Failure Diagnosis System Research Articles

Coordinated Decentralized Protocols for Failure Diagnosisof Discrete Event Systems

We address the problem of failure diagnosis in discrete event systems with decentralized information. We propose a coordinated decentralized architecture consisting of local sites communicating with a coordinator that is responsible for diagnosing the failures occurring in the system. We extend the notion of diagnosability, originally introduced in Sampath et al. (1995) for centralized systems, to the proposed coordinated decentralized architecture. We specify three protocols that realize the proposed architecture; each protocol is defined by the diagnostic information generated at the local sites, the communication rules used by the local sites, and the coordinator‘s decision rule. We analyze the diagnostic properties of each protocol. We also state and prove conditions for a language to be diagnosable under each protocol. These conditions are checkable off-line. The on-line diagnostic process is carried out using the diagnosers introduced in Sampath et al. (1995) or a slight variation of these diagnosers. The key features of the proposed protocols are: (i) they achieve, each under a set of assumptions, the same diagnostic performance as the centralized diagnoser; and (ii) they highlight the ’’performance vs. complexity‘‘ tradeoff that arises in coordinated decentralized architectures. The correctness of two of the protocols relies on some stringent global ordering assumptions on message reception at the coordinator‘s site, the relaxation of which is briefly discussed.

SAFETY ASSESSMENT MODEL OF SOFTWARE FOR FAILURE DIAGNOSIS SYSTEMS

The software for a failure diagnosis system can be represented in terms of production rules; the condition part represents a system failure condition and the conclusion part corresponds to a cause of the system failure or an appropriate protective action to be taken. This paper proposes a safety assessment model of the software to evaluate its contribution to the risk caused by the entire failure diagnosis system. The proposed risk criterion considers not only the reliability of hardware components of the failure diagnosis system, but also the reliability characteristics of the system to be monitored. Conventional verification and validation methods of rule-based systems assume that the software reliability can be achieved by maintaining the consistent relation between condition parts and conclusion parts. However, the risk criterion derived in this paper shows that the software for a failure diagnosis system cannot be optimized without considering these environmental factors.

A Failure Diagnosis System for the Burner of Absorption Chillers.

Sensors utilized for diagnosing the flame current of a flame-detector, for instance, are limited because the burner of the absorption chiller is small or medium in size. Subtle differences in the conditions can affect the readings of these sensors and make them unstable. The knowledge of burner daignosis based on such unstable sensors has several features. 1. The structure of the knowledge is relatively simple. 2. The knowledge is probabilistic because of the unstable readings and the non-deterministic nature of the knowledge possessed by engineers in the domain 3. The knowledge has a low quantitative accuracy. In this paper, we propose a system that employs simple binary logic to determine failure modes, and probalistic multi-valued logic to evaluate the probabilities of failure causes. This system achieves suitable expressions for knowledge of domain-engineers and features a compact inference engine that can be implemented into small-size computers such as a notebook-PCs.

Axiomatic Design Theory for Systems

A general theory for system design is presented based on axiomatic design. The theory is applicable to many different kinds of systems, including machines, large systems, software systems, organizations, and systems consisting of a combination of hardware and software. Systems are represented by means of a system architecture, which takes the form of the {FR}/{DP}/{PV}hierarchies, a ‘junction-module’ diagram, and the ‘flow diagram’. The ‘flow diagram’ for system architecture concisely represents the system design, the relationship among modules, and the control sequence in operating systems. The flow diagram of the system architecture can be used for many different tasks: design, construction, operation, modification, and maintenance of the system. It should also be useful for distributed design and operation of systems, diagnosis of system failures, and for archival documentation.

Active diagnosis of discrete-event systems

The need for accurate and timely diagnosis of system failures and the advantages of automated diagnostic systems are well appreciated. However, diagnosability considerations are often not explicitly taken into account in the system design. In particular, design of the controller and that of the diagnostic subsystem are decoupled, and this may significantly affect the diagnosability properties of a system. The authors present an integrated approach to control and diagnosis. More specifically, they present an approach for the design of diagnosable systems by design of the system controller. This problem, which they refer to as the active diagnosis problem, is studied in the framework of discrete-event systems (DESs); it is based on prior and new results on the theory of diagnosis for DESs and on existing results in supervisory control under partial observations. They formulate the active diagnosis problem as a supervisory control problem where the legal language is an appropriate regular sublanguage of the regular language generated by the system. They present an iterative procedure for determining the supremal controllable, observable, and diagnosable sublanguage of the legal language and for obtaining the supervisor that synthesizes this language. This procedure provides both a controller that ensures diagnosability of the closed-loop system and a diagnoser for online failure diagnosis. The procedure can be implemented using finite-state machines and is guaranteed to converge in a finite number of iterations. The authors illustrate their approach using a simple pump-valve system.

