Fault Diagnosis Task Research Articles

Diagnóstico de faltas em sistemas de potência: definição do problema e abordagens via inteligência artificial

O constante aumento na demanda, sem o correspondente aumento na capacidade de geração e transmissão de energia elétrica, tem levado os sistemas de potência a operarem cada vez mais próximos dos seus limites. Este fato torna imprescindível o uso de ferramentas de apoio à operação, capazes de aumentar a confiabilidade do sistema, reduzindo o número e/ou duração de paradas não programadas das instalações. Na tentativa de reduzir a duração das interrupções, têm sido publicados trabalhos envolvendo ferramentas desenvolvidas para uso na operação em tempo real, com a finalidade de dar apoio ao diagnóstico de faltas. Tais ferramentas seriam empregadas após contingências com desligamentos definitivos, antes de dar início ao processo de recomposição da rede. Este artigo discute aspectos importantes do diagnóstico de faltas e as técnicas mais utilizadas para modelar e tratar este problema.

EVALUATING THE EFFECTIVENESS OF KNOWLEDGE-RETRIEVAL AIDS FOR FAULT-DIAGNOSIS TASKS

ABSTRACT The mechanism of knowledge retrieval is the key factor regarding whether users can efficiently apply a knowledge-based system. The effects case-based reasoning (CBR) retrieval and traditional key-term knowledge retrieval have on the performance of fault-diagnosis tasks were compared. The study was conducted using a split-plot factorial design: Two knowledge-retrieval aids by two problem complexities. Twenty subjects participated in the experiment and were randomly assigned to one of two groups. ANOVA results show that the fault-diagnosis times of the subjects in the CBR knowledge-retrieval group were significantly shorter than those of the key-term knowledge-retrieval group. In addition, the correct fault-diagnosis rates of the CBR knowledge-retrieval group were significantly higher than those of the key-term knowledge-retrieval group. The differences were more significant when the diagnosis tasks were of higher degrees of complexity. The results demonstrate that the CBR retrieval aid can improve the performance of fault-diagnosis tasks. It is therefore suggested that algorithm-based knowledge browsing and retrieval aids should be included in the development of knowledge-based systems to support users in interacting with these systems.

The effects of presenting functionally abstracted information in fault diagnosis tasks

Abstract With respect to the design of visual information display (VID) for process control, this study experimentally evaluated the effectiveness of functionally abstracted information in the task of fault diagnosis. The benefits of functional properties of work domain have been emphasized by ecological interface design (EID), a relatively new design framework for human–machine interfaces. According to the concept of EID, multilevel information representation based on abstraction hierarchy of work domain is expected to be advantageous for supporting the operator's problem solving. To investigate the advantage of EID application, an experiment was conducted using a computer-based simulation of the secondary cooling system of nuclear power plants. Three interfaces were compared: the first representing only the physical properties of the process, the second representing purpose-related generalized functions (GFs) in addition to the physical properties, and the third representing abstract functions (AFs) governing the GFs in addition to the physical properties. The results showed that the diagnostic performance was improved by displaying functionally abstracted information at both levels, and that the usefulness of the abstract information was dependent on the complexity of the diagnosis problems.

Sensor validation and process fault diagnosis for FCC units under MPC feedback

Abstract Model predictive control (MPC) has been widely applied in industry, especially in the refining industry. As all feedback controllers require correct sensor measurements, unreliable sensors can cause the MPC controller to move the process in an erroneous manner. Data validation of sensor measurements is a prerequisite in applying advanced control, particularly multivariable control which depends on many sensors. However, little research work is available on how feedback controllers like MPC complicate the task of sensor validation and process fault diagnosis. In theory, a controller can transfer the effect of a sensor fault in a controlled variable to the manipulated variables. In this paper, principal component analysis (PCA) is applied to detect, identify and reconstruct faulty sensors in a simulated FCC unit. A base PCA model is generated by perturbing the process throughout the operating region. Performance of MPC with and without data validation is compared. The same base PCA model is applied to detect and identify dynamic process faults. We demonstrate that process faults can be detected and diagnosed at an early stage.

