- Email: [email protected]
Fault diagnosis and prevention by fuzzy sets
1190
World Abstracts on Microelectronics and Reliability
taken from 0.3-14 percent strain range. A conventional crack criterion was used as the fatigue metric in the thermal cycle experiments. A simple finite element model of a 68 post molded leaded chip carrier (PLCC) was analyzed to examine expected stress loading of the solder joints under typical
operating conditions. Transient thermal stress loading and low stress creep of the solder joint are discussed in light of this model and known creep properties of solder, it is shown that during typical operating cycles, the total creep strain can account for the majority of plastic or nonrecoverable strain
3. CIRCUIT AND SYSTEMS RELIABILITY, MAINTENANCE AND REDUNDANCY Fault diagnosis and prevention by fuzzy sets. IGOR GAZDIK. IEEE Trans. Reliab. R-34 (4), 382 (1985). Fault diagnosis and prevention in engineering systems are associated with imprecision which can often be overcome by fuzzy-set theoretic techniques. One such technique is based on the concept of conditioned fuzzy sets: a set of fuzzified influences acting on a system, conditions a fuzzy relation existing between those influences and the symptoms of the state of the system. The conditioning materializes as a fuzzy intersection (here extended to cover real phenomena), and a partial ordering of the influence parameters by their importance for each symptom of the state of the system. This paper explains the application of the conditioning fuzzy technique to a real engineering device. In particular, the paper focuses on how to use pertinent fuzzy-set theoretic concepts for creating a fault diagnosis and prevention model of an engineering system, and how to develop the application induced fuzzy intersection formula. A step-by-step summary of the computational procedure, including a sample calculation, concludes the paper. The fuzzy-set approach to fault diagnosis and prevention has several advantages: The use of linguistic hedges makes it possible to develop the model of a system based on verbal formulation, and subsequent quantification, of the state of the system under study. That formulation follows from observations of the system either on the test bench, or in actual operation. The knowledge base thus generated can be refined, as more experience of the behavior of the system is gained, to yield more dependable results. Simulations of various aspects of the fault prevention process are possible. The fuzzy-set theoretic operations are simple and easy to program for the computer. Many problems in fault diagnosis and prevention in engineering systems can, indeed, be solved by fuzzy sets only.
Fault tree analysis, methods, and applications--a review. W. S. LEE, D. L. GROSH, F. A. TILLMAN and C. H. LIE. IEEE Trans. Reliab. R-M(3), 194 (1985). This paper reviews and classifies fault-tree analysis methods developed since 1960 for system safety and reliability. Fault-tree analysis is a useful analytic tool for the reliability and safety of complex systems. The literature on fault-tree analysis is, for the most part, scattered through conference proceedings and company reports. We have classified the literature according to system definition, fault-tree construction, qualitative evaluation, quantitative evaluation, and available computer codes for fault-tree analysis.
A simplified design procedure for life tests bused on KnlibuckLeibler information. Hmosm OWrA and IKUO ARIZONO. IEEE Trans. Reliab. R-34(4), 363 (1985). This paper describes a new simple, practical design procedure for life tests based on the Kullback-Leibler information. We assume that the length of life has an exponential distribution. In the reliability literature, the exponential distribution is the underlying distribution of the length of life.
sensors are reporting symptoms and the expert system uses these in a Bayes or modified Bayes mode as evidence to help compute the posterior estimate of the source and/or nature of the fault. One of the complaints of the expert-system developer community is that Bayes formula can rarely be applied in pure form due to lack of data from which to compute the priors and likelihood ratio elements; less defensible evidential reasoning models are becoming prevalent. For some applications, sufficiently large pools of such data do exist; however, their validity is suspect due to the lack of built-in test or built-in sensor reporting. That is, these failure data-bases depend on human operator reporting of failure events and causes. This article explores humanoperator-introdcued validity problems as part of an attempt to develop a strategy for compensating for failure data invalidities to the point where a Bayes approach can be possible. After elaborating on the Bayes formulation, the design of the experiments are reviewed. Results are then presented and discussed along with suggestions for further research.
