本文研究使用Bayes分類方法，分析平面磨床振動信號，由研磨加工過程中之砂輪主轴振動監測完工品質狀態。以信號的統計參數萃取出分佈狀態之特徵值，包含波形、峰值、裕度、歪度及峭度，將特徵值建立於訓練様本中，計算其 協方差矩陣作為分類工具並建立完工品質狀態之訓練様本，進行完工品質狀態分類。
研磨加工實驗結果可區分出不同狀態下之研磨完工品質，包括鏡面、霧面、顫紋、刮傷及燒傷，利用Bayes分類器作為研磨完工狀態判別，可達到品質監測之目的。

In this paper, Bayes classifier method is used to analyze vibration signal from surface grinder and the finished quality is also monitored through process of grinding wheel spindle vibration. Based on Bayes classifier, the extraction of statistical parameters from the signal characteristics in the distribution of state values including waveform, peak value, margin value, skewness, and kurtosis is established to the training samples to calculate its covariance matrix which plays as classification tools for the establishment of the state of finished quality.
Experimental results could be distinguished among the different states of finishing quality of grinding and classified into mirror, matte, chatter, scratch, and burning by using Bayes classifier method for their quality monitoring.