【摘要】 研究利用小波变换提取过程检测信号的频带分布特征，并提出一种新的模式分类算法用于对提取的特征归纳分类，并将分类结果用于实际过程的工况监测.应用实例研究证实了所提方法的可行性和有效性．更多还原

【Abstract】 The problem of process monitoring was defined by using the principle of pattem recognition.After the analysis of time-frequency character of discrete wavelet transform,a means to extract the features of frequency band in real process measurement siwgnals based on discrete wavelet transform was stru ied．Meanwhile a new pattem classification algorithm was proposed to induce the relationship of the extracted features and corresponding process conditions.The induction result was applied to process monitoring,which verified the feasibility and effectiveness of the method.更多还原