J. Wesley Hines

Use of Autoassociative Neural Networks for Signal Validation

Recently, the use of Autoassociative Neural Networks (AANNs) to perform on-line calibration monitoring of process sensors has been shown to not only be feasible, but practical as well. This paper summarizes the results of applying AANNs to instrument surveillance and calibration monitoring at Florida Power Corporations Crystal River #3 Nuclear Power Plant and at the Oak Ridge National Laboratory High Flux Isotope Reactor. In both cases sensor drifts are detectable at a nominal level of 0.5% instruments full scale range. This paper will discuss the selection of a five layer neural network architecture, a robust training paradigm, the input selection criteria, and a retuning algorithm.

Current Computational Trends in Equipment Prognostics

The article overviews current trends in research studies related to reliability prediction and prognostics. The trends are organized into three major types of prognostic models: failure data models, stressor models, and degradation models. Methods in each of these categories are presented and examples are given. Additionally, three particular computational prognostic approaches are considered; these are Markov chain-based models, general path models, and shock models. A Bayesian technique is then presented which integrates the prognostic types by incorporate prior reliability knowledge into the prognostic models. Finally, the article also discusses the usage of diagnostic/prognostic predictions for optimal control.

Regularization of Feedwater Flow Rate Evaluation for Venturi Meter Fouling Problem in Nuclear Power Plants

Abstract Inferential sensing is a method that can be used to evaluate parameters of a physical system based on a set of measurements related to these parameters. The most common method of inferential sensing uses mathematical models to infer a parameter value from correlated sensor values. However, since inferential sensing is an inverse problem, it can produce inconsistent results due to minor perturbations in the data. This research shows that regularization can be used in inferential sensing to produce consistent results. Data from Florida Power Corporation’s Crystal River nuclear power plant (NPP) are used to give an important example of monitoring NPP feedwater flow rate.

Regularization Methods for Inferential Sensing in Nuclear Power Plants

Inferential sensing is the use of information related to a plant parameter to infer its actual value. The most common method of inferential sensing uses a mathematical model to infer a parameter value from correlated sensor values. Collinearity in the predictor variables leads to an ill posed problem that causes inconsistent results when data based models such as linear regression and neural networks are used. This chapter presents several linear and non-linear inferential sensing methods including linear regression and neural networks. Both of these methods can be modified from their original form to solve ill posed problems and produce more consistent results.

Process and equipment monitoring methodologies applied to sensor calibration monitoring

Traditionally, the calibration of safety critical nuclear instrumentation has been performed at each refueling cycle. However, many nuclear plants have moved toward condition-directed rather than time-directed calibration. This condition-directed calibration is accomplished through the use of on-line monitoring. On-line monitoring (OLM) uses an autoassociative empirical modeling architecture to assess instrument channel performance. An autoassociative architecture predicts a group of correct sensor values when supplied a group of sensor values that is usually corrupted with process and instrument noise, and could also contain faults such as sensor drift or complete failure. The Process and Equipment Monitoring (PEM) Toolbox, which was developed at The University of Tennessee, is a set of MATLAB based tools, which have been developed to support the design of process and equipment condition monitoring systems. Its purpose is to provide the necessary functionality so that different empirical modeling and uncertainty estimation methods may be investigated and compared. This paper discusses the purpose and architecture of the PEM Toolbox, as well as presenting an example application to power plant sensor calibration monitoring using actual plant data. Copyright

Online Sensor Calibration Monitoring Uncertainty Estimation

Abstract Empirical modeling techniques have been applied to online process monitoring to detect equipment and instrumentation degradations. However, few applications provide prediction uncertainty estimates, which can provide a measure of confidence in decisions. This paper presents the development of analytical prediction interval estimation methods for three common nonlinear empirical modeling strategies: artificial neural networks, neural network partial least squares, and local polynomial regression. The techniques are applied to nuclear power plant operational data for sensor calibration monitoring, and the prediction intervals are verified via bootstrap simulation studies.

