论文部分内容阅读
针对航空发动机健康评估过程中存在的多工况、非线性和小子样问题,提出了一种基于相似性度量的发动机健康评估的变精度粗集决策方法.首先给出了变精度粗糙集中正确分类率的确定原则和方法,提取了决策规则;其次,提出了基于交叉信息熵的属性相对重要度计算方法,并给出了属性的客观权重,结合专家经验和运行工况形成了属性的综合权重;最后,提出了基于属性综合权重的加权相似性度量方法,通过比较当前对象与约简后的决策系统中所有对象的相似程度,给出决策结果,完成对发动机的健康评估.实例表明,在结合专家经验和运行环境对评估结果进行修正后,该方法对发动机的健康评估更贴近实际,方法具有一定的普适性.
Aiming at the problems of multi-working conditions, non-linearity and small sub-samples in the process of aero-engine health assessment, a variable-precision rough set decision-making method based on similarity measure for engine health assessment is proposed. Secondly, the method of calculating the relative importance of attributes based on cross entropy is proposed, and the objective weight of attributes is given. Combined with the expert experience and operating conditions, the comprehensive weight of attributes is formed Finally, a weighted similarity measure method based on attribute comprehensive weight is proposed, and the similarity degree between the current object and all objects in the decision system after reduction is presented, and the decision result is given to complete the engine health assessment. Combined with the expert experience and operating environment to amend the assessment results, the method of the engine health assessment closer to reality, the method has a certain degree of universality.