Research On Online Monitoring And Rating Of Transmission Accuracy Of RV Reducers For Industrial Robots

Industrial robots are increasingly widely used in modern enterprise production.RV reducer is the core part of industrial robot,and the transmission accuracy directly affects the normal operation of production.With the automation,digitization and intelligent upgrade of the production of enterprises,the transmission accuracy detection of RV reducer of industrial robot is put forward higher requirements.In order to meet the needs of industrial production,scholars at home and abroad have carried out meaningful research on the transmission accuracy of RV reducer and obtained experience and results.The research on transmission accuracy on-line monitoring and rating of RV reducer is still in its infancy.The degradation of transmission accuracy life of diren RV20E-121 reducer in the operation of domestic six-axis industrial robot is studied.The rated life of transmission accuracy of RV reducer estimated by formula method is compared with the test function of transmission accuracy degradation of RV reducer to estimate its life.The shortcomings of these methods are analyzed.Finally,the three-direction dynamometer and three-axis acceleration vibration sensor are used as output signal,Constructing a real-time monitoring system for RV reducer.Through comparative analysis,Sym5 wavelet basis,soft threshold function and SUREShrink threshold estimation method are selected to complete wavelet threshold denoising.The reconstructed signals are analyzed in time domain,frequency domain and wavelet packet,and a SBS search strategy combined with support vector machine non-independent criterion is used to realize the transformation from high dimensional signals to low dimensional signals,which reduces the disaster of data dimension in a certain program.In the experiment part,RV reducer with different transmission accuracy is divided into three life degradation state zones,including green health zone,yellow warning zone and red failure zone.The experimental factors are load and speed.The three-level orthogonal test table for transmission accuracy monitoring of RV reducer is made,which reduces the number of experiments.GWO algorithm is introduced to optimize the penalty factor and kernel parameters of RBF kernel function of support vector machine.The 20-dimensional feature vectors(729 sets of data)corresponding to three kinds of different transmission accuracy were used as training samples and test samples of support vector machine.By comparing the RATING methods such as SBS-GA-SVM,the recognition accuracy of the RATING method of SBS-GMO-SVM adopted in this thesis has been improved to some extent.