Knowledge-based Process Control Research Articles

Straightening rollformed profiles by controlled partial rolling

Roll forming is a sheet metal forming process in which flat sheets are incrementally formed into a desired profile geometry. The process is characterized by high material utilization and output. Along with these advantages, process-related longitudinal strains occur, causing profile defects such as bowing and twisting of the profile, which require subsequent straightening. Conventional straightening methods use overbending and twisting to eliminate existing profile defects. However, this approach is purely experience-based and iterative due to the complex interactions between the various bending and twisting operations. Therefore, a closed-loop control of the straightening process is currently not implemented. This can be remedied by the novel knowledge-based straightening method of partial rolling.In this paper, the basic idea of partial rolling will be introduced and the idea of knowledge-based process control will be demonstrated. A simulation model is then presented to test different process control strategies and evaluate their impact on the profile defects. Finally, an outlook is given on how a closed-loop controlled partial rolling process could be set up in a real process.

Automatic Process Control in Press Shops

The manufacturing of automotive body components in press lines is a sensitive process. The quality characteristics of body components vary. These fluctuations are rooted in the fact that the factors influencing the component quality are varying, e.g., fluctuations of batches regarding material quality, abrasion or heating of the tool during the production cycle. If a certain quality characteristic exceeds a predefined range an intervention in the process is necessary. This intervention is based upon the subjective know-how of the machine operator. Objective information about the state of the process, like tool temperature or the material quality of the semi-finished product is not available. Therefore, a lack of knowledge emerges in the interrelations between the tuning parameters of the system press-tool and the component quality during different stages of the process (material quality, temperature…). In this paper a complete concept for an automatic process control in press shops is described. The concept will be realized in a pilot plant for mass production in the press shop of AUDI AG. The mechanisms of occurrence of quality defects are shown in the paper, as well as the essential factors influencing the quality during the mass production of body components in the automotive industry and their variation. A sensor-system for continuous measurement of influencing variables during the mass production is presented. The key element of the concept is the non-destructive identification of material-properties for every single blank. By associating the sensor-data with the respective quality, a knowledge-based process control can be realized. The purpose is to create a failure prediction algorithm and make optimal adjustments for each stroke of the moulding press, respectively. The potential of existing actuators in modern press lines as well as new, tool integrated proposals for actuators are highlighted.

Knowledge based process control supervision and diagnosis: the AEROLID approach

The artificial intelligence incidence in process control, although an active area in the researchers community and even with some implementations at industrial environment, is not sufficiently evaluated in numerical terms for the long term. The present article shows such an evaluation of a knowledge based system, developing supervisory control tasks in the sugar production from sugar-beet, and paying particular attention to fault detection and diagnosis. A way of conceiving supervision for continuous processes is presented and supported with this industrial application. The expert system carrying out supervisory tasks operates in a VAX ® workstation, directly over the distributed control system. The expert system development tool is G2 ® which has real-time facilities. Although the core system was developed in G2, it also consists of some external modules because it combines both analytical and artificial intelligence problem resolution techniques. The global architecture, as well as the implementation details of the modules necessary for fault identification, are presented altogether with the experimental results obtained from the factory field.

A Result on the Generalized Predictive Controller and its Role in Knowledge-Based Process Control

ABSTRACTIn this paper, we present a key new result relating the choice of the parameters of the Generalized Predictive Controller for plants with similar dynamics but with different dead-times. The result is proved in the paper, and a discussion is provided on how the result may be used as “deep knowledge” in the context of a knowledge-based predictive controller. An application example is also included where the result is utilized in the red-time control of the fluid level of a coupled-tanks system.

Temporal management using relative time in knowledge-based process control

In this paper, a knowledge-based approach is proposed for the management of temporal information in process control. A common-sense theory of temporal constraints over processes/events, allowing relative temporal knowledge, is employed here as the temporal basis for the system. This theory supports duration reasoning and consistency checking, and accepts relative temporal knowledge which is in a form normally used by human operators. An architecture for process control is proposed which centres on an historical database consisting of events and processes, together with the qualitative temporal relationships between their occurrences. The dynamics of the system is expressed by means of three types of rule: database updating rules, process control rules, and data deletion rules. An example is provided in the form of a life scheduler, to illustrate the database and the rule sets. The example demonstrates the transitions of the database over time, and identifies the procedure in terms of a state transition model for the application. The dividing instant problem for logical inference is discussed with reference to this process control example, and it is shown how the temporal theory employed can be used to deal with the problem.

A "CLOSED LOOP" SYSTEM FOR THE REALTIME CONTROL OF SOLDER PASTE STENCIL PRINTING

The move towards fine and ultrafine pitch components in surface mount Printed Wiring Board (PWB) assembly has increased the importance of the solder paste printing process. The quality of the solder paste deposit is influenced by factors such as the characteristics of the board, the solder paste selected, the printer used, etc. These inter-related process and material factors complicate the procedures and the reasoning used to identify and set-up the stencil printing parameters. The increased use of ultrafine pitch components (with lead pitch down to 0.4 mm) in PWB assembly has almost made it mandatory to inspect the solder paste deposit prior to component placement. This inspection process provides the necessary information to correct the printing process in a realtime fashion if printing is out of control or is exhibiting a tendency to go out of control.This research designed and developed a "closed-loop" system for realtime control of the solder paste stencil printing process. This knowledge based process control system interfaces with a stencil printer and a laser based solder paste inspection system to form the "closed-loop". This "closed loop" realtime system insures both proactive and reactive control of the stencil printing process.