Bayes estimation of the reliability functiou of normal distribution. S. K. SINHA. IEEE Trans. Reliab. R-34(4), 360 (1985). Bayes estimates of the survivor function of the snormal distribution are obtained using the predictive distribution and some recent results on Bayes approximations due to Lindley. Monte Carlo simulation was used. These estimates have smaller rms error than that of their maximum likelihood counterparts. Except for large samples, the exact estimators based on the predictive distribution perform better than the Lindley approximate estimators.
A heuristic method to upgrade system availability for hot or cold standby and voting systems. KAREN M. YASUTAKEBAKERand ERNESTJ. HENLEY.IEEE Trans. Reliab. R-34(4), 389 (1985). This paper details a new heuristic-based method for optimizing system availability by redundant allocation. Hot or cold standby, majority voting, repair crew, or other type of logic systems are easily handled. A computer program, Upgrading of System Availability Package (USAP), has been developed and is available from the authors. This program assesses and upgrades system availability when the component availability and system configuration are known. If the system availability needs to be improved and funds are available for such an improvement, then USAP suggests a system configuration that has the desired system availability within a given cost constraint. The algorithm is based on cut sets, so both fault trees and block diagrams can be handled. No special subroutines which specify the functional form of the system availability are required. This makes the program easy to implement on a personal computer, since no linking and compiling of special routines are needed. The upgrading program in USAP provides good results as compared to other upgrading methods found in the literature. This simple and intuitive analysis gives practical solutions for very large, complicated systems.
Chi-square probabilities are Poisson probabilities in di~aise. Expert fault-diagnosis under humn-repurting bias. BARRYG. SILVERMAN and ALEXANDER G. TSOLAKIS. IEEE Trans. Reliab. R-34 (4), 366 (I 985). An important class of problems is the application of expert systems to fault diagnosis where
ANDREW GORSKI. IEEE Trans. Reliab. R-34(3), 209 (1985). Since chi-square probabilities and Poisson probabilities are different forms of the same mathematical function, it is reasonable to use the chi-square table for obtaining Poisson
World Abstracts on Microelectronics and Reliability
taken from 0.3-14 percent strain range. A conventional crack criterion was used as the fatigue metric in the thermal cycle experiments. A simple finite element model of a 68 post molded leaded chip carrier (PLCC) was analyzed to examine expected stress loading of the solder joints under typical
operating conditions. Transient thermal stress loading and low stress creep of the solder joint are discussed in light of this model and known creep properties of solder, it is shown that during typical operating cycles, the total creep strain can account for the majority of plastic or nonrecoverable strain
3. CIRCUIT AND SYSTEMS RELIABILITY, MAINTENANCE AND REDUNDANCY Fault diagnosis and prevention by fuzzy sets. IGOR GAZDIK. IEEE Trans. Reliab. R-34 (4), 382 (1985). Fault diagnosis and prevention in engineering systems are associated with imprecision which can often be overcome by fuzzy-set theoretic techniques. One such technique is based on the concept of conditioned fuzzy sets: a set of fuzzified influences acting on a system, conditions a fuzzy relation existing between those influences and the symptoms of the state of the system. The conditioning materializes as a fuzzy intersection (here extended to cover real phenomena), and a partial ordering of the influence parameters by their importance for each symptom of the state of the system. This paper explains the application of the conditioning fuzzy technique to a real engineering device. In particular, the paper focuses on how to use pertinent fuzzy-set theoretic concepts for creating a fault diagnosis and prevention model of an engineering system, and how to develop the application induced fuzzy intersection formula. A step-by-step summary of the computational procedure, including a sample calculation, concludes the paper. The fuzzy-set approach to fault diagnosis and prevention has several advantages: The use of linguistic hedges makes it possible to develop the model of a system based on verbal formulation, and subsequent quantification, of the state of the system under study. That formulation follows from observations of the system either on the test bench, or in actual operation. The knowledge base thus generated can be refined, as more experience of the behavior of the system is gained, to yield more dependable results. Simulations of various aspects of the fault prevention process are possible. The fuzzy-set theoretic operations are simple and easy to program for the computer. Many problems in fault diagnosis and prevention in engineering systems can, indeed, be solved by fuzzy sets only.