Dynamic Compensation in a Failure Diagnosis System for Absorption Chillers/Heaters.

Recently, environmentally-friendly absorption chillers/heaters have become popular as air conditioning/heating sources. This system is free from CFC use. In this way next improvements such as a system for maintenance, including failure diagnosis, are needed to be considered. The authors earlier proposed a failure diagnosis system for absorption chillers/heaters that is compact and sufficiently accurate. The system uses linear or curvilinear quantitative relationships derived from experiments grounded on qualitative knowledge. These quantitative relationships, however, are only effective in steady-state. Therefore, under the condition of large state-transitions, the possibility of erroneous daignosis exists. In this paper, we propose a compensation method that allows the failure diagnosis system to be effective even under large state-transitions. The method eliminates possibil- ities of fault judgements.

Predictors of fault-finding skill

In modern-day, complex industrial systems, the role of human operators has become ever more vital in the safe running and diagnosis of system failures. Previous research has, however, failed to identify consistent predictors of novel fault-finding skill. It is argued that this failure has arisen because researchers have adopted an atheoretical approach to identifying predictors of diagnostic skill. In this paper, a study is reported which examines the development of diagnostic skill from the perspective of Ackerman's (1988) theory of the ability determinants of skill acquisition (TADSA). Forty-two subjects completed 600 fault-diagnosis problems, and their performance was predicted by measures of ability that tapped aspects of cognitive performance which included cognitive capacity, perceptual speed, and psychomotor speed. As predicted by TADSA, fault-diagnosis performance is predictable by different measures cognitive performance at different stages of practice.

A Petri‐net based machine tool maintenance management system

Considers an integrated approach to maintenance management, based on the development of an advanced machine tool failure diagnosis and recovery system. Outlines how such systems can be utilized to effectively monitor the ever more complicated machining facilities used in modern manufacturing. Illustrates how a Petri‐net based system has been developed to meet the needs of a modern computer‐based maintenance facility and the potential that such systems provide in a move towards total productive maintenance. Deals with the steps taken to ensure that such a system can be successfully deployed, and considers how the resulting integrated maintenance management system can benefit existing users of advanced manufacturing technology.

Least-cost failure diagnosis in uncertain reliability systems

In many textbook solutions, for systems failure diagnosis problems studied using reliability theory and artificial intelligence, the prior probabilities of different failure states can be estimated and used to guide the sequential search for failed components after the whole system fails. In practice, however, both the component failure probabilities and the structure function of the system being examined—i.e., the mapping between the states of its components and the state of the system—may not be known with certainty. At best:, the probabilities of different hypothesized system descriptions, each specifying the component failure probabilities and the system's structure function, may be known to a useful approximation, perhaps based on sample data and previous experience. Cost-effective diagnosis of the system's failure state is then a challenging problem. Although the probabilities of component failures are aleatory, uncertainties about these probabilities and about the system structure function are epistemic. This paper examines how to make best use of both epistemic prior probabilities for system descriptions and the information gleaned from costly inspections of component states after the system fails, to minimize the average cost of identifying the failure state. Two approaches are introduced for systems dominated by aleatory uncertainties, one motivated by information theory and the other based on the idea of trying to prove a hypothesis about the identity of the failure state as efficiently as possible. While the general problem of cost-effective failure diagnosis is computationally intractable (NP-hard), both heuristics provide useful approximations on small to moderate sized problems and optimal results for certain common types of reliability systems, including series, parallel, parallel-series, and k-out-of- n systems. A hybrid heuristic that adaptively chooses which heuristic to apply next after any sequence of observations (component test results) appears to give excellent results. Several computational experiments are summarized in support of these conclusions, and extensions to reliability systems with repair are briefly considered. Next, it is shown that diagnosis can proceed when aleatory and epistemic uncertainties are both present using the same techniques developed for aleatory probabilities alone. If only the epistemic probability distribution of system descriptions is known, then the same heuristics that are used to diagnose a system's failure state for systems with known descriptions can also be used to identify the system and diagnose its failure state when there is epistemic uncertainty about the identity of the system. This result suggests a unified approach to least-cost failure diagnosis in reliability systems with both aleatory probabilities of component failures and epistemic probabilities for system descriptions.