A self-validating control system based approach to plant fault detection and diagnosis

An approach is proposed in which fault detection and diagnosis (FDD) tasks are distributed to separate FDD modules associated with each control system located throughout a plant. Intended specifically for those control systems that inherently eliminate steady state error, it is modular, steady state based, requires very little process specific information and therefore should be attractive to control systems implementers who seek economies of scale. The approach is applicable to virtually all types of process plant, whether they are open loop stable or not, have a type or class number of zero or not and so on. Based on qualitative reasoning, the approach is founded on the application of control systems theory to single and cascade control systems with integral action. This results in the derivation of cause–effect knowledge and fault isolation procedures that take into account factors like interactions between control systems, and the availability of non-control-loop-based sensors.

A genetic algorithm approach to measurement prescription in fault diagnosis

To fully discriminate among all possible diagnoses in a fault diagnosis task, one needs to take measurements from the system being diagnosed. The primary effects of taking one measurement in diagnosis based on first principles were presented in A. Reiter [Artificial Intelligence (32) (1987) 57–95] and a more detailed, formal account was given in A. Hou [Artificial Intelligence (65) (1994) 281–328]. However, the order in which measurements are to be taken is an issue. We propose a genetic algorithm to determine a good measurement order for a diagnosis task. The method applies operators such as selection, crossover, and mutation to evolve an initial population of measurement sequences. The quality of a measurement sequence is evaluated based on the cost taken for the measurement sequence to find the final diagnosis. Experiments on testing circuits have shown that the quality of measurement sequences is greatly improved after evolution.

A Fault Detection And Diagnosis Methodology For Chemical Processes

A fault detection and diagnosis system, which has been design to cope with transient behaviours of the process, abrupt and incipient faults, is described in this paper. The system consists of a knowledge-based system coupled with a fuzzy neural network. Fault detection is performed through the knowledge-based system, where fault detection heuristic rules have been generated from deep and shallow knowledge of the process. The fuzzy neural network performs the fault diagnosis task. The fault detection and diagnosis system has been implemented and successfully applied in studies of an industrial chemical process. Single and multiple simultaneous faults have been considered.

A Model-Based Diagnosis System for Identifying Faulty Components in Digital Circuits

We describe the ideas and implementation of a model-based diagnosis system for digital circuits. Our work is based on Reiter‘s theory of diagnosis from first principles [14], incorporated with Hou‘s theory of measurements [17], to derive possible diagnoses in a fault diagnosis task. To determine the best order in which measurements are to be taken, a measurement selection strategy using the genetic algorithm (MSSGA) is proposed. A circuit description language for describing circuits hierarchically is given. An efficient propositional logic prover used for consistency checking based on the trie structure is developed [22]. An example run is given to illustrate the working of the system. Finally, a comparison with other systems is discussed, and possible extensions to our system are described.

A graphic modeling and analysis tool for human fault diagnosis tasks

This study presents a graphic modeling and analysis tool for use in constructing an operator's mental model in fault diagnosis tasks. In most automatic and complicated process control systems, human fault diagnosis tasks have become increasingly complex and specialized. The system designer should consider the cognitive process of human operator to avert failure of implement action owing to a lack of compatibility between humans and aiding system interface. Here, an experiment is performed to investigate the nature of human fault diagnosis. A graphic modeling and analysis tool is then proposed to model the continuous process of human fault diagnosis. The approach proposed herein exploits both the line-chart and Petri nets to demonstrate the operator's thoughts and actions. Moreover, results in this study are integrated into an adaptive standard diagnosis model that can assess the operators' mental workload and accurately depict the interactions between human operator and aiding system. Relevance to industry Automatic intelligent diagnosis systems can not provide satisfactory operating performance. Human diagnosticians are more effective than computer ones. Results in this study offer further insight into an operator behavior in graphic form and also how to design a better aiding system.

