论文部分内容阅读
采用实例推理和遗传算法相结合的方法,研究了航空机械零部件失效模式识别的问题。对用于识别的失效属性的选择、检索相似度计算、训练用遗传算法的适应度函数设计以及训练策略的影响进行了较为详细的描述。应用测试表明,对包含分布均衡的3种模式的情况取得了高于74.67%的识别率,所获得的最佳权值向量对另外2种模式具有很好的识别精度(大于73.3%),对混合多模式情况也具有较好的推广能力。验证了该方法对航空零部件失效模式的识别是可行的。
The combination of case-based reasoning and genetic algorithm is used to study the problem of failure mode identification of aeronautic machinery components. The selection of failure attributes for identification, the similarity calculation of retrieval, the fitness function of training genetic algorithm, and the influence of training strategy are described in detail. The application tests show that the recognition rate of more than 74.67% is achieved for the three modes with balanced distribution, and the best weight vector obtained has good recognition accuracy (more than 73.3%) for the other two modes. Mixed multi-mode situation also has good promotion capabilities. It is verified that this method is feasible to identify the failure modes of aerospace components.