Patricia A. S. Ralston

Computer-based monitoring and fault diagnosis: a chemical process case study

Abstract Principal component analysis (PCA) for process modeling and multivariate statistical techniques for monitoring, fault detection, and diagnosis are becoming more common in published research, but are still underutilized in practice. This paper summarizes an in-depth case study on a chemical process with 20 monitored process variables, one of which reflects product quality. Data from intervals of “good” operation times are used to determine a PCA model, and then data sets from intervals of “bad” operation times are compared to the model to detect the faulty variable and determine those variables responsible for the poor product quality. The analysis is performed using the PLS_Toolbox 2.01 with MATLAB. The methods used are reviewed summarily, and then results are shown based on typical application of the multivariate statistical techniques. An enhancement is made by using confidence limits on the residuals of each variable for fault detection rather than just confidence limits on an overall residual. Results show that the time required for fault detection is reduced. This approach was suggested in the literature, but its efficacy not demonstrated. Finally, ways to more effectively monitor processes and to more promptly detect and diagnose faults when they occur are identified.

High-Achieving High School Students and Not So High-Achieving College Students A Look at Lack of Self-Control, Academic Ability, and Performance in College

This study investigated the relationship among lack of self-control, academic ability, and academic performance for a cohort of freshman engineering students who were, with a few exceptions, extremely high achievers in high school. Structural equation modeling analysis led to the conclusion that lack of self-control in high school, as measured by the frequency of illegal and irresponsible behaviors, had an inverse relationship with first semester grade point average (GPA), whereas academic ability, as measured by ACT scores, had a positive relationship with college GPA. The correlation between the residual error for one of the indicators of self-control, homework behaviors in high school, and the residual error for first semester GPA was also significant. Research on the relationships between self-control, homework behaviors in high school, and performance in college should continue; meanwhile, parents and teachers would be advised to emphasize the importance of developing self-control and positive homework behaviors in academically advanced high school students.

Fuzzy logic control of aggregate production planning

Abstract Rinks used linguistic variables of the sort proposed by Zadeh for cost, inventory, production, and work force. He operated on these with fuzzy if-ten rules in a manner like that used by Mamdani and Assilan to control a steam engine. The Rinks membership functions and rules were subjective creations. Nevertheless, he obtained results that compared favorably to both prior and subsequent heuristic treatments of the HMMS data. We have developed a C language fuzzy logic controller (FLC) that uses the Rinks discrete membership functions and closely reproduces his results. We have used this FLC to investigate the effect of granularity of the discretization on the quality of solution. We have also used Wang-Mendel learning to improve the rule base. This paper discusses these results and suggests areas for further investigation.

Graphical enhancement to support PCA-based process monitoring and fault diagnosis.

Principal component analysis (PCA) for process modeling and multivariate statistical techniques for monitoring, fault detection, and diagnosis are becoming more common in published research, but are still underutilized in practice. This paper summarizes an in-depth case study on a chemical process with 20 monitored process variables, one of which reflects product quality. The analysis is performed using the PLS_Toolbox 2.01 with MATLAB, augmented with software which automates the analysis and implements a statistical enhancement that uses confidence limits on the residuals of each variable for fault detection rather than just confidence limits on an overall residual. The newly developed graphical interface identifies and displays each variables contribution to the faulty behavior of the process; and it aids greatly in analyzing results. The case study analyzed within shows that using the statistical enhancement can reduce the fault detection time, and the automated graphical interface implements the enhancement easily.

Tablet PCs in engineering mathematics courses at the J.B. Speed School of Engineering

In fall 2007, J.B. Speed School of Engineering at the University of Louisville joined the ranks of universities requiring the purchase of Tablet PCs for all new entering students. This article presents a description of how the Department of Engineering Fundamentals incorporated Tablet PCs into their instruction, a review of the literature pertaining to the use of Tablet PCs for instruction and preliminary survey results from the students in engineering mathematics courses at the end of the first year, after students had been exposed to Tablet PCs for 1 year. Results show that a large majority of students in the Department of Engineering Fundamentals agree that presentation of engineering mathematics material using the Tablet PCs and DyKnow software is a vast improvement over overhead projector, blackboard, or PowerPoint lectures and course packs. However, students are split as to whether the Tablet PC is something they actually want to use for their own note-taking. Finally, a plan for assessment of tablet impact on student learning is presented.

