基于状态条件概率的设备剩余寿命预测

Remaining useful life prediction for equipment based on conditional probability of states

针对机载设备剩余使用寿命预测中存在的不确定性因素,建立了基于状态条件概率分布的机载设备剩余寿命模型.首先,引入状态条件概率矢量对隐马尔科夫模型(HMM,Hidden Markov Model)进行不确定性改进,并推导了其计算形式.其次,给出了近似确定状态退化转移时间的方法,由此得到了以状态条件概率矢量为协变量的条件可靠度函数及剩余寿命模型.最后,以飞机发动机温控放大器为应用对象进行仿真计算.仿真结果表明该模型预测精度高,能够较大程度地降低不确定性因素的影响.

To conquer the question of uncertain factors in the remaining useful life（ RUL） prediction of airborne equipment,a RUL model was presented based on the conditional probability distribution of states. Firstly,the conditional probability vector of states was introduced to improve the ability of hidden Markov model（ HMM） to deal with uncertainties,and its computation form was deduced. Secondly,a method was presented to approximately determine the degradation transfer time of states,thus the conditional reliability function and RUL model were obtained,in which the conditional probability vector of states was taken as the covariate. Lastly,the aero-engine temperature control amplifier was applied to validate the model. The results show that the prediction precision of this model is high,and this model can reduce the influence of uncertain factors greatly.