Fault tree analysis, methods, and applications--a review. W. S. LEE, D. L. GROSH, F. A. TILLMAN and C. H. LIE. IEEE Trans. Reliab. R-M(3), 194 (1985). This paper reviews and classifies fault-tree analysis methods developed since 1960 for system safety and reliability. Fault-tree analysis is a useful analytic tool for the reliability and safety of complex systems. The literature on fault-tree analysis is, for the most part, scattered through conference proceedings and company reports. We have classified the literature according to system definition, fault-tree construction, qualitative evaluation, quantitative evaluation, and available computer codes for fault-tree analysis.
A simplified design procedure for life tests bused on KnlibuckLeibler information. Hmosm OWrA and IKUO ARIZONO. IEEE Trans. Reliab. R-34(4), 363 (1985). This paper describes a new simple, practical design procedure for life tests based on the Kullback-Leibler information. We assume that the length of life has an exponential distribution. In the reliability literature, the exponential distribution is the underlying distribution of the length of life.
sensors are reporting symptoms and the expert system uses these in a Bayes or modified Bayes mode as evidence to help compute the posterior estimate of the source and/or nature of the fault. One of the complaints of the expert-system developer community is that Bayes formula can rarely be applied in pure form due to lack of data from which to compute the priors and likelihood ratio elements; less defensible evidential reasoning models are becoming prevalent. For some applications, sufficiently large pools of such data do exist; however, their validity is suspect due to the lack of built-in test or built-in sensor reporting. That is, these failure data-bases depend on human operator reporting of failure events and causes. This article explores humanoperator-introdcued validity problems as part of an attempt to develop a strategy for compensating for failure data invalidities to the point where a Bayes approach can be possible. After elaborating on the Bayes formulation, the design of the experiments are reviewed. Results are then presented and discussed along with suggestions for further research.
Bayes estimation of the reliability functiou of normal distribution. S. K. SINHA. IEEE Trans. Reliab. R-34(4), 360 (1985). Bayes estimates of the survivor function of the snormal distribution are obtained using the predictive distribution and some recent results on Bayes approximations due to Lindley. Monte Carlo simulation was used. These estimates have smaller rms error than that of their maximum likelihood counterparts. Except for large samples, the exact estimators based on the predictive distribution perform better than the Lindley approximate estimators.
A heuristic method to upgrade system availability for hot or cold standby and voting systems. KAREN M. YASUTAKEBAKERand ERNESTJ. HENLEY.IEEE Trans. Reliab. R-34(4), 389 (1985). This paper details a new heuristic-based method for optimizing system availability by redundant allocation. Hot or cold standby, majority voting, repair crew, or other type of logic systems are easily handled. A computer program, Upgrading of System Availability Package (USAP), has been developed and is available from the authors. This program assesses and upgrades system availability when the component availability and system configuration are known. If the system availability needs to be improved and funds are available for such an improvement, then USAP suggests a system configuration that has the desired system availability within a given cost constraint. The algorithm is based on cut sets, so both fault trees and block diagrams can be handled. No special subroutines which specify the functional form of the system availability are required. This makes the program easy to implement on a personal computer, since no linking and compiling of special routines are needed. The upgrading program in USAP provides good results as compared to other upgrading methods found in the literature. This simple and intuitive analysis gives practical solutions for very large, complicated systems.
Chi-square probabilities are Poisson probabilities in di~aise. Expert fault-diagnosis under humn-repurting bias. BARRYG. SILVERMAN and ALEXANDER G. TSOLAKIS. IEEE Trans. Reliab. R-34 (4), 366 (I 985). An important class of problems is the application of expert systems to fault diagnosis where
ANDREW GORSKI. IEEE Trans. Reliab. R-34(3), 209 (1985). Since chi-square probabilities and Poisson probabilities are different forms of the same mathematical function, it is reasonable to use the chi-square table for obtaining Poisson