基于时频分析与神经网络的实时智能故障诊断系统的软件设计--总体构架与实时数据采集功能的开发

介绍了以振动/声响信号为传感器的采集对象,基于小波(包)及联合时频信号处理和模糊神经网络分类的多用途智能故障诊断系统中的实时数据采集、信号分析与识别诊断等模块的程序设计与实现.作者所开发的实用软件包在现代汽车实时故障诊断中的成功应用证明了其理论与软件开发的正确性和有效性.这里着重介绍智能故障诊断系统的软件框架的设计以及实时数据采集部分的设计、实现方法.

Software Design of Real-Time Fault Diagnosis System Based on Joint Time-Frequency Analysis and Fuzzy Neural Networks I. The Total Software Frame and Implementation of Real-Time Data Collection Module.

This paper introduces an approach combining time-frequency analysis and fuzzy neural networks for pattern recognition and intelligent fault diagnosis based on real-time vibration signal processing. We focus on the software developing of real-time data collection, signal analysis and classification in fault diagnosis software system, which is based on the vibration or sound signal, by means of various feature extraction methods such as wavelet transform, STFT and Wigner-Ville distribution function, and fuzzy neural networks for signal classification and fault recognition. This real-time diagnosis system can be used in modern automobile intelligent diagnosis, human electrocardiogram-based medical diagnosis and speech recognition, and so on. The practical application on automobile real-time diagnosis verifies the correctness and effectiveness of our theories and programs. This paper emphasizes the total software frame of the fault diagnosis system, and the implementation of its real-time data collection module.