摘要
提出了一种有效的磨粒类型识别方法,该方法除了选用传统的磨粒形态特征参数,将表面粗糙度和表面纹理指数也作为重要的磨粒识别参数,选用面积、长度、圆度、纤维比率、体态比、边界分形维数、表面粗糙度Sa、Sq和表面纹理指数(Stdi)等9个参数,采用人工神经网络来识别磨粒类型,应用示例表明效果良好,提高了磨粒类型识别的精确度。
The recognition of types of wear debris is the key objective of wear debris analysis. An effective approach to recognize types of wear debris was proposed. In addition to conventional boundary morphology parameters of wear debris, surface roughness parameters and surface texture index were adopted as important parameters. Total 9 input parameters, area, length, roundness, fiber ratio, aspect ratio, boundary fractal dimension, surface roughness parameters （ Sa, and Sq ） and surface texture index （Stdi）, were used to be as input parameters of BP neural network to recognize types of wear debris. The application example demonstrates that the developed method can be used to recognize types of wear debris precisely.
出处
《润滑与密封》
CAS
CSCD
北大核心
2007年第3期21-23,46,共4页
Lubrication Engineering
基金
湖北省自然科学基金项目(2006ABA315)
关键词
磨粒
磨粒类型
识别
表面纹理
wear debris
types of wear debris
recognition
surface texture