An incremental retrieval mechanism for case-based electronic fault diagnosis

One problem with using CBR for diagnosis is that a full case description may not be available at the beginning of the diagnosis. The standard CBR methodology requires a detailed case description in order to perform case retrieval and this is often not practical in diagnosis. We describe two fault diagnosis tasks where many features may make up a case description but only a few features are required in an individual diagnosis. We evaluate an incremental CBR mechanism that can initiate case retrieval with a skeletal case description and will elicit extra discriminating information during the diagnostic process.

Information Aiding of Real-Time Systems for Human Fault Diagnosis Tasks

This research aims to develop and evaluate a real-time diagnosis assistant system that based on operator behavior recognition, hierarchical neural networks and expert explanation systems. The independent variables of were aiding levels and scenario types. The results showed that either partial or entire aiding approach could improve the performance significantly, and there was no significant difference between the partial aid and the entire aid. However, there was a significant effect of the level of aiding on the transitions of diagnostic strategy. These results revealed that with a direct and short information aiding, one could save diagnostic time and test steps.

A multi-model method to fault detection and diagnosis: Bayesian solution. An introductory treatise

In the paper, a method for solving fault detection and diagnosis problems in sampled-data stochastic systems is presented. As a main methodology tool the Bayesian view on uncertainty is exploited. The method can be classified as of a multi-model type. It requires to supply mathematical models of the system dynamics, each describing the situation when a particualr fault separately acts on the system or when the system behaves normally (i.e. as desired). At the stage of research, discrete-time stochastic causal input–output non-parametrized models are supported. As discrete-time courses, model of the actual system behaviour is recursively estimated and a decision on the actually acting fault is given. The presented method solves both the fault detection and diagnosis tasks simultaneously. Three illustrative examples show the method in action, possibly demonstrating a range of its applications. © 1998 John Wiley & Sons, Ltd.

Intelligent system for fault diagnosis on low voltage distribution networks

The task of fault diagnosis on the distribution networks operating at 11 kV and lower voltages is very different from the higher voltage levels because very little telemetered data is available and knowledge of a fault occurrence often depends on customers complaining of loss of supply. Unfortunately, the connectivity of customers to the network is not normally available on geographic information systems (GIS), so connectivity is indirectly established through post codes. A novel kind of object oriented blackboard architecture using deep and shallow knowledge has been developed. The notion of object-oriented active databases has also been introduced. The intelligent system developed can assist the control engineer in diagnosing the fault quickly to ameliorate the immediate effects of the fault. The special feature of this research is that the developed architecture is suitable for any fault diagnosis situation where information of the system is incomplete and is accrued over a period of time. Partial solutions are obtained as the first piece of information is received and these partial solutions are improved as further information is received.

Intelligent Diagnosis Systems

This paper examines and compares several different approaches to design of intelligent systems for diagnosis applications. These include expert systems (or knowledge based systems), truth (or reason) maintenance systems, case based reasoning systems, and inductive approaches like decision trees, neural networks (or connectionist systems), and statistical pattern classification systems. Each of these approaches is demonstrated through the design of a system for a simple automobile fault diagnosis task. The paper also discusses the domain characteristics that influence the choice of a specific technique (or combination of techniques) for a given application.

Real Time Fault Detection and Diagnosis of an Industrial Process

A real time fault detection and diagnosis system, which has been design to cope with transient behaviours of the process, as well as abrupt and incipient faults, is described in this paper. The computational system consists of a knowledge based approach coupled with a fuzzy neural network. Fault detection is performed through the knowledge based system. The fuzzy neural network performs the fault diagnosis task. The real time fault detection and diagnosis system has been implemented through a TURBO C++ program and successfully applied in simulation studies of an industrial process. Single and multiple simultaneous faults have been considered.

