Study On The Fatigue Reliability And Neural Net Models Of Laser Cladding Part Under Repeated Impact Load

The paper analyzes the fatigue and invalidation of the laser cladding parts under repeated impact load, and researches the model of performance prediction.The paper chooses cobalt-base self-fusible alloy, nickel-base self-fusible alloy, and iron-base self-fusible alloy as coating material, chooses Cr18Ni9Ti, 45 steels and 40Cr steels as basal body material. Laser melting experiments were carried out on the CO₂ laser(HGL—90),the repeated impact experiments were carried out on self making repeated impact machine. The experiment sequence was determined by orthogonal way and the every sequence was determined by the group test way. The cladding samples under repeated impact load have been macro-analyzed and micro-analyzed. The results show that the forms of invalidation were various. There are mainly crack, depression, upset, pit on the surface, breakout, etc. The invalidation experiences three stages, including early distortion stage, steady stage and aging invalidation stage. The invalidation of laser cladding under repeated impact load is cumulate and belongs to the fatigue category. This paper analyzes the results of the invalidation and comes to the conclusion that the invalidation relates to many factors, including the matching of coating material and basal body material, thickness of cladding and the technology of laser cladding, etc.On the basis of the tests, the reliability model of laser cladding under repeated impact load is made. The paper finds that the fatigue life of laser cladding submit the logarithmic normal distribution. The paper dose the orthogonal analyses and gets the following results. The influence of coating material plays the most important role to the fatigue strength of the laser cladding layer under repeated impact load. Basal body material plays the less important role. And layer thickness plays the least important role. The layer is the thicker the better in these tests; If the layer is the nicket base alloy, the laser cladding will have the higher fatigue strength; in the middle life, the fatigue strength is related to the static compression strength. Generally speaking, if the layer has the higher static compression strength, the fatiguestrength will be higher. If the basal body material has the good ductility, the cladding sample will have the higher fatigue strength than the cladding sample with the brittle basal body. The basal body material, the cladding material and cladding thickness interact, but the interaction among them to the fatigue strength will decrease with the increasing of the cycle number. The paper fits P-S-N curve and analyzes the relation between the stress of the repeated impact and fatigue life. These results can help to choose the suitable material during the laser cladding and to predict the reliable fatigue life on the work condition.On the basis of the reliability analyse, the BP neural network is made to forecast the fatigue capability. The neural network is 6-11-1,the inputs are the layer material, the basal body material, the stress and the layer thickness, the output is the fatigue life. The paper analyses this neural net by the factor means and finds that the influence of stress plays the more important role to the fatigue strength of the laser cladding layer under repeated impact load than the layer thickness. The precision of this neural net is 0.000999993,the error is less than 4%. The neural network is precise enough to predict the fatigue capability of laser cladding under repeated impact load.In the MATLAB, the interface of the reliability analyse, neural network to forecast, the factor analyse of the neural network and the modeling of neural network are set up which establish the basic for the more analyses and researches.