturing has a good energy saving and emission reduction effect.The uncertainty of rough embryos in traditional remanufacturing has formed a bottleneck in the development of remanufacturing technology.Active remanufacturing can solve this bottleneck well,and the key problem of active remanufacturing is the determination of the timing of active remanufacturing.This article takes data modeling as an entry point and builds an active remanufacturing timing decision model based on product life cycle theory.The main research content includes the following three aspects:(1)Considering of the characteristics of the machine tool monitoring signal that is susceptible to external interference and low signal-to-noise ratio,the monitoring signal is de-noised through wavelet analysis to ensure the quality of the signal data.Feature extraction is performed on the denoised signal to obtain the time domain index and frequency domain index that characterize the machine tool operating status information.In order to avoid dimensional disasters,a principal feature extraction method based on principal component analysis is proposed,which reduces the computational complexity by fusing multidimensional feature indexes.(2)Based on the extracted main feature indicators,a GRNN-based remaining life prediction model of the machine tool is constructed to avoid the influence of human subjective judgment on the prediction results to the greatest extent in traditional signal analysis.In order to improve the prediction accuracy of the remaining life prediction model,by analyzing the performance of the neural network,the PSO algorithm is introduced to optimize the structural parameters of the neural network,and a smooth parameter optimized GRNN prediction model is constructed.(3)Based on the product life cycle theory,comprehensive consideration of machine tool purchase cost,remanufacturing cost and maintenance cost,the two-parameter Weibull distribution is used to characterize the reliability function of the machine tool,and the expected cost per unit time is taken as the goal to construct an active remanufacturing recovery opportunity for active machine tools Forecasting model.Through the conversion of genotype and phenotype in the Gene Expression Programming(GEP)algorithm,a remanufacturing cost prediction model based on GEP was constructed,and the complex functional relationship between machine tool running time and remanufacturing cost was tapped.Finally,the feasibility of the built model was verified by predicting the timing of remanufacturing a batch of machine tools in a machine tool factory.