Modeling for reliability optimization of system design and maintenance based on Markov chain theory
Modeling for reliability optimization of system design and maintenance based on Markov chain theory
复制标题
DOI：
10.1016/j.compchemeng.2019.02.016
复制
发表时间：
2019-05-08
影响因子：
4.3
通讯作者：
Ramaswamy, Sivaraman
中科院分区：
文献类型：
--
作者：
Ye, Yixin;Grossmann, Ignacio E.;Ramaswamy, Sivaraman
本文提出了一个MINLP模型，表示系统故障和维修的随机过程作为一个连续时间的马尔可夫链，在此基础上，它优化冗余的选择和频率的检查和维护任务的利润最大化。该模型明确地说明了系统的每一种可能状态。提出了有效的分解和情景约简方法。解决了一个小的例子，两个处理阶段，以证明将维护考虑的影响。一个分解方法和一个场景减少方法应用到这个例子中，并大大减少了计算工作量。一个更大的例子，四个阶段，这是不能直接求解，也成功地解决了使用所提出的算法。最后，我们表明，所提出的模型和算法是能够解决实际问题的基础上的空气分离过程的例子，激励我们的工作，其中具有多个阶段，潜在的单位和故障模式。(C)2019爱思唯尔有限公司版权所有。
This paper proposes an MINLP model that represents the stochastic process of system failures and repairs as a continuous-time Markov chain, based on which it optimizes the selection of redundancy and the frequency of inspection and maintenance tasks for maximum profit. The model explicitly accounts for every possible state of the system. Effective decomposition and scenario reduction methods are also proposed. A small example with two processing stage is solved to demonstrate the impact of incorporating maintenance considerations. A decomposition method and a scenario reduction method are applied to this example and are shown to drastically reduce the computational effort. A larger example with four stages, which is not directly solvable, is also successfully solved using the proposed algorithm. Lastly, we show that the proposed model and algorithms are capable of solving a practical problem based on the air separation process example that motivated our work, which features multiple stages, potential units and failure modes. (C) 2019 Elsevier Ltd. All rights reserved.