Wes Hines - Academia.edu (original) (raw)

Papers by Wes Hines

Applied Computational Intelligence - Proceedings of the 6th International FLINS Conference, 2004

The investigation and application of on-line monitoring programs has been ongoing for over two de... more The investigation and application of on-line monitoring programs has been ongoing for over two decades by the U.S. nuclear industry and researchers. To this date, only limited pilot installations have been demonstrated and the original objectives have changed significantly. Much of the early work centered on safety critical sensor calibration monitoring and reduction. The current focus is on both sensor and equipment monitoring. This paper presents the major lessons learned that contributed to the lengthy development process including model development and implementation issues, and the results of a recently completed cost benefit analysis.

Power Systems, 2002

Most data-based predictive modeling techniques have an inherent weakness in that they may give un... more Most data-based predictive modeling techniques have an inherent weakness in that they may give unstable or inconsistent results when the predictor data is highly correlated. Predictive modeling problems of this design are usually under constrained and are termed ill-posed. This paper presents several examples of ill-posed diagnostic problems and regularization methods necessary for getting accurate and consistent prediction results. The examples include plant-wide sensor calibration monitoring and the inferential sensing of nuclear power plant feedwater flow using neural networks, and non-linear partial least squares techniques, and linear regularization techniques implementing ridge regression and informational complexity measures.

International Journal of Intelligent Systems, 2002

Most data-based predictive modeling techniques have an inherent weakness in that they may give un... more Most data-based predictive modeling techniques have an inherent weakness in that they may give unstable or inconsistent results when the predictor data is highly correlated. Predictive modeling problems of this design are usually under constrained and are termed ill-posed. This paper presents several examples of ill-posed surveillance and diagnostic problems and regularization methods necessary for getting accurate and consistent prediction

AAAI Fall Symposium on Artificial Intelligence for …, 2007

Most modern prognostic algorithms are founded on a simple abstraction of device degradation, for ... more Most modern prognostic algorithms are founded on a simple abstraction of device degradation, for an individual device there exists a degradation signal that progresses along a unique path until it crosses a critical failure threshold. While this abstraction has been shown to be valid for well understood failure modes under controlled stress conditions, its viability in "real world" devices being exposed to "real world" stresses is questionable. Because the complexity of degradation should scale in a similar fashion as the complexity of the device, the applicability of a simple abstraction of degradation is increasingly arguable for modern devices. This paper will propose an alternative to the current abstraction of degradation, which is founded on the premise that degradation data should be allowed to speak for itself. In this way, many different forms of information can be incorporated into a prognoser's estimate of a device's remaining useful life (RUL). More specifically, this paper will outline a methodology for implementing a dynamic prognoser that can be incrementally trained to learn general (physical model output, expert opinion, etc.) and specific ("real world" data) degradation trends. This work will demonstrate the viability of the proposed method by applying a particular embodiment, namely the path classification and estimation (PACE) model, to data collected from a deep-well oil exploration drill. To begin, expert opinion will be used to develop a PACE prognoser. Next, data collected from individual drills will be used to incrementally train the prognoser to learn specific degradation trends.

by Chris Black, J. Wesley Hines, and …, 1997

An autoassociative artificial neural network (AANN) instrument channel monitoring technique has b... more An autoassociative artificial neural network (AANN) instrument channel monitoring technique has been developed for sensor and associated instrument channel online calibration verification. Several AANN models, each modeling a group of interrelated signals, are ...

Ideally, health monitoring of new, complex engineering systems should occur from initial operatio... more Ideally, health monitoring of new, complex engineering systems should occur from initial operation to decommissioning. Health monitoring typically involves a suite of modules, including system monitoring, fault detection, fault diagnostics, and system prognostics. However, for systems which have not yet operated, this is challenging. Most available health monitoring modules are empirically based, meaning they are derived from available historic data. For new system designs, such data simply does not exist. This research proposes an adaptive modeling system which initially builds empirical models

