摘要
针对机车二系弹簧载荷均匀性分配调整的建模问题,提出综合运用机理建模和神经网络建模的混合建模方法。该方法在刚性车体假定下采用经典力学和数学方法建立机车车体-二系弹簧系统的机理模型,作为调簧主规律模型;用人工神经网络方法建立BP网络误差补偿模型来弥补机理模型的建模误差;二者并联组成混合模型,其输出为机理模型和BP网络模型输出的叠加。研究结果表明:混合建模方法用于二系调簧的多维连续空间系统建模,可大幅提高模型精度;实际调簧过程中使用混合模型可进一步提高调簧精度和效率,使载荷分布最大误差较机理模型减少8%~15%,平均调簧时间缩短25%以上。
A hybrid modeling method for adjusting the distribution of locomotive secondary spring loads was proposed based on the combination of mechanism modeling method and neural network. In this method, a mechanism model based on the rigid body assumption was established as the master-rule model for spring load adjustment by using mechanical and mathematical methods. An error compensation model was constituted by BP neural network to compensate for the error of the mechanism model. The hybrid model was the parallel connection of the two models above and its output was the summation of their outputs. The results show that the hybrid modeling method can make further improvement on the accuracy and efficiency of spring loads adjustment in solving the continuous multi-dimension space modeling problem. Compared with the mechanism model, the maximum deviation of secondary spring loads is reduced by 8%-15% and the average adjusting time is reduced by more than 25%.
出处
《中南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2012年第1期378-383,共6页
Journal of Central South University:Science and Technology
基金
铁道部科技研究开发计划项目(Z2007-079)
关键词
机车二系载荷
调整
混合建模
BP神经网络
locomotive secondary spring load
adjustment
hybrid modeling method
BP neural network