摘要
研究了利用小波变换提取过程检测信号变化的多尺度定性特征 ,分析和证明了提取的特征具有稳定性和表征工况区域的完备性 .提出一种多尺度因果推理模型 ,用于对提取的特征进行因果推理分析 .应用实例研究证实了所提出方法的可行性和有效性 .
Process monitoring is one of the most important problems in modern industrial processes. How to extract the operating features from the measurement data, determine the change of operating region according to the extracted features and infer its effect in real time is the key problem of process monitoring.Extraction of the qualitative features from process measurement signals with wavelet transform has been studied and a model based approach for casual reasoning is proposed to deduce the causality of the extracted qualitative features. The deductive results are applied to real process monitoring. The theoretical analysis and application results show that the proposed method can effectively describe the process behavior knowledge in multi scale and eliminate the virtual character of expert system in describing the time states. Therefore the proposed method improves the reliability of real time process monitoring system.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2000年第6期746-750,共5页
CIESC Journal
基金
国家自然科学基金!(No .6 97740 2 2 )
中国博士后科学基金
关键词
过程监测
小波变换
特征提取
因果推理
化工生产
process monitoring, wavelet transform, feature extraction, casual reasoning