2008 Annual Reliability and Maintainability Symposium, 2008

ABSTRACT Equipment, process, and system prognostic techniques can be classified as belonging to o... more ABSTRACT Equipment, process, and system prognostic techniques can be classified as belonging to one of three major classes of methods: 1) conventional reliability-based using failure times (Weibull), 2) population based with environmental considerations (e.g. proportional hazards modeling), and 3) individual based (e.g. general path model). A new individual-based prognostic algorithm, termed the path classification and estimation (PACE) model, has been developed and is based entirely on failure data. This model recasts the general path model (GPM), which is the foundation of the majority of the modern individual based prognosis algorithms, as a classification problem, where a current device's degradation path is classified according to a series of exemplar paths and the results of the classification are used to estimate the remaining useful life (RUL) of the device. The requirement of the existence of a failure threshold is removed, thereby enabling the PACE to be applied to ldquoreal worldrdquo systems, where a single failure threshold is not likely to occur. If the failure threshold is known, simple formatting may be applied to the degradation paths such that they can be easily used with the PACE. The newly proposed method was applied to data collected from the hydraulic steering system of a drill used for deep oil exploration with the objective of detecting, diagnosing, and prognosing faults. The PACE was used to predict the RUL for several failure modes using actual data. For this work, a three tiered architecture was implemented, where conventional reliability methods were used to estimate the population-based RUL, PACE population-based prognosers were trained to map the cause of a failure mode to the RUL, and PACE individual prognosers were trained to map the effects of a failure mode to the RUL. It was found that the population based prognoser produced RUL estimates with large errors (75 hours) and uncertainties (261 hours). The individual prognosers were found to si- gnificantly outperform the population based prognoser, with errors ranging from 1.2 to 11.4 hours with 95% confidence intervals ranging from 0.67 to 32.02 hours.

Energy system on-line-monitoring is becoming a crucial component of improving safety, reliability... more Energy system on-line-monitoring is becoming a crucial component of improving safety, reliability, and profitability. The Holy Grail is the development prognostic methodologies to accurately predict the Remaining Useful Life (RUL) of a system or component for predictive maintenance and effective risk mitigation. Calculating precise RUL estimates requires both system specific maintenance information and performance data to develop representative lifecycle models. Current conventional prognostic methods focus on process data and do not utilize maintenance data to directly influence the modeling and data analysis. However, equipment maintenance impacts future system degradation and is dependent on the maintenance actions taken. Differences in the amount of degradation removed from a system are common for repaired equipment compared to replacements. This talk discusses methods of incorporating maintenance information into Lifecycle Prognostics and the effect it has on prediction error a...

The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life ... more The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life of individual systems or components based on their use and performance. This class of prognostic algorithms is termed effects-based, or Type III, prognostics. A unit-specific prognostic model, called the General Path Model, involve identifying an appropriate degradation measure to characterize the system's progression to failure. A functional fit of this parameter is then extrapolated to a pre-defined failure threshold to estimate the remaining useful life of the system or component. This paper proposes a specific formulation of the General Path Model with dynamic Bayesian updating as one effects-based prognostic algorithm. The method is illustrated with an application to the prognostics challenge problem posed at PHM '08.

Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 2014

As nuclear power plants seek to extend their licenses and maintain a high level of performance an... more As nuclear power plants seek to extend their licenses and maintain a high level of performance and safety, online monitoring and assessment of system degradation are becoming a crucial consideration. A goal of the DOE Light Water Reactor Sustainability program is the accurate estimation of the remaining useful life of nuclear power plant systems, structures, and components. Effective prognostic systems should seamlessly predict the remaining useful life from beginning of component life to end of component life, so-called Lifecycle Prognostics. When a component is first put into operation, the only information available may be past failure times of similar components or the expected distribution of failure times derived from reliability analyses of these data (Type I Prognostics). These data provide an estimated life for the average component operating under average usage conditions. As the component operates, it begins to consume its available life at a rate largely influenced by th...

Applied Artificial Intelligence, 2006

2008 International Conference on Prognostics and Health Management, 2008

Prognostic algorithms can be divided into three major categories. The most basic methods model th... more Prognostic algorithms can be divided into three major categories. The most basic methods model the component or system reliability using failure time data and conventional models such as the Weibull. When information pertaining to the operating condition and environmental stressors are available, stress-based techniques can be used. The third type of prognostics is termed effects-based. It is truly an individual

Proceedings of International Conference on Neural Networks (ICNN'96)

Multilayer feedforward neural networks with sigmoidal activation functions have been termed “univ... more Multilayer feedforward neural networks with sigmoidal activation functions have been termed “universal function approximators”. Although these types of networks can approximate any continuous function to a desired degree of accuracy, this approximation may require an inordinate number of hidden nodes and is only accurate over a finite interval. These short comings are due to the standard multilayer perceptron's (MLP) architecture

