基于灵敏度分析的系统可靠性稳健分配优化方法

在系统可靠性分配中,考虑单元可靠度的不确定性已是可靠性分配的现实需要.为了提高系统可靠性分配优化的质量,将稳健理论引入可靠性分配中,提出基于单元可靠性灵敏度的系统可靠性稳健分配方法.将单元可靠性灵敏度溶入系统可靠性分配模型之中,建立系统可靠性稳健分配模型.在此基础上,采用粒子群-序列二次规划算法对该模型进行优化设计,该混合算法既保持了粒子群算法全局收敛的特点,又补充了序列二次规划法精确求解的能力,因此该混合算法可以快速获取全局最优解.通过对发动机曲柄连杆机构进行可靠性稳健分配设计,验证了可靠性稳健分配模型的合理性和混合算法的寻优能力.对结果分析表明,所提方法可以较好解决单元可靠度不确定时的可靠性分配问题,混合算法具有较强的全局搜索能力,分配优化结果具有较强的稳健性.

Robust allocation optimization method for system reliability based on sensitivity analysis

It is the practical need of reliability allocation to consider the uncertainty of the unit reliability in reliability distribution of system. In order to improve the quality of system reliability allocation, a robust allocation method for system reliability based on unit reliability sensitivity is presented through introducing the robust theory into reliability allocation. The allocation model for system reliability is established after incorporating unit reliability sensitivity into system reliability allocation model. On this basis, the particle swarm optimization-sequential quadratic programming algorithm is used to optimize this model. The hybrid algorithm can maintain not only the characteristics of the global convergence of the particle swarm algorithm, but also the ability of accurate solution of sequential quadratic programming algorithm. Therefore, the global optimal solution can be obtained quickly with this hybrid algorithm. The design of robust reliability allocation for engine crank rod mechanism is carried out, which verifies the rationality of reliability allocation model and optimal capability capacity of hybrid algorithm. The analytical results showed the strong global search capability of hybrid algorithm and robustness of allocation optimization results, and then, the problem of reliability allocation could be solved with the proposed method when the reliability of unit is uncertain.