针对风机变速箱的劣化特性和维修策略问题,将变速箱的劣化过程看成一个Gamma过程,把维修方式和检测时间间隔作为决策变量,同时又考虑系统处于劣化状态下的劣化代价和停机代价的影响,建立基于半马尔可夫决策过程的状态维修优化模型。使用Q学习算法和策略迭代算法分别对该问题进行求解,得到最优维修策略。另外还通过对仿真数据结果进行分析,证明了该模型的可用性和经济性。%Considering the fan gearbox ’ s degradation and maintenance strategy , and through regarding degra-dation process as a Gamma process and taking both maintenance manner and detection time ’ s interval as deci-sion variables , as well as thinking of the influence of system degradation and downtime , the semi-Markov deci-sion process-based optimal maintenance strategy was established .Q-learning and policy iteration algorithm were employed to solve this problem so as to gain optimal maintenance strategy .Analyzing the simulation re-sults proves both practicality and feasibility of this model .
展开▼