2005 Annual Conference Proceedings

This paper presents the methods and results of delivering a basic Statics course to Pellissippi S... more This paper presents the methods and results of delivering a basic Statics course to Pellissippi State Technical Community College (PSTCC) students located in Knoxville, Tennessee over the Internet. All aspects of the course, including textbook, lectures, class meetings, student discussions, homework and tests were conducted through the Internet. The online course material included animations, simulations, narrations, and graphics. Homework and tests were also conducted over the Internet. For comparison purposes, a traditional Statics class was conducted by the same instructors (the authors) at PSTCC using traditional delivery methods with oncampus lectures and office hours. The two classes covered the same material at the same rate, and students took the same multiple choice tests and final exam. Both classes had access to identical course content on the Internet as well as a standard print textbook. Comparison of the test results of the two classes shows that the online delivery of basic engineering content through the Internet provides as good, if not better, education than traditional delivery methods. In summary, the online class students performed better on the exams by over a half-grade level.

Nuclear Engineering and Technology, 2017

Nuclear Technology, 2019

This paper discusses the design of a fast spectrum subcritical assembly utilizing a genetic algor... more This paper discusses the design of a fast spectrum subcritical assembly utilizing a genetic algorithm. The facility proposed in this paper would be a flexible platform for expanding the knowledge of fast spectrum neutron cross sections needed for next-generation fast reactor designs. The Fast Neutron Source (FNS) would be composed of both a fast and a thermal region to minimize the amount of uranium fuel and reduce overall material costs while maintaining flexibility for many potential fast neutron crosssection experiments. The FNS would be customizable and interchangeable down to 1 × 1 × 10-in.-volume sections. An optimal core design requires the adjustment of many factors to both reduce the cost and accurately reproduce the spectra of interest during an experiment. A genetic algorithm was developed to optimize this complex design problem while reducing design time and expert judgment. The genetic algorithm was able to vary multiple design factors in an unattended fashion from a random initial population of designs and arrived at a design comparable to an expertly designed assembly.

Nuclear Technology, 2011

Using Hilbert transforms, we establish two families of sum rules involving Bessel moments, which ... more Using Hilbert transforms, we establish two families of sum rules involving Bessel moments, which are integrals associated with Feynman diagrams in two-dimensional quantum field theory. With these linear relations among Bessel moments, we verify and generalize two conjectures by Bailey-Borwein-Broadhurst-Glasser and Broadhurst-Mellit.

Applied Computational Intelligence - Proceedings of the 6th International FLINS Conference, 2004

The investigation and application of on-line monitoring programs has been ongoing for over two de... more The investigation and application of on-line monitoring programs has been ongoing for over two decades by the U.S. nuclear industry and researchers. To this date, only limited pilot installations have been demonstrated and the original objectives have changed significantly. Much of the early work centered on safety critical sensor calibration monitoring and reduction. The current focus is on both sensor and equipment monitoring. This paper presents the major lessons learned that contributed to the lengthy development process including model development and implementation issues, and the results of a recently completed cost benefit analysis.

Power Systems, 2002

Most data-based predictive modeling techniques have an inherent weakness in that they may give un... more Most data-based predictive modeling techniques have an inherent weakness in that they may give unstable or inconsistent results when the predictor data is highly correlated. Predictive modeling problems of this design are usually under constrained and are termed ill-posed. This paper presents several examples of ill-posed diagnostic problems and regularization methods necessary for getting accurate and consistent prediction results. The examples include plant-wide sensor calibration monitoring and the inferential sensing of nuclear power plant feedwater flow using neural networks, and non-linear partial least squares techniques, and linear regularization techniques implementing ridge regression and informational complexity measures.

International Journal of Intelligent Systems, 2002

Most data-based predictive modeling techniques have an inherent weakness in that they may give un... more Most data-based predictive modeling techniques have an inherent weakness in that they may give unstable or inconsistent results when the predictor data is highly correlated. Predictive modeling problems of this design are usually under constrained and are termed ill-posed. This paper presents several examples of ill-posed surveillance and diagnostic problems and regularization methods necessary for getting accurate and consistent prediction

