作　　者： ; ; ; ;

机构地区： 中国科学院沈阳自动化研究所工业信息学重点实验室

出　　处： 《仪器仪表学报》 2009年第4期711-716,共6页

摘　　要： 鉴于多重分形理论在定量描述复杂系统非线性运行规律方面具有的独特优势,将多重分形理论引入到工况特征识别研究中来,确认了水泥回转窑窑电流信号的多重分形特性。在此基础上,研究了窑电流多重分形谱参数随工况变化的情况,发现多重分形谱参数的变化趋势与回转窑内工况状态的变化趋势之间具有较强的关联性,进而提出了基于多重分形谱参数进行水泥回转窑异常工况特征提取的新方法,为水泥生产过程中工况状态的在线监控和预报提供了有力支持。 Considering the advantages of multifractal method in the field of complexity researches, the multifractal theory was introduced to the research of working condition recognition. The muhifraetal characteristics of cement rotary kiln current were confirmed using multifractal analysis. Furthermore, a time-dependent muhifractal analysis of cement rotary kiln current was performed using muhifractal method. It is found that there is a strong correlation between the muhifraetal spectrum parameters of kiln current and the working condition. And it is found that when the working condition appears abnormal, a strong variability is clearly characterized by the multifractal spectrum parameters. Therefore, a new method of feature extraction is developed to identify the working condition of cement rotary kiln current using multifractal analysis. This will lay a foundation for online monitoring and forecasting.

关 键 词： 非线性 多重分形 窑电流 工况识别 特征提取

领　　域： [自动化与计算机技术] [自动化与计算机技术]