摘要
状态评级是柴油机状态监测的主要目的。常规的以特征参数变化倍数评级的方法对于经常拆迁的设备效果不理想;而模糊C(center)-均值聚类算法不能自动对聚类结果进行等级排序。文中提出的波动法与模糊C-均值聚类相结合的状态评级则有效地解决了上述问题。波动法原理为柴油机各缸的特征参数波动越小,则整机状态越好。选取与柴油机状态密切相关的7个参数组成特征向量,用现场采集的PZ12V190柴油机的35个样本建立聚类标准;将另10台柴油机与标准逐一再聚类,其结果与实际情况吻合得很好。表明该方法对多缸柴油机状态评级的有效性和实用性。
The condition evaluation is the main purpose of condition monitoring of diesel engine. The usual method evaluates the condition using the multiple of current parameters to the standard ones, which is not ideal for the equipments needing disassembling, moving and reassembling frequently. On the other hand, fuzzy C（center）-means clustering can not automatically sort the diesel engines by their conditions. A new method integrated with fluctuation method and fuzzy C-means clustering was put forward and solved the above difficult problems. The main principle of fluctuation method is that the fluctuation of characteristic parameters among cylinders is lower, the condition of the diesel engine is better. Then 7 parameters closely relative to the condition were selected to compose characteristic vector and 35 samples of model PZ12V190 diesel engine were collected to establish clustering criterion. The other 10 diesel engines were respectively clustered with the criterion and the results accorded with the actual condition very well, which shows that the integrated method is valid and practicable for the condition evaluation of multi-cylinder diesel engine.
出处
《机械强度》
EI
CAS
CSCD
北大核心
2005年第5期567-570,共4页
Journal of Mechanical Strength
基金
国家自然科学基金资助项目(50105015
50375103)
北京市科技新星基金资助项目(2003B33)。~~