AAAI Fall Symposium on Artificial Intelligence for …, 2007

Most modern prognostic algorithms are founded on a simple abstraction of device degradation, for ... more Most modern prognostic algorithms are founded on a simple abstraction of device degradation, for an individual device there exists a degradation signal that progresses along a unique path until it crosses a critical failure threshold. While this abstraction has been shown to be valid for well understood failure modes under controlled stress conditions, its viability in "real world" devices being exposed to "real world" stresses is questionable. Because the complexity of degradation should scale in a similar fashion as the complexity of the device, the applicability of a simple abstraction of degradation is increasingly arguable for modern devices. This paper will propose an alternative to the current abstraction of degradation, which is founded on the premise that degradation data should be allowed to speak for itself. In this way, many different forms of information can be incorporated into a prognoser's estimate of a device's remaining useful life (RUL). More specifically, this paper will outline a methodology for implementing a dynamic prognoser that can be incrementally trained to learn general (physical model output, expert opinion, etc.) and specific ("real world" data) degradation trends. This work will demonstrate the viability of the proposed method by applying a particular embodiment, namely the path classification and estimation (PACE) model, to data collected from a deep-well oil exploration drill. To begin, expert opinion will be used to develop a PACE prognoser. Next, data collected from individual drills will be used to incrementally train the prognoser to learn specific degradation trends.

by Chris Black, J. Wesley Hines, and …, 1997

An autoassociative artificial neural network (AANN) instrument channel monitoring technique has b... more An autoassociative artificial neural network (AANN) instrument channel monitoring technique has been developed for sensor and associated instrument channel online calibration verification. Several AANN models, each modeling a group of interrelated signals, are ...

Ideally, health monitoring of new, complex engineering systems should occur from initial operatio... more Ideally, health monitoring of new, complex engineering systems should occur from initial operation to decommissioning. Health monitoring typically involves a suite of modules, including system monitoring, fault detection, fault diagnostics, and system prognostics. However, for systems which have not yet operated, this is challenging. Most available health monitoring modules are empirically based, meaning they are derived from available historic data. For new system designs, such data simply does not exist. This research proposes an adaptive modeling system which initially builds empirical models

2008 Annual Reliability and Maintainability Symposium, 2008

ABSTRACT Equipment, process, and system prognostic techniques can be classified as belonging to o... more ABSTRACT Equipment, process, and system prognostic techniques can be classified as belonging to one of three major classes of methods: 1) conventional reliability-based using failure times (Weibull), 2) population based with environmental considerations (e.g. proportional hazards modeling), and 3) individual based (e.g. general path model). A new individual-based prognostic algorithm, termed the path classification and estimation (PACE) model, has been developed and is based entirely on failure data. This model recasts the general path model (GPM), which is the foundation of the majority of the modern individual based prognosis algorithms, as a classification problem, where a current device's degradation path is classified according to a series of exemplar paths and the results of the classification are used to estimate the remaining useful life (RUL) of the device. The requirement of the existence of a failure threshold is removed, thereby enabling the PACE to be applied to ldquoreal worldrdquo systems, where a single failure threshold is not likely to occur. If the failure threshold is known, simple formatting may be applied to the degradation paths such that they can be easily used with the PACE. The newly proposed method was applied to data collected from the hydraulic steering system of a drill used for deep oil exploration with the objective of detecting, diagnosing, and prognosing faults. The PACE was used to predict the RUL for several failure modes using actual data. For this work, a three tiered architecture was implemented, where conventional reliability methods were used to estimate the population-based RUL, PACE population-based prognosers were trained to map the cause of a failure mode to the RUL, and PACE individual prognosers were trained to map the effects of a failure mode to the RUL. It was found that the population based prognoser produced RUL estimates with large errors (75 hours) and uncertainties (261 hours). The individual prognosers were found to si- gnificantly outperform the population based prognoser, with errors ranging from 1.2 to 11.4 hours with 95% confidence intervals ranging from 0.67 to 32.02 hours.

Energy system on-line-monitoring is becoming a crucial component of improving safety, reliability... more Energy system on-line-monitoring is becoming a crucial component of improving safety, reliability, and profitability. The Holy Grail is the development prognostic methodologies to accurately predict the Remaining Useful Life (RUL) of a system or component for predictive maintenance and effective risk mitigation. Calculating precise RUL estimates requires both system specific maintenance information and performance data to develop representative lifecycle models. Current conventional prognostic methods focus on process data and do not utilize maintenance data to directly influence the modeling and data analysis. However, equipment maintenance impacts future system degradation and is dependent on the maintenance actions taken. Differences in the amount of degradation removed from a system are common for repaired equipment compared to replacements. This talk discusses methods of incorporating maintenance information into Lifecycle Prognostics and the effect it has on prediction error a...

