A methodology for fault diagnosis in large chemical processes and an application to a multistage flash desalination process: Part II

In Part I, an efficient method for identifying faults in large processes was presented. The whole plant is divided into sectors by using structural, functional, or causal decomposition. A signed directed graph (SDG) is the model used for each sector. The SDG represents interactions among process variables. This qualitative model is used to carry out qualitative simulation for all possible faults. The output of this step is information about the process behaviour. This information is used to build rules. When a symptom is detected in one sector, its rules are evaluated using on-line data and fuzzy logic to yield the diagnosis. In this paper the proposed methodology is applied to a multiple stage flash (MSF) desalination process. This process is composed of sequential flash chambers. It was designed for a pilot plant that produces drinkable water for a community in Argentina; that is, it is a real case. Due to the large number of variables, recycles, phase changes, etc., this process is a good challenge for the proposed diagnosis method.     

A fault diagnosis prototype for a bioreactor for bioinsecticide production

The objective of this work is to develop an algorithm for fault diagnosis in a process of animal cell cultivation, for bioinsecticide production. Generally, these processes are batch processes. It is a fact that the diagnosis for a batch process involves a division of the process evolution (time horizon) into partial processes, which are defined as pseudocontinuous blocks. Therefore, a PCB represents the evolution of the system in a time interval where it has a qualitative behavior similar to a continuous one. Thus, each PCB, in which the process is divided, can be handled in a conventional way (like continuous processes). The process model, for each PCB, is a Signed Directed Graph (SDG). To achieve generality and to allow the computational implementation, the modular approach was used in the synthesis of the bioreactor digraph. After that, the SDGs were used to carry out qualitative simulations of faults. The achieved results are the fault patterns. A special fault symptom dictionary —SM—has been adopted as data base organization for fault patterns storage. An effective algorithm is presented for the searching process of fault patterns. The system studied, as a particular application, is a bioreactor for cell cultivation for bioinsecticide production. During this work, we concentrate on the SDG construction, and 3btaining real fault patterns by the elimination of spurious patterns. The algorithm has proved to be effective in both senses, resolution and accuracy, to diagnose different kinds of simulated faults.     

基于关联规则分类的冷水机组故障诊断研究

鉴于现有基于数据驱动的故障诊断方法多以黑箱模型为主,诊断过程和结果难以解释的问题,本文提出一种基于关联规则分类的冷水机组故障诊断和故障作用机理解释的方法,在保证较好故障诊断精度的前提下,利用故障诊断模型中的规则库对诊断过程进行逆向分析,解析故障作用机理和模型的诊断过程,提升了基于数据驱动的故障诊断方法的可靠性。通过ASHRAE研究项目1043的实验数据对该方法进行验证。结果表明,基于关联规则分类的冷水机组故障诊断方法可以有效地识别7种典型冷水机组故障,平均故障诊断准确率高达90.84％。此外,提取的规则能够较好地吻合制冷原理及热力学相关知识,可用于故障作用机理分析与故障诊断的进一步研究。     

Rotor fault diagnosis system based on individual neural networks and fuzzy synthesized engine

Abstract

This paper proposes a robust fault diagnosis system for rotating machines adapting machine learning technology. The kernel of this diagnosis system includes a set of individual neural networks (INNs) and a fuzzy synthesized reasoning engine. First, the frequency characteristics from differential fault signals, including full spectrum, principal axis of full spectrum and slope of cascade, are used to feed into the INNs corresponding to assigned faults to emphasize the phenomenon of each fault. Especially, Taguchi’ s method is applied to quickly get optimal parameters of the above INNs. In the final step of the proposed diagnosis system, the evaluated indexes from each INN are synthesized by a fuzzy reasoning engine to identify the faults in the rotor system. Finally, five common faults of rotor systems, imbalance, misalignment, bow, whirl and whip, are generated from a rotor kit to verify the performance of this diagnosis system. The advantages of this diagnosis system are that the training epoch can be dramatically reduced, the over fitting problem can be avoided and diagnosis accuracy can be improved.     

大型风力机组远程智能监测与诊断系统的研究与开发

该文研究了大型风力机组的远程智能监测与诊断系统的关键技术问题,介绍了系统的开发情况。整个系统采用分布式架构,由数据采集与处理、实时数据存储、智能监测与诊断和人机交互4个子系统组成。智能监测与诊断子系统采用了知识库/推理机架构,推理机是一个自主开发的基于模糊Rete算法的模糊专家系统,知识库中存储了来源于风力机故障实验研究的常见振动故障的诊断知识。通过故障仿真,验证了整套系统的有效性。     

基于专家系统与神经网络集成的故障诊断的应用研究

本文针对工业生产中使用的直流电动机，应用人工智能的相关理论对其故障进行了广泛深入的研究。在此基础上，探讨了专家系统与人工神经网络相集成的电动机故障智能诊断方法并加以实现。实践证明，网络的学习时间显著缩短，整个系统的推理效率明显提高，并验证了集成式专家系统的诊断效果比传统的专家系统或神经网络更为全面、准确和迅速。电动机故障的集成式智能诊断方法是一个既有理论研究意义又有实际使用价值的课题与方向。     

基于模糊神经网络的机载设备智能故障诊断系统

针对某型飞机机载设备故障多,且具有模糊性、复杂性的特点,本文将模糊逻辑和神经网络相结合,采用模糊隶属函数来描述这些故障的程度,建立了模糊神经网络故障诊断模型.采用图形化编程技术,开发了一种故障诊断推理流程图,方便了用户的开发.该系统依据专家知识和测试数据,可将故障隔离到内场可更换单元(SRU)或某个功能电路.实践证明该诊断系统是有效的,具有推广应用价值.     

基于专家系统与神经网络集成的故障诊断的应用研究

本文针对工业生产中使用的直流电动机,应用人工智能的相关理论对其故障进行了广泛深入地研究。在此基础上,探讨了专家系统与人工神经网络相集成的电动机故障智能诊断方法并加以实现。实践证明,网络的学习时间显著缩短,整个系统的推理效率明显提高,并验证了集成式专家系统的诊断效果比传统的专家系统或神经网络更为全面、准确和迅速。电动机故障的集成式智能诊断方法是一个既有理论研究意义又有实际使用价值的课题与方向。     

非平稳信号特征提取在航空发动机故障诊断中的应用

在航空发动机早期故障诊断中,特征提取是早期诊断的重要过程之一.文中以航空发动机转子故障为研究对象,给出了基于经验模式分解、小波分析为核心的故障特征提取方法,并作了针对性的比较研究.在matlab7.0环境下开发了一个故障特征提取软件系统.研究结果表明:基于经验模式分解的时频分析方法可以很有效地提取到非平稳故障特征信号,是一种适合于非线性信号处理的方法.     

Signal interpretation for monitoring and diagnosis, a coolingsystem testbed

This paper discusses a method for fault detection and isolation (FDI) in continuous dynamic systems. A key aspect of this approach is the coupling of a qualitative diagnosis engine and a monitoring system that computes symbolic feature values through a signal-to-symbol transformation on the continuously sampled measurement data. Signal analysis techniques with a sound statistical basis are employed to generate reliable symbolic data. The methodology is evaluated on the diagnosis of engineered faults in the cooling system of an automobile engine that has been instrumented with temperature and pressure sensors. Results show the interdependency between modeling for diagnosis and the feature extraction system