基于RBF神经网络的某型号行星变速器传动效率试验及预测研究

Transmission Efficiency Test and Prediction Research of a Planetary Gearbox Based on RBF Neural Network

为开展行星变速器的传动效率检测研究,同时节省时间和成本,采用试验测试和仿真预测相结合的方法,首先,基于专有测试平台监测行星变速器在1挡转动时的转速和转矩变化,并计算传动功率和传动效率;然后,取各工况下共66组传动效率测试值作为样本数据,其中59组作为训练样本,剩下7组为验证样本,采用RBF神经网络建立行星变速器传动效率预测模型,并成功进行训练和验证。研究表明:GJ002型行星变速器的传动效率满足设计使用要求;基于RBF神经网络模型预测的传动效率最大差值不超过0.22,该模型预测精度较高。

In order to realize the transmission efficiency detection research of planetary transmission and save time and cost at the same time,this paper adopts the method of combining experimental test and simulation prediction.Based on the proprietary test platform,the speed and torque changes of the planetary tachometer are monitored when it rotates in the first gear,and the transmission power and transmission efficiency are calculated.66 sets of transmission efficiency test values under each working condition are taken as the sample data,of which 59 sets are training samples,the remaining 7 groups are validation samples.Using RBF neural network,the planetary transmission efficiency prediction model is established,and training and verification are successfully carried out.Research shows that:the transmission efficiency of the GJ002 planetary transmission meets the design and use requirements,the maximum difference in transmission efficiency predicted based on the RBF neural network model does not exceed 0.22,and the prediction accuracy of this model is relatively high.