随机故障注入结合神经网络法的机电系统可靠性计算方法

发布时间：2018-04-14 15:39

本文选题：神经网络 + 动态　；参考：《机械工程学报》2017年06期

【摘要】：将随机故障注入方法与神经网络技术相结合,提出机电系统多失效模式可靠性计算方法。以液压缸内外泄漏故障为事例,将虚拟故障信息注入活塞杆线性定位系统一体化仿真模型中。利用神经网络较强的函数逼近功能,得到关键敏感特征参数与系统状态信号间的显式极限状态方程,将系统的可靠性概率约束转化为一个等价的确定型,避免了机电系统动态响应的多次遍历运算。再结合随机模拟,避开了各失效模式极限状态函数间复杂的相关性讨论。采用正交试验设计方法对机电系统进行参数灵敏度分析并精简样本。基于随机故障注入-神经网络法得到了关键敏感特征参数的改变对机电系统可靠性的影响规律,进而获得了参数的可靠性区间及失效临界值。为机电系统的可靠性分析和设计提供了参考和依据。
[Abstract]:Combining random fault injection method with neural network technology, a method for calculating the reliability of multiple failure modes in electromechanical systems is proposed.Taking the leakage fault inside and outside the hydraulic cylinder as an example, the virtual fault information is injected into the integrated simulation model of piston rod linear positioning system.Using the strong function approximation function of neural network, the explicit limit state equation between the key sensitive characteristic parameters and the system state signal is obtained, and the reliability probability constraint of the system is transformed into an equivalent deterministic type.The multiple ergodic operation of dynamic response of electromechanical system is avoided.Combined with stochastic simulation, the complex correlation among the limit state functions of each failure mode is avoided.The orthogonal design method is used to analyze the sensitivity of the parameters of the electromechanical system and to simplify the samples.Based on the stochastic fault injection-neural network method, the influence of the change of the key sensitive characteristic parameters on the reliability of the electromechanical system is obtained, and the reliability interval and the failure critical value of the parameters are obtained.It provides reference and basis for reliability analysis and design of electromechanical system.
【作者单位】：西北工业大学飞行器可靠性工程研究所;北京机械设备研究所;
【基金】：国家自然科学基金(51305350) 陕西省自然科学基础研究计划(2013JM6011)资助项目
【分类号】：TH17;TP183

【相似文献】