The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life ... more The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life of individual systems or components based on their use and performance. This class of prognostic algorithms is termed effects-based, or Type III, prognostics. A unit-specific prognostic model, called the General Path Model, involve identifying an appropriate degradation measure to characterize the system's progression to failure. A functional fit of this parameter is then extrapolated to a pre-defined failure threshold to estimate the remaining useful life of the system or component. This paper proposes a specific formulation of the General Path Model with dynamic Bayesian updating as one effects-based prognostic algorithm. The method is illustrated with an application to the prognostics challenge problem posed at PHM '08.

Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 2014

As nuclear power plants seek to extend their licenses and maintain a high level of performance an... more As nuclear power plants seek to extend their licenses and maintain a high level of performance and safety, online monitoring and assessment of system degradation are becoming a crucial consideration. A goal of the DOE Light Water Reactor Sustainability program is the accurate estimation of the remaining useful life of nuclear power plant systems, structures, and components. Effective prognostic systems should seamlessly predict the remaining useful life from beginning of component life to end of component life, so-called Lifecycle Prognostics. When a component is first put into operation, the only information available may be past failure times of similar components or the expected distribution of failure times derived from reliability analyses of these data (Type I Prognostics). These data provide an estimated life for the average component operating under average usage conditions. As the component operates, it begins to consume its available life at a rate largely influenced by th...

Applied Artificial Intelligence, 2006

2008 International Conference on Prognostics and Health Management, 2008

Prognostic algorithms can be divided into three major categories. The most basic methods model th... more Prognostic algorithms can be divided into three major categories. The most basic methods model the component or system reliability using failure time data and conventional models such as the Weibull. When information pertaining to the operating condition and environmental stressors are available, stress-based techniques can be used. The third type of prognostics is termed effects-based. It is truly an individual

Proceedings of International Conference on Neural Networks (ICNN'96)

Multilayer feedforward neural networks with sigmoidal activation functions have been termed “univ... more Multilayer feedforward neural networks with sigmoidal activation functions have been termed “universal function approximators”. Although these types of networks can approximate any continuous function to a desired degree of accuracy, this approximation may require an inordinate number of hidden nodes and is only accurate over a finite interval. These short comings are due to the standard multilayer perceptron's (MLP) architecture

2005 Annual Conference Proceedings

This paper presents the methods and results of delivering a basic Statics course to Pellissippi S... more This paper presents the methods and results of delivering a basic Statics course to Pellissippi State Technical Community College (PSTCC) students located in Knoxville, Tennessee over the Internet. All aspects of the course, including textbook, lectures, class meetings, student discussions, homework and tests were conducted through the Internet. The online course material included animations, simulations, narrations, and graphics. Homework and tests were also conducted over the Internet. For comparison purposes, a traditional Statics class was conducted by the same instructors (the authors) at PSTCC using traditional delivery methods with oncampus lectures and office hours. The two classes covered the same material at the same rate, and students took the same multiple choice tests and final exam. Both classes had access to identical course content on the Internet as well as a standard print textbook. Comparison of the test results of the two classes shows that the online delivery of basic engineering content through the Internet provides as good, if not better, education than traditional delivery methods. In summary, the online class students performed better on the exams by over a half-grade level.

Nuclear Engineering and Technology, 2017

Nuclear Technology, 2019

This paper discusses the design of a fast spectrum subcritical assembly utilizing a genetic algor... more This paper discusses the design of a fast spectrum subcritical assembly utilizing a genetic algorithm. The facility proposed in this paper would be a flexible platform for expanding the knowledge of fast spectrum neutron cross sections needed for next-generation fast reactor designs. The Fast Neutron Source (FNS) would be composed of both a fast and a thermal region to minimize the amount of uranium fuel and reduce overall material costs while maintaining flexibility for many potential fast neutron crosssection experiments. The FNS would be customizable and interchangeable down to 1 × 1 × 10-in.-volume sections. An optimal core design requires the adjustment of many factors to both reduce the cost and accurately reproduce the spectra of interest during an experiment. A genetic algorithm was developed to optimize this complex design problem while reducing design time and expert judgment. The genetic algorithm was able to vary multiple design factors in an unattended fashion from a random initial population of designs and arrived at a design comparable to an expertly designed assembly.

Nuclear Technology, 2011

Using Hilbert transforms, we establish two families of sum rules involving Bessel moments, which ... more Using Hilbert transforms, we establish two families of sum rules involving Bessel moments, which are integrals associated with Feynman diagrams in two-dimensional quantum field theory. With these linear relations among Bessel moments, we verify and generalize two conjectures by Bailey-Borwein-Broadhurst-Glasser and Broadhurst-Mellit.