• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2016, Vol. 52 ›› Issue (15): 204-212.doi: 10.3901/JME.2016.15.204

• 制造工艺与装备 • 上一篇    

切削加工表面粗糙度的多维多规则云预测方法*

阴艳超1, 丁卫刚2   

  1. 1. 昆明理工大学机电工程学院 昆明 650500
    2. 太原太航科技有限公司 太原 030006
  • 出版日期:2016-08-05 发布日期:2016-08-05
  • 作者简介:

    阴艳超,女,1977年出生,博士,副教授,硕士研究生导师。主要研究方向为智能制造、制造业信息化、知识工程。

    E-mail:yinyc@163.com

  • 基金资助:
    * 国家自然科学基金(51365022)和云南省教育厅重点(2014Z032)资助项目; 20150818收到初稿,20160526收到修改稿;

A Novel Cloud Model Prediction for Surface Roughness Based on Multidimensional & Multi-rules Reasoning

YIN Yanchao1, DING Weigang2   

  1. 1. Faculty of Mechanical and Electrical Engineering, Kunming University of Science and Technology,Kunming 650500
    2. Taiyuan Taihang Technology Co.,Ltd., Taiyuan 030006
  • Online:2016-08-05 Published:2016-08-05

摘要:

针对目前常用切削加工表面粗糙度预测方法存在预测精度不高、泛化能力不强的问题,提出一种融合模糊和随机性的粗糙度云预测新方法。在分析大量试验数据的基础上给出了多维预测云的数字特征和数学模型,设计粗糙度预测的多维多规则定性推理发生器,将各切削用量作为前件云输入,分析其对粗糙度后件云的影响关系,并通过对推理规则的多元组合,实现不同加工工艺规范下对工件表面质量的精准预测,揭示加工表面质量随切削参数变化的规律。试验结果表明,在相同样本条件下所提方法的平均相对预测误差低于4.78%,求解速度和预测精度方面均有所提高,且预测范围更加广泛。

关键词: 多规则推理器, 精准预测, 云模型, 粗糙度

Abstract:

:Aiming at the problems of lower prediction accuracy and narrower prediction range for the common prediction method, a new surface roughness prediction method based on multidimensional & multi-rules reasoning of cloud model is put forward. Based on analyzing a large quantity of test data, the digital characteristic of a multidimensional cloud is represented and the generator on multidimensional & multi-rules for qualitative reasoning is designed firstly. Then through combining the reasoning rules in various modes, the accurate surface roughness prediction is realized with the cutting speed, feed rate and cut depth as the input conditions and the predicted roughness values as output, thus the variation law of quality of machined surface following milling parameters can be obtained. Finally, experimental results show that the presented method is more accurate with the average relative prediction error below to 4.78% in the same conditions. Compared with other prediction models, the presented model makes the information classification and accuracy prediction more precise, and the predicted range wider.

Key words: accuracy prediction, cloud model, multi-rules reasoning generator, roughness