Analysis of population size in the accuracy and performance of genetic training for rule-based control systems

Abstract In off-line training of a rule-based controller, the significant measure of successful training is the quality of control provided by the generated rule set. In an adaptive or on-line control environment, performance is also measured by the ability to accurately maintain a satisfactory rule set, but within constraints on speed and/or resource availability. Very small population genetic algorithms, or microGAs, have been proposed as a means of capitalizing on the hill climbing characteristics of faster local optimizatiion techniques while requiring less memory and retaining much of the robustness of traditional, larger population genetic search. A traditional genetic algorithm and a similar microGA are developed and applied to two control problems. The performance of these algorithms is analyzed with respect to (1) the quality of the rules learned, (2) the rate at which learning occurs, and (3) the memory resources required during learning.

Fuzzy logic control of chip form during turning

Abstract Chips produced during lathe turning should be broken so as to prevent snarling and to enhance heat removal from the tool. At present, conservatively written part programs for computer numerically controlled lathes are fine tuned by a human machine operator to produce desirable chip length. K. W. Yee and coworkers have produced an acoustic emission chip form monitor. The use of such an acoustic emission sensor in a simulated chip length control system is described. In this paper a fuzzy logic controller is then substituted for the conventional linear controller.

Predicting performance in a first engineering calculus course: implications for interventions

At the University of Louisville, a large, urban institution in the south-east United States, undergraduate engineering students take their mathematics courses from the school of engineering. In the fall of their freshman year, engineering students take Engineering Analysis I, a calculus-based engineering analysis course. After the first two weeks of the semester, many students end up leaving Engineering Analysis I and moving to a mathematics intervention course. In an effort to retain more students in Engineering Analysis I, the department collaborated with university academic support services to create a summer intervention programme. Students were targeted for the summer programme based on their score on an algebra readiness exam (ARE). In a previous study, the ARE scores were found to be a significant predictor of retention and performance in Engineering Analysis I. This study continues that work, analysing data from students who entered the engineering school in the fall of 2012. The predictive validity of the ARE was verified, and a hierarchical linear regression model was created using math American College Testing (ACT) scores, ARE scores, summer intervention participation, and several metacognitive and motivational factors as measured by subscales of the Motivated Strategies for Learning Questionnaire. In the regression model, ARE score explained an additional 5.1% of the variation in exam performance in Engineering Analysis I beyond math ACT score. Students took the ARE before and after the summer interventions and scores were significantly higher following the intervention. However, intervention participants nonetheless had lower exam scores in Engineering Analysis I. The following factors related to motivation and learning strategies were found to significantly predict exam scores in Engineering Analysis I: time and study environment management, internal goal orientation, and test anxiety. The adjusted R2 for the full model was 0.42, meaning that the model could explain 42% of the variation in Engineering Analysis I exam scores.

Web-based grading of symbolic answers

This paper describes a system in which the computer algebra system Maple running on a remote server is used to check the correctness of symbolic answers submitted via the World Wide Web. The goal is to have a computer grade the kinds of symbolic algebraic, trigonometric, and calculus expressions that arise naturally in engineering and scientific applications rather than just numerical or multiple-choice options. The ability to grade symbolic answers can be used synchronously (on-line quizzes or exams under monitored conditions) or asynchronously (e.g., homework, tutorials, or readiness tests) for a broad range of quantitative courses and levels. A database keeps extensive, detailed records of student performance. This paper describes a system for symbolic grading via a Web browser. After giving the background and a brief overview of the system, it presents two simple computer-gradable problems and the steps used to grade an illustrative problem. The paper concludes with a brief summary of the work in progress toward a system that can be used for a variety of Web-based grading activities over a broad spectrum of on-site or remote quantitative courses.

Intelligent control of machines and processes

Abstract This review will consider computer based feedback control systems that rely on rule bases, knowledge bases, machine learning, and other artificial intelligence (AI) techniques. Traditional adaptive control is specifically excluded.