James W. Howse

Least squares estimation techniques for position tracking of radioactive sources

This paper describes least squares estimation algorithms used for tracking the physical location of radioactive sources in real time as they are moved around in a facility. We present both recursive and moving horizon nonlinear least squares estimation algorithms that consider both the change in the source location and the deviation between measurements and model predictions. The measurements used to estimate position consist of four count rates reported by four different gamma ray detectors. There is an uncertainty in the source location due to the large variance of the detected count rate, and the uncertainty in the background count rate. This work represents part of a suite of tools which will partially automate security and safety assessments, allow some assessments to be done remotely, and provide additional sensor modalities with which to make assessments.

Blast furnace stove control

This paper outlines the process model and model-based control techniques implemented on the hot blast stoves for the No. 7 Blast Furnace at the Inland Steel facility in East Chicago, Indiana. A detailed heat transfer model of the stoves is developed. It is then used as part of a predictive control scheme to determine the minimum amount of fuel necessary to achieve the blast air requirements. The controller also considers maximum and minimum temperature constraints within the stove.

Comparison of recursive estimation techniques for position tracking radioactive sources

This paper compares the performance of recursive state estimation techniques for tracking the physical location of a radioactive source based on radiation measurements obtained from a series of detectors at fixed locations. Specifically, the first order, iterated, and a second order extended Kalman filter performance is compared to nonlinear least squares estimation. The results of this study indicate that least squares estimation significantly outperforms the extended Kalman filter implementations in this application due to the nature of the model nonlinearities.

Model-based control of a thermal regenerator. Part 1: dynamic model

Abstract This paper describes the dynamic process model and solution technique developed for the hot blast stoves used with the No. 7 blast furnace at the Ispat Inland Steel facility in East Chicago, IN, USA. A detailed, distributed parameter, heat transfer model of this thermal regenerator system is developed and verified using plant data. The model is capable of predicting accurately the temperature and energy content of the stoves during the thermal regenerative cycles. It was developed for a predictive controller that determines the minimum amount of fuel necessary to achieve the energy requirements from the system.

Model-based control of a thermal regenerator. Part 2: control and estimation

Abstract This paper describes the model-based control and estimation techniques implemented on the hot blast stoves for the number 7 blast furnace at the Ispat Inland Steel facility in East Chicago, IN. The process model is a detailed heat transfer model of this thermal regenerator system used as part of a predictive control scheme to determine the minimum amount of fuel necessary to achieve the energy requirements. Batch nonlinear least squares estimation is used to update the predicted temperature profile and heat transfer coefficients. These estimated parameters are then used by the model-based controller to determine the minimum fuel required for the subsequent regenerative cycle.

An evaluation of detection metrics for an integrated catalyst controller and diagnostic monitor

Abstract An integrated, model-based methodology for three-way automotive catalyst control and diagnostic monitoring utilizing a limited integrator model with an adaptive integral gain is outlined in this work. This adaptive gain, which is a measure of the catalyst oxygen storage capacity, is used both by the controller to provide information on the dynamic catalyst behavior and by the diagnostic monitor to provide information on long-term catalyst deactivation and short-term emission control device failure. Nonparametric test statistics using various metrics computed from a moving window sample of the adaptive gain are compared to determine their ability to detect changes in catalyst system performance with a number of differently aged catalysts. These diagnostic monitoring metrics have been applied to 4.6 liter ULEV II gasoline engine data tested over an EPA Federal Test Procedure drive cycle.

Adaptive security systems - combining expert systems with adaptive technologies

The adaptive multisensor integrated security system (AMISS) uses a variety of computational intelligence techniques to reason from raw sensor data through an array of processing layers to arrive at an assessment for alarm/alert conditions based on human behavior within a secure facility. In this paper, we give an overview of the system and briefly describe some of the major components of the system. This system is currently under development and testing in a realistic facility setting.

Model-Based Fault Detection for Three-Way Automotive Catalyst Systems

Abstract A model-based three-way automotive catalyst monitoring and fault detection strategy is presented in this work. A simplified oxygen storage and reversible catalyst deactivation model is employed to predict the measured postcatalyst air fuel ratio. A fault is assumed to be present in the system when the current distribution of the post-catalyst air fuel ratio prediction error differs from the base operating distribution. Changes in the post-catalyst air fuel ratio prediction error distribution are indicative of both long-term catalyst poisoning effects and short-term emission control device failures. These changes are detected based on the results of a Kolmogorov-Smirnov test Using sampled cumulative distribution functions. A moving horizon approach is used to determine the current error distribution.

Lagrangian solution methods for nonlinear model predictive control

This work presents a simultaneous approach to the solution of the receding horizon, open-loop optimal model predictive control law for nonlinear systems using first-order Lagrangian methods. The nonlinear model considered is a general form of the initial value advective-diffusion parabolic partial differential equation. Others forms may be considered in a similar manner. The Lagrangian is formed from the discretized objective function, model and constraint equations. A finite volume approach is used to discretize the partial differential model equations. Inequality constraints on the model states and control inputs are handled with an active set method. The nonlinear equations resulting from the first order necessary conditions are then solved directly using a Newton-Krylov technique.

Probability density diagnostic metric for an integrated three-way catalyst controller and monitor

Abstract An integrated model-based methodology for three-way automotive catalyst control and diagnostic monitoring is presented in this work. The catalyst controller and monitor both utilize a limited integrator catalyst oxygen storage model with an adaptive integral gain. This adaptive catalyst gain, which is a measure of the catalyst oxygen storage capacity, is used by the controller to provide information on the dynamic catalyst behaviour and by the diagnostic monitor to provide information on long-term catalyst deactivation and short-term emission control device failure. A statistical classification technique based on the fraction of time that the catalyst gain values in a moving window are within a threshold of zero is employed as the test metric for on-board diagnostic monitoring. The performance of the catalyst monitor is demonstrated with experimental vehicle test data from a 4.6 l ULEV II gasoline engine operated over a series of Environmental Protection Agency Federal Test Procedure drive cycles with differently aged catalysts. Preliminary results indicate that it is possible to perform very accurate discrimination between catalyst operation, even near the on-board diagnostic detection threshold, using this technique.