摘要
利用多传感器信息融合技术 ,通过模糊分类的方法对不同的磨削条件进行模糊化处理 ,构建了砂轮钝化监测多传感器融合系统结构 ;应用BP神经网络将磨削过程中声发射、磨削力和功率传感器信号合理融合 ,提出了自适应变学习率策略 ,将其神经网络输出的信号特征值作为表征砂轮钝化状态识别的判据 ,进行了砂轮钝化监测实验·结果表明 ,使用多传感器信息融合方法比使用单一传感器方法识别率高 ,监测效果好 。
The multi sensor fusion system was established to monitor the passivation status of grinding wheel. BP neural network was used to interfuse reasonably the acoustic emission signal, grinding force signal and power sensor signal. The output eigenvalue of the neural network was considered as a criterion to distinguish the passivation status of grinding wheel. A self adaptive learning strategy was brought forward. A grinding wheel passivation experiment was done. The multi sensor infusion method can get higher recognition rate and better monitoring effect than a single sensor. In this way, intelligent monitor can be realized and the grinding wheel can be dressed in time.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2003年第3期248-251,共4页
Journal of Northeastern University(Natural Science)
基金
教育部科学技术研究重点资助项目 ( 2 0 0 3 2 )
关键词
多传感器融合
砂轮钝化
模糊分类
BP神经网络
智能监测
磨削
multi sensor fusion
wheel passivatio
fuzzy classifying
BP neural network
intelligent monitoring
grinding