传统的遗传算法存在早熟现象严重和局部搜索精度较低的固有缺陷，容易导致分析结果与实际情况不相符，不能很好地用于船用核动力装置概率因果故障诊断。提出了一组综合改进策略，首先定义了奇异个体判断指标；而后设计了一种自适应交叉、变异策略和自适应局部搜索策略，并通过经典案例测试改进算法的有效性；最后构建改进算法与概率因果故障诊断模型，进行船用核动力装置故障诊断实例分析。分析结果对船用核动力装置故障诊断具有重要的指导意义，改进遗传算法是进行船用核动力装置故障诊断有效而实用的方法。%The traditional genetic algorithms have inherent defects such as serious prematurity phenome⁃non and low accuracy in local search,which may cause disagreement between analytic results and practi⁃cal situations. Therefore,they cannot be well applied to the fault diagnosis of probabilistic causal models for marine nuclear power plant. Aiming at the problem,this paper presents a series of comprehensive im⁃provements for the traditional genetic algorithms. Firstly,the judgment index of singular individuals is de⁃fined and a self-adaptive crossover,mutation and local search strategies are developed. Secondly,the va⁃lidity of the improved algorithm is tested against a classic case. Finally,the improved algorithm together with a probabilistic causal fault diagnosis model is constructed for marine nuclear power plant. The analy⁃sis result is of great significance,since the improved genetic algorithm is proved to be an effective and practical way to perform the fault diagnose for marine nuclear power plant.
展开▼
