基于改进PNCC的水泵机组声纹信号特征提取OA

As a piece of key equipment in pumping station,the water pump unit plays a vital role in agricultural and urban industrial water supply.The voiceprint feature extraction technology related to automatic speech recognition(ASR)is continuously improving and currently used in many fields.However,there are few studies on water pumps.In addition,when the units are operating on site,the signals are are susceptible to interference from other noises,easily leading to unstable extracted feature parameters.In order to better reflect the characteristics of the pump units during operation,this paper proposes a feature extraction method based on improved PNCC.This method extracts the energy spectrum envelope of the unit voiceprint signals,and screens out some irrelevant detailed features to obtain feature extraction parameters based on improved PNCC.The stability of characteristic parameters extracted by the improved PNCC method is compared with those by PNCC,MFCC,and GFCC methods.The results show that the characteristic parameters extracted by the improved PNCC method are more stable.