【摘要】 机械加工过程优化控制,通过控制图模式智能识别过程。由于控制图是一种反映加工质量的工具,受多种因素影响,加工过程具有时变性、非线性等特点,传统线性方法无法识别特点,控制图识别正确率较低。为了提高了控制图识别正确率,将多种智能方法组合在一起,提出一种混合智能算法的控制图识别模型(WA-PCA-PSO-SVM)。首先采用小波变换对数据进行分解和重构,消除数据中的"噪声",然后采用主成分分析提取控制图样本的关键特征信息,降低分类器复杂度,最后采用粒子群算法优化SVM建立控制图分类器。仿真结果表明,WA-PCA-PSO-SVM可以准确的控制图变化规律,克服了传统模型的缺陷,提高了控制图模式识别的正确率,识别结果符合生产的实际情况。更多还原

【Abstract】 In the optimization control of machining process,chart pattern recognition by intelligent control was applied.Since the control chart is a tool which reflect the quality of machining,affected by many factors,and the process has a time-varying,nonlinear and other characteristics.The traditional linear methods cannot identify the features,and the correct identification rate of control charts is low.In order to improve the accuracy of control chart recognition,we combined multiple intelligence approach and proposed a control chart recognition model(WA-PCA-PSO-SVM) based on hybrid intelligent algorithm.Firstly,wavelet transform was used to make data decomposition and reconstruction and eliminate the"noise"in data.Then we used principal component analysis to extract the key features information of control charts sample and reduce the complexity of the classifier.Finally,particle swarm optimization SVM was applied to establish control chart classifier.Simulation results show that,WA-PCA-PSO-SVM can accurately control chart variation to overcome the shortcomings of traditional models and improve the accuracy of control chart pattern recognition.The recognition results are in accord with production of the actual situation.更多还原