带隐含退化过程非线性动态系统预测维护策略

Predictive maintenance policies for nonlinear dynamic systems with hidden degradation process

考虑带有隐含退化过程和不确定外界干扰的非线性动态系统的预测维护.根据系统的预测可靠性安全阈值M是否已知,提出了相应的预测维护策略来最小化系统的期望成本率.当M未知时,将其作为决策变量进行决策;当M已知时,充分利用预测可靠性及其变化趋势给出系统的维护策略.系统的替换依据监测的信息决定,从而使得维护策略对不同的工况有很好的适应能力.利用三容水箱模型进行数值仿真,充分验证了这两种维护策略的可行性和有效性.

Predictive Maintenance for nonlinear dynamic systems with hidden degradation and uncertain disturbance is considered in this paper.Two predictive maintenance policies are proposed to minimize the expected cost rate corresponding to whether the threshold of system predictive reliability M is given or not.When M is not given,it is recognized as a decision variable;when M is given,the predictive reliability and its development are both considered in maintenance policy.Replacement is decided by the monitoring information so that the maintenance policies have good adaptability for various working conditions.Numerical experiment is carried out in a three-tank vessels model to verify the feasibility and effectiveness of two maintenance policies.