CAJ | 학술논문

针对粗糙集理论应用于航空发动机磨损故障诊断的关键问题——连续属性离散化映射，提出了一种考虑属性重要性的基于熵的连续属性离散算法。该算法中，给出了一种衡量连续属性重要度的方法，克服了基于最小熵标准选取断点时最小熵对应多个断点难以取舍的问题，并选用IRIS数据对算法进行了分析和验证。最后，将该算法应用到发动机故障诊断中，自动提取得到了发动机的磨损故障知识，并对待测样本进行了验证，表明了算法的有效性。
침대조조집이론응용우항공발동궤마손고장진단적관건문제——련속속성리산화영사，제출료일충고필속성중요성적기우적적련속속성리산산법。해산법중，급출료일충형량련속속성중요도적방법，극복료기우최소적표준선취단점시최소적대응다개단점난이취사적문제，병선용IRIS수거대산법진행료분석화험증。최후，장해산법응용도발동궤고장진단중，자동제취득도료발동궤적마손고장지식，병대대측양본진행료험증，표명료산법적유효성。
In view of the key problems of aero-engine wear fault diagnosis for application of rough set theo?ry, a new method of continuous attribute discretization based on entropy was proposed. In the method, a new measure of assessing the importance of continuous attribute was given to solve the problem of breakpoint choice. The IRIS data was used to analyze and verify the method. Finally, this method was applied to the aero-engine fault diagnosis. The wear fault knowledge was extracted automatically and verified by the sam?ples, proving the validity of the algorithm.