Depression and Problem-Solving Performance on a Fault-Diagnosis Task

Dysphoric and control undergraduates selected on the basis of Beck Depression Inventory scores were compared on a problem-solving task involving fault-diagnosis. The stimuli were wiring diagrams showing boxes or units connected to indicators, and the aim was to diagnose which one of the units was faulty. Dysphoric subjects did not differ from controls on problems requiring them to use the information available at the outset of the problem to identify all the potentially faulty units. They were impaired relative to controls in testing connections that they believed to be potentially faulty until the faulty unit was diagnosed; they made more repetition errors, carried out more redundant tests, and were slower than controls, although the groups did not differ in the number of correct solutions achieved. The findings are discussed in terms of theoretical models that posit impaired strategic processing in depression. Language: en

Depression and Problem-Solving Performance on a Fault-Diagnosis Task

Depression and Problem-Solving Performance on a Fault-Diagnosis Task

Individual differences and the effects of an information aid in performance of a fault diagnosis task

Two experiments are reported that investigated the performance of operators in a fault diagnosis task. Naval engineers working in the Ship's Control Centre (SCC) on board of frigates of the Royal Netherlands Navy were asked to solve a number of unfamiliar fault problems. In the first experiment, it was shown that there was a high correlation (0·85) between the number of problems correctly solved and the availability of system knowledge. In addition there was an effect of the level of formal education. In a second experiment it was shown that the provision of a simple help facility that compensates for a lack of system knowledge leads to a substantial increase in the number of problems that were correctly solved. The percentage of correct solutions increased from 62 to 89%. This increase was the same for operators with either an electrical engineering or a mechanical engineering background although the former group performed slightly better than the latter. These results show that the conclusion of Morris and Rouse (1985a) that instruction in the theoretical principles on which a system is based is less effective than the training of diagnostic procedures, underestimates the importance of system knowledge for the solution of fault diagnosis problems.

FAULT DETECTION AND DIAGNOSIS IN BATCH AND SEMI-BATCH PROCESSES USING ARTIFICIAL NEURAL NETWORKS

Abstract The objective of this work is to assess the feasibility of adopting artificial neural networks (ANNs) in fault detection and diagnosis for batch and semi-batch processes. Although there is a large volume of related publications available, most of them used steady-state data to train ANNs and, as such, the task of fault diagnosis can only be implemented in continuous operations. Based upon the concept of analytical redundancy, the framework of a two-stage fault monitoring system is proposed in this paper. In the first stage, a hybrid ANN is adopted to predict the long-term dynamic behaviors of the output variables under normal condition. The occurrence of fault(s) can be detected by inspecting the residuals, i.e. the differences between the measured and the predicted values of outputs. A second feedforward neural network is then used for the purpose of differentiating the residual patterns caused by various faults. In addition to the fact the results of pilot tests are quite satisfactory, it is al...

Automated extraction of attribute hierarchies for an improved decision-tree classifier

A method to automatically extract inherent hierarchies of attribute values from data is proposed to improve the performance of a decision-tree classifier. When attributes used for a decision tree take continuous values, simple decision rules such as “IF a certain attribute is less than a certain value” yield good results in many cases. The rationale of this type of rule is the natural tendency that two nearby attribute values have a “similar meaning” in the sense that they suggest the same class with high probability. When only discrete-type, unordered attributes are available for tree generation, however, such efficient decision rules are hard to obtain because “distance relationships” between the attribute values are seldom known beforehand. In order to solve this problem, the proposed method estimates from the training data set a conceptual distance between each pair of attribute values, and by iteratively grouping two attribute values with minimum distance, generates hierarchies of attribute values, which are then utilized for making the decision tree. This method is applied to the task of fault diagnosis of a certain printed circuit board, and it is indicated that the generated attribute hierarchies can reduce the size of a decision tree sufficiently, which results in a significant improvement of its classification accuracy.