Factors Associated With Change in Resuscitation Preference of Seriously Ill Patients

Background: During serious illness, patient preferences regarding life-sustaining treatments play an important role in medical decisions. However, little is known about life-sustaining preference stability in this population or about factors associated with preference change. Methods: We evaluated 2-month cardiopulmonary resuscitation (CPR) preference stability in a cohort of 1590 seriously ill hospitalized patients at 5 acute care teaching hospitals. Using multiple logistic regression, we measured the association of patient demographic and health-related factors (quality of life, function, depression, prognosis, and diagnostic group) with change in CPR preference between interviews. Results: Of 1590 patients analyzed, 73% of patients preferred CPR at baseline interview and 70% chose CPR at follow-up. Preference stability was 80% overall—85% in patients initially preferring CPR and 69% in those initially choosing do not resuscitate (DNR). For patients initially preferring CPR, older age, non—African American race, and greater depression at baseline were independently associated with a change to preferring DNR at follow-up. For patients initially preferring DNR, younger age, male gender, less depression at baseline, improvement in depression between interviews, and an initial admission diagnosis of acute respiratory failure or multiorgan system failure were associated with a change to preferring CPR at follow-up. For patients initially preferring DNR, patients with substantial improvements in depression score between interviews were more than 5 times as likely to change preference to CPR than were patients with substantial worsening in depression score. Conclusions: More than two thirds of seriously ill patients prefer CPR for cardiac arrest and 80% had stable preferences over 2 months. Factors associated with preference change suggest that depression may lead patients to refuse life-sustaining care. Providers should evaluate mood state when eliciting patients' preferences for life-sustaining treatments. Arch Intern Med. 1996;156:1558-1564

Factors associated with change in resuscitation preference of seriously ill patients. The SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments

During serious illness, patient preferences regarding life-sustaining treatments play an important role in medical decisions. However, little is known about life-sustaining preference stability in this population or about factors associated with preference change. We evaluated 2-month cardiopulmonary resuscitation (CPR) preference stability in a cohort of 1590 seriously ill hospitalized patients at 5 acute care teaching hospitals. Using multiple logistic regression, we measured the association of patient demographic and health-related factors (quality of life, function, depression, prognosis, and diagnostic group) with change in CPR preference between interviews. Of 1590 patients analyzed, 73% of patients preferred CPR at baseline interview and 70% chose CPR at follow-up. Preference stability was 80% overall-85% in patients initially preferring CPR and 69% in those initially choosing do not resuscitate (DNR). For patients initially preferring CPR, older age, non-African American race, and greater depression at baseline were independently associated with a change to preferring DNR at follow-up. For patients initially preferring DNR, younger age, male gender, less depression at baseline, improvement in depression between interviews, and an initial admission diagnosis of acute respiratory failure or multiorgan system failure were associated with a change to preferring CPR at follow-up. For patients initially preferring DNR, patients with substantial improvements in depression score between interviews were more than 5 times as likely to change preference to CPR than were patients with substantial worsening in depression score. More than two thirds of seriously ill patients prefer CPR for cardiac arrest and 80% had stable preferences over 2 months. Factors associated with preference change suggest that depression may lead patients to refuse life-sustaining care. Providers should evaluate mood state when eliciting patients' preferences for life-sustaining treatments.

Failure diagnosis using discrete-event models

Detection and isolation of failures in large, complex systems is a crucial and challenging task. The increasingly stringent requirements on performance and reliability of complex technological systems have necessitated the development of sophisticated and systematic methods for the timely and accurate diagnosis of system failures. We propose a discrete-event systems (DES) approach to the failure diagnosis problem. This approach is applicable to systems that fall naturally in the class of DES; moreover, for the purpose of diagnosis, continuous-variable dynamic systems can often be viewed as DES at a higher level of abstraction. We present a methodology for modeling physical systems in a DES framework and illustrate this method with examples. We discuss the notion of diagnosability, the construction procedure of the diagnoser, and necessary and sufficient conditions for diagnosability. Finally, we illustrate our approach using realistic models of two different heating, ventilation, and air conditioning (HVAC) systems, one diagnosable and the other not diagnosable. While the modeling methodology presented here has been developed for the purpose of failure diagnosis, its scope is not restricted to this problem; it can also be used to develop DES models for other purposes such as control.

Optimization of failure diagnosis procedure based on diagnosis error loss

AbstractBased on the results of its inspection items, the failure diagnosis system identifies the cause of system failure. The role of failure diagnosis is to identify the system state correctly. However, diagnosis errors may be caused by sensor failure or operator errors, which lead to serious damage or unnecessary protective actions. The effect of an error depends on the system condition. In the design of a failure diagnosis procedure, not only diagnosis error probabilities, but also the effect of diagnosis error on the system must be considered.This paper proposes the optimization of failure diagnosis procedures based on diagnosis error loss. The optimal procedure that minimizes the expected loss caused by inspection errors is developed for two cases: one where all results of inspection items are given simultaneously, and one where an inspection item is examined sequentially. In the former case the optimal failure diagnosis logic is obtained, while in the latter case the optimal test sequence is developed. Illustrative examples show the merits and details of the proposed method.

