论文部分内容阅读
引入了一套磨粒的形态学参数描述体系 ,实现了磨损颗粒图象的数值描述 ,并将灰色系统中的定权聚类技术用于磨损颗粒的自动识别 ,编制了相应的计算机模拟程序。在识别过程中根据某型航空发动机的滑油磨粒监测试验 ,确定了磨损颗粒各形态参数的聚类权值和灰类白化权函数。应用此方法对一组测试磨粒进行了模拟识别 ,识别正确率在 90 %以上 ,并且识别速度很快 ,大大优于传统的磨粒识别方法。
A set of description system of the morphological parameters of abrasive grains was introduced to realize the numerical description of the worn-out grain images. The fixed-weight clustering technique in the gray system was applied to the automatic identification of the worn grains, and the corresponding computer simulation program was compiled. In the identification process, according to the test of lubricating oil particles in a certain type of aeroengine, the clustering weight and the whitening weight function of ash particles were determined. Applying this method to simulate a group of test abrasive particles, the correct recognition rate is more than 90% and the recognition speed is very fast, which is much better than the traditional method of abrasive particle identification.