Modification Optimization Design Of Multi-stage Gear Transmission System Based On Machine Learning Under Multiple Working Conditions

With the problems of resource exhaustion and environmental pollution increasingly prominent,new energy vehicles are gradually active in the public view.Different from the driving devices of traditional fuel vehicles,new energy vehicles are driven by integrated motor,and the noise of motor is far less than that of internal combustion engine.Due to the absence of the masking effect of engine noise,the vibration noise produced by the transmission system is more prominent.In order to pursue higher driving speed,the high speed input of motor leads to more strict design of gear transmission system,and the vibration and noise of multi-stage gear transmission system become the focus of attention.In this paper,the multi-stage gear transmission system of a new energy vehicle is taken as the research object,and the main purpose is to improve the vibration and noise of the multi-stage gear transmission system.(1)The three-dimensional model of multi-stage gear transmission system is established by using Romax software to carry out multi-working condition simulation analysis,and the dynamic and vibration noise characteristics of the transmission system are studied.According to the gear meshing characteristics,the three-dimensional modification mode of tooth direction combined with tooth profile is determined,and the value range of each modification parameter is calculated.Sampling was carried out based on the optimal Latin hypercube method,and parametric modeling of modification was carried out to obtain the corresponding maximum vibration acceleration value,which provided data support for the subsequent process.(2)XGBoost machine learning algorithm is used to explore the mapping relationship between the modification parameters and vibration acceleration,and a modification parametervibration noise prediction model is created.The model is trained to make the accuracy of the model reach 98%.Taking the minimum vibration acceleration as the optimization objective,the standard particle swarm algorithm is used to solve the prediction model to obtain the optimal solution of the modification parameters.Under the condition of constant velocity and acceleration,the modified parameterized three-dimensional model was created for simulation analysis.The results show that the transmission error and load distribution of the gear pair are greatly improved,the maximum vibration acceleration is reduced by 63% and 76% respectively,and the error between predicted and simulated values is less than 2%,which verifies the effectiveness of the optimal solution and the feasibility of the noise reduction design.(3)A multi-condition shape modification noise reduction design is proposed based on the combination of the constant velocity and acceleration conditions.The combination of multicondition comprehensive modification parameters is obtained by weighting the optimal modification parameters under the condition of constant velocity and acceleration,taking the proportion of working time and the contribution of vibration as the influencing factors.The multi-objective fuzzy optimization evaluation model is established.The transmission error,maximum unit length load and maximum vibration acceleration of two-stage gear pair are taken as the objective.Based on the comprehensive evaluation of the overall shape modification effect under the constant speed and acceleration conditions,the feasibility of the multi-condition comprehensive modification optimization method is verified by simulation analysis.