Strategies of failure diagnosis in computer-controlled manufacturing systems: empirical analysis and implications for the design of adaptive decision support systems

This study investigates strategies in failure diagnosis at cutting-machine-tools with a verbal knowledge acquisition technique. Sixty-nine semi-structured interviews were performed with mechanical and electrical maintenance technicians, and a protocol analysis was conducted. Strategies were analysed in dependence of the technician's job experience, his familiarity with the problem and problem complexity. The technicians were categorized into three groups, novices, advanced, and experts, based upon level of experience. Results show that typical strategies of failure diagnosis are "Historical information", "Least effort", "Reconstruction", and "Sensory check". Strategies that lead to a binary reduction of the problem space, such as "Information uncertainty" and "Split half", play only a minor role in real-life failure diagnosis. Job experience and the familiarity with the problem significantly influenced the occurrence of strategies. In addition to "Symptomatic search" and "Topographic search", results show frequent use of case-based strategies, particularly in routine failures. In novel situations, technicians usually used "Topographic search". A software design method, the strategy-based software design (SSD) is proposed, that uses strategies to derive decision support systems, that are adaptive to the different working styles and the changing levels of experience in user groups. The methodology is briefly described and illustrated by the development of an information support system for maintenance and repair.

A Petri‐Net‐based machine tool failure diagnosis system

Describes the ongoing development of a Petri‐Net‐based machine tool failure diagnosis system. The described system incorporates condition monitoring systems and expert system‐based diagnostic tools. Brings these together with the Petri‐Net monitoring system to form an integrated fault location and rectification package and an associated maintenance management system.

Failure diagnosis system using ARTMAP neural networks

A novel scheme for diagnosing failures in dynamic systems is presented. After a failure is detected by the state chi-square test, two ARTMAP neural networks are applied to recognize patterns formed by data of the state chisquare test in order to determine when the failure happened, how serious the failure is, and where the failure is located. The efficacy of the proposed scheme is demonstrated by simulation examples where soft failures in an integrated global positioning system/inertial navigation system are successfully diagnosed. An effective approach is also developed to overcome the difficulty of distinguishing gyro failures from accelerometer failures.

Automated fault detection and accommodation: a learning systems approach

The detection, diagnosis, and accommodation of system failures or degradations are becoming increasingly more important in modern engineering problems. A system failure often causes changes in critical system parameters, or even, changes in the nonlinear dynamics of the system. This paper presents a general framework for constructing automated fault diagnosis and accommodation architectures using on-line approximators and adaptation/learning schemes. In this framework, neural network models constitute an important class of on-line approximators. Changes in the system dynamics are monitored by an on-line approximation model, which is used not only for detecting but also for accommodating failures. A systematic procedure for constructing nonlinear estimation algorithms is developed, and a stable learning scheme is derived using Lyapunov theory. Simulation studies are used to illustrate the results and to gain intuition into the selection of design parameters. >

Failure Diagnosis System for Absorption Chiller/Heater.

Recently, the absorption chiller/heater is attracting interest because it delivers airconditioning without using fluorocarbons. Practically, it requires periodical or non-periodical maintenance. It is difficult to detect failures except in summertime with large refrigerant load and, moreover, human expertise is needed for discrimination of their origin. This paper reports on a system that automatically predicts the possibility of failures and discriminates their origins, regardless of the driving situation.

Failure Diagnostic System for Air-Operated Control Valves Using Neural Network

A prototype failure diagnosis system has been developed using neural network technology for the actuators of air-operated valves. Because actual failure data were not easily available, the data of 30 failure patterns were experimentally obtained using more than 10 sensors. The time series data of sensor signals are Fourier transformed. The data of magnitude spectrum, phase difference and other quantities are used as the characteristic parameters in our failure diagnosis. From the data, appropriate information for use in failure diagnosis was extracted. Furthermore, similarities among failure characteristics were found by fuzzy clustering and statistical analysis. The new system that we developed consists of many sub-networks and one main network. Each sub-network is related to one specific sensor signal, and deals with the magnitude spectra from the sensor signal. The main network makes the final decision according to the outputs from the sub-networks and other data. In our system, the number of network connections can be reduced by approximately 40% without degradation of the recognition capability in comparison with the conventional system that uses only one neural network.

Optimization of Failure Diagnosis Procedure Based on Diagnosis Error Loss.

Based on the results of its inspection items, the failure diagnosis system identifies the cause of system failure. The role of failure diagnosis is to identify the system state currectly. However, diagnosis errors may be caused by sensor failure or operator errors, which lead to serious damage or unnecessary protective actions. The effect of an error depends on the system condition. In the design of failure diagnosis procedure, not only diagnosis error probability, but also its effect on the system must be considered.This paper proposes the optimization of failure diagnosis procedures based on diagnosis error loss. The optimal procedure that minimizes the expected loss caused by inspection errors is developed for two cases: one where all results of inspection items are given simultaneously and one where an inspection item is examinied sequentially. In the former case the optimal failure diagnosis logic is obtained, while in the latter case the optimal test sequence is obtained. Illustrative examples show the merits and details of the proposed method.