本論文主要目為在一般辦公室環境噪音下，建立一套可判別風扇是否為良品之儀器設備。由某電子公司提供於無響室所測得五種不同型號風扇之相關資料，如：聲音檔、時域訊號、頻域訊號及風扇頻譜圖，觀察各風扇頻譜圖並配合人耳所聽到的聲音與心理感受，建立不同型號風扇異音判別指標，且整合找出風扇通的用判別指標。 接著實際進行風扇量測，由兩支麥克風，一支量測風扇聲音，一支量測背景噪音，將各別所測得之時域訊號進行訊號處理，並搭配本文所提出之排除背景噪音方法後呈現出風扇頻譜圖，同時也探討縮短量測秒數之風扇實驗結果。將所有訊號處理過程及風扇判別指標撰寫入LabVIEW程式，並用燈號表式是否通過各判別指標，最後實際呈現LabVIEW程式對風扇之量測結果。 最後以本文撰寫完成之LabVIEW程式實際量測風扇，扣除擷取訊號秒數，從訊號處理到燈號呈現出風扇在各判別指標為OK或NG僅花費2到3秒，且使用儀器量測是經由訊號處理完後轉成風扇頻譜圖，再經由所選取之判別指標，故能判別出人耳所不能辨別的異音，提升風扇品質。 ;The purpose of this research is to establish the measuring equipment, used within an environment with its level similar to an office, which can determine if a fan is in good condition. The company provided the measurement data, in an anechoic chamber, for five different models of fans. Based on the frequency domain data, one can choose several discrimination indices for examining the Fans. The corresponding critical values of the indices can also be determined. The goal of designing the equipment is to measure the fan noise in a quiet room and make the decision between fair or foul in a short period. Two microphones were used. One measures the sound from a fan and the other one measures the background noise. The time domain signals then were processed by a LabVIEW program for signal processing. Based on the technique proposed in this research, the processed signal can remove the background noise. Meanwhile, the fan experimental results of the shortened time of measurement would be discussed. All signal processing formulation and the fan discrimination indices were written in a LabVIEW program which shows lights to indicate whether the fans pass the examination. The entire processing time takes about 2-3 seconds plus the measurement time. This equipment can improve the quality of quality control since it distinguishes the sound humans’ ears cannot tell.
