CAJ | 학술논문

在对少齿差行星齿轮进行结构分析的基础上,根据其传动特点和设计要求,运用模糊数学的原理进行了模糊可靠性分析,建立了可靠性数学模型,将模糊设计优化模型转化为了常规的优化模型。通过所建立的模型可以实现最优参数选取,同时针对传统BP神经网络的不足,将模拟退火和BP网络相结合,设计了一种新型的改进神经网络。实验结果表明,此种算法得出的绝对误差和相对误差都较小。
재대소치차행성치륜진행결구분석적기출상,근거기전동특점화설계요구,운용모호수학적원리진행료모호가고성분석,건립료가고성수학모형,장모호설계우화모형전화위료상규적우화모형。통과소건립적모형가이실현최우삼수선취,동시침대전통BP신경망락적불족,장모의퇴화화BP망락상결합,설계료일충신형적개진신경망락。실험결과표명,차충산법득출적절대오차화상대오차도교소。
This paper, on the basis of analyzing few - tooth - difference planetary gear structure and according to its transmission characteristics and design requirements, builds a reliability mathematical model based on ambiguity reliability analysis through ambiguity reliability principle, which transforms the ambiguity design optimization model into the common optimization model. Through this model, the optimal parameters can be selected. And considering the deficiency of the traditional BP neural net- work, this paper designs a new type of improved neural network by combining stimulated annealing and BP neural network. The results showed that the absolute error and relative error through this algo-rithm are smaller. This is a comparatively accurate method for optimization design.