摘要
为准确评估滚珠丝杠副性能的退化程度,提出基于量子遗传算法和灰色神经网络的滚珠丝杠副性能退化评估方法。以CINCINNATIV5-3000加工中心的滚珠丝杠副为研究对象,设计了丝杠在线监测系统,利用动态聚类数据处理技术对采集的海量数据进行预处理,提取信号的时域、频域及时频域特征,通过主分量分析方法压缩特征数量,构建了丝杠振动信号特征向量,采用量子遗传算法优化灰色神经网络的初始化参数,将特征向量输入到灰色神经网络进行训练,进而得到丝杠性能退化模型。实践运行结果表明,所建立的丝杠性能退化模型能够有效评估数控机床的丝杠的性能,研究成果具有重要的工业推广价值。
To evaluate the performance of ball screw, the screw performance degradation assessment method based on quantum genetic algorithm and grey neural network was studied. An online monitoring system was designed based on the ball screw of CINCINNATIVS-3000 machining center. The mass data were processed by dynamic clustering data processing technology. Time domain, frequency domain and time-frequency domain features were extracted, and the screw vibration signal feature vectors were constructed by compressing the feature data with principal com- ponent analysis method. Initialization parameters of grey neural network were optimized by means of quantum genet- ic algorithm. The feature vectors were inputted to gray neural network for training, and the screw performance deg- radation model was obtained. Practice results showed that the screw performance degradation model could effectively evaluate the performance of NC machine screw, while it was with industrial prospects.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2015年第5期1293-1300,共8页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(51275426)~~
关键词
丝杠
性能
灰色神经网络
遗传算法
量子
screws
performance
grey neural network
genetic algorithms
quantum