Prediction interval estimation techniques for empirical modeling strategies and their applications to signal validation tasks

The basis of this work was to evaluate both parametric and non-parametric empirical modeling strategies applied to signal validation or on-line monitoring tasks. On-line monitoring methods assess signal channel performance to aid in making instrument calibration decisions, enabling the use of condition-based calibration schedules. The three non-linear empirical modeling strategies studied were: artificial neural networks (ANN), neural network partial least squares (NNPLS), and local polynomial regression (LPR). The evaluation of the empirical modeling strategies includes the presentation and derivation of prediction intervals for each of three different model types studied. An estimate and its corresponding prediction interval contain the measurements with a specified certainty, usually 95%. The prediction interval estimates were compared to results obtained from bootstrapping via Monte Carlo resampling, to validate their expected accuracy. Properly determined prediction interval estimates were obtained that consistently captured the uncertainty of the given model such that the level of certainty of the intervals closely matched the observed level of coverage of the prediction intervals over the measured values. In most cases the expected level of coverage of the measured values within the prediction intervals was 95%. The prediction intervals were required to perform adequately under conditions of model misspecification. The results also indicate that instrument channel drifts are identifiable by observing the drop in the level of coverage of the prediction intervals to relatively low values, e.g. 30%. A comparative evaluation of the different empirical models was also performed. The evaluation considers the average estimation errors and the stability of the models under repeated Monte Carlo resampling. The results indicate the large uncertainty of ANN models applied to collinear data, and the utility of the NNPLS model for the same purpose. While the results from the LPR models remained consistent for data with or without collinearity, assuming proper regularization was applied. All of the methods studied herein were applied to a simulated data set for an initial evaluation of the methods, and data from two different U.S. nuclear power plants for the purposes of signal validation for on-line monitoring tasks.

A novel approach to process modeling for instrument surveillance and calibration verification

Abstract This work presents an empirical modeling approach combining a bilinear modeling technique, partial least squares, with the universal function approximation abilities of single hidden layer nonlinear artificial neural networks. This approach, referred to as neural network partial least squares (NNPLS), is compared to the common autoassociative artificial neural network. The NNPLS model is embedded into a graphical user interface and implemented at the Electrical Power Research Institute’s Instrumentation and Control Center located at Tennessee Valley Authority’s Kingston fossil power plant. Results are presented for 51 process signals with an average absolute estimation error of ~1.7% of the mean value, and sample drift detection performances are shown.

Heuristic, systematic, and informational regularization for process monitoring

Most data‐based predictive modeling techniques have an inherent weakness in that they might 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 article presents several examples of ill‐posed surveillance and 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 the following techniques: neural networks, nonlinear partial least squares techniques, linear regularization techniques implementing ridge regression, and informational complexity measures.

Verification of helical tomotherapy delivery using autoassociative kernel regression.

Quality assurance (QA) is a topic of major concern in the field of intensity modulated radiation therapy (IMRT). The standard of practice for IMRT is to perform QA testing for individual patients to verify that the dose distribution will be delivered to the patient. The purpose of this study was to develop a new technique that could eventually be used to automatically evaluate helical tomotherapy treatments during delivery using exit detector data. This technique uses an autoassociative kernel regression (AAKR) model to detect errors in tomotherapy delivery. AAKR is a novel nonparametric model that is known to predict a group of correct sensor values when supplied a group of sensor values that is usually corrupted or contains faults such as machine failure. This modeling scheme is especially suited for the problem of monitoring the fluence values found in the exit detector data because it is able to learn the complex detector data relationships. This scheme still applies when detector data are summed over many frames with a low temporal resolution and a variable beam attenuation resulting from patient movement. Delivery sequences from three archived patients (prostate, lung, and head and neck) were used in this study. Each delivery sequence was modified by reducing the opening time for random individual multileaf collimator (MLC) leaves by random amounts. The error and error-free treatments were delivered with different phantoms in the path of the beam. Multiple autoassociative kernel regression (AAKR) models were developed and tested by the investigators using combinations of the stored exit detector data sets from each delivery. The models proved robust and were able to predict the correct or error-free values for a projection, which had a single MLC leaf decrease its opening time by less than 10msec. The model also was able to determine machine output errors. The average uncertainty value for the unfaulted projections ranged from 0.4% to 1.8% of the detector signal. The low model uncertainty indicates that the AAKR model is extremely accurate in its predictions and also suggests that the model may be able to detect errors that cause the fluence to change by less than 2%. However, additional evaluation of the AAKR technique is needed to determine the minimum detectable error threshold from the compressed helical tomotherapy detector data. Further research also needs to explore applying this technique to electronic portal imaging detector data.