定性趋势分析在故障诊断中的应用研究系统工程专业论文.docxVIP

摘要定性趋势分析在故障诊断中的应用研究 摘要 定性趋势分析在故障诊断中的应用研究 摘要 过程监测和快速故障检测与诊断对于过程操作的有效管理非常重要， 而定性趋势分析(Qualitative Trend Analysis，QTA)是故障检测与诊断的一 种重要的技术。趋势分析是状态监测和故障诊断的重要一环，趋势能够反 映重要参数的运行状态，发展速度和趋势，确定故障的出发时刻，预测早 期故障的发生，为故障的早期判定提供一个有效的手段。用计算机对传感 器数据的瞬态演化进行建模，定性趋势分析(QTA)的基本思想是把已测 的信号作为一种趋势序列(这种趋势序列是基于基元的)，定义三种基元： {不变，上升，下降)，用这三个基元组合成七种与时间有关的形状来描述 过程的趋势：A不变，B上升，C下降，D正步，E负步，F上升／下降瞬 变，G下降／上升瞬变。本文系统介绍用于过程故障诊断的定性趋势分析 (Qualitative Trend Analysis，QTA)两个重要算法：趋势提取算法和趋势 识别算法。本文用最小二乘法，实时地从数据中捕获并识别数据的趋势。 为了使得趋势提取算法和趋势识别算法有更大的适应性，采用了分级阈值 的办法来动态地获得趋势提取和识别过程中所需要的阈值。获得过程的趋 势后，用一种趋势匹配的策略构架来诊断故障。在研究实例中应用定性趋 势分析，并针对趋势分析故障匹配分辨率不高的不足，引入了定量的信息， 从而提高故障匹配分辨率。对于过程工业的故障诊断，定性趋势分析的优 点是能够简洁、快速地诊断故障。 I 关键词：定性趋势分析(QTA)；趋势提取；趋势识别；故障检测与诊断Ⅱ 关键词：定性趋势分析(QTA)；趋势提取；趋势识别；故障检测与诊断 Ⅱ APPLIED APPLIED RESEARCH OF QUALITATIVE TREND ANAI YSIS IN PRoCESS FAUl月DIAGNOSIS ABSTRACT Process monitoring and quick fault detection and diagnosis are very important for effective management of process operations management，and Qualitative Trend Analysis is one of the important techniques of fault detection and diagnosis．Trend analysis is very important for fault detection and diagnosis，trend can reflect the running states，the development speed and trend of the important parameters，make sure the starting time of the fault， predict the early occurrence of the fault．Qualitative process fiend representation is a useful approach to model temporal evolution of sensor data by computer,Trend base on the three primitives：{Steady,Increasing， Decreasing}，using these seven shapes combined by three primitives to describe the process trend．Two important algorithm of Qualitative Trend ^ Analysis(QTA)for process fault diagnosis is systematically introduced trend extraction algorithm and trend identification algorithm．To make the algorithm have a larger adaptability,the method of grading threshold is introduced to achieve the threshold dynamically required in trend extraction and identification．A framework of trend match methodology is used to diagnosis faul