Intelligent decision support. Handbook of applications and advances of the rough sets theory

Preface Z. Pawlak. Scope and Goals of the Book R. Slowinski. Part I: Applications of the Rough Sets Approach to Intelligent Decision Support. 1. LERS -- A System for Learning from Examples Based on Rough Sets J.W. Grzymala-Busse. 2. Rough Sets in Computer Implemetation of Rule-Based Control of Industrial Process A. Mrozek. 3. Analysis of Diagnostic Symptoms in Vibroacoustic Diagnostic by Means of the Rough Sets Theory R. Nowicki, R. Slowinski, J. Stefanowski. 4. Knowledge-Based Process Control Using Rough Sets A.J. Szladown, W.P. Ziarko. 5. Acquisition of Control Algorithms from Operation Data W.P. Ziarko. 6. Rough Classification of HSV Patients K. Slowinski. 7. Surgical Wound Infection -- Conducive Factors and their Mutual Dependencies M. Kandulski, J. Marciniec, K. Tukallo. 8. Fuzzy Inference System Based on Rough Sets and its Application to Medical Diagnosis H. Tanaka, H. Ishibuchi, T. Shigenaga. 9. Analysis of Structure-Activity Relationships of Quaternary Ammonium Compounds J. Krysinski. 10. Rough Sets-Based Study of Voter Preference in 1988 U.S.A. Presidential Election M. Hadjimichale, A. Wasilewska. 11. An Application of Rough Set Theory in the Control of Water Conditions in a Polder A. Reinhard, B. Stawski, T. Weber, U. Wybraniec-Skardowska. 12. Use of 'Rough Sets' Methods to draw Premonitory Factors for Earthquakes by emphasising Gas Geochemistry: The Case of a Low Seismic Activity Context in Belgium J. Teghem, J.-M. Charlet. 13. Rough Sets and Some Aspects of Logic Synthesis T. Luba,J. Rybnik. Part II: Comparison with Related Methodologies. 1. Putting Rough Sets and Fuzzy Sets together D. Dubois, H. Prade. 2. Applications of Fuzzy-Rough Classification to Logics A. Nakamura. 3. Comparison of the Rough Sets Approach and Probalistic Data Analysis Techniques on a Common Set of Medical Data E. Krusinska, A. Babic, R. Slowinski, J. Stefanowski. 4. Some Experiments to Compare Rough Sets Theory and Ordinal Statistical Methods J. Teghem, M. Benjelloun. 5. Topological and Fuzzy Rough Sets T. Lin. 6. On Convergence of Rough Sets L.T. Polkowski. Part III: Further Developments. 1. Maintenance of Knowledge in Dynamic Systems M.E. Orlowska, M.W. Orlowski. 2. The Discernibility Matrices and Functions in Information Systems A. Skowron, C. Rauszer. 3. Sensitivity of Rough Classification to Changes in Norms of Attributes K. Slowinski, R. Slowinksi. 4. Discretization of Condition Attributes Space A. Lenarcik, Z. Piasta. 5. Consequence Relations and Information Systems D. Vakarelov. 6. Rough Grammar for High Performance Management of Processes on a Distributed System Z.M. Wojcik, B.E. Wojcik. 7. Learning Classification Rules from Database in the Context of Knowledge-Acquisition and Representation R. Yasdi. 8. 'RoughDAS' and 'RoughClass' Software Implementations of the Rough Sets Approach R. Slowinski, J. Stefanowski. Appendix: Glossary of Basic Concepts. Subject Index.

An Investigation into the Control of Brushed Yarn Properties: The Application of Machine Vision and Knowledge-based Systems. Part III: The Knowledge-based Process Control System

The paper describes the development of a computer machine interface utilising software development and hardware devices and the incorporation of a machine vision system for on-line quality and proc...

On the limits of expert systems and engineering models in process control

Abstract The review is based on an analysis of current literature of expert systems and of system engineering models in dynamic process control. It starts with an analysis of the mental operations and cognitive requirements needed for supervisory control. Mental models are discussed as a function of situational requirements as well as of personal strategies. Systems engineering models and expert systems are briefly described and their function as decision support tools evaluated. Criteria are the overall functionality, similarity of knowledge bases and reasoning strategies of the human and the support system, adaptability to the operator's skill level and self-explanation of the support system in the interaction mode. As a result, system engineering models are only of limited value for knowledge-based process control. Expert systems seem to be very valuable tools for augmenting human decision making in process control, if the interaction problem can be solved.

An Investigation of the “Von Restorff” Phenomenon in Post-Test Workload Ratings

The present experiment was designed to examine the possibility of a “von Restorff” effect occurring during post-task ratings of difficulty, effort and workload as a result of performing tasks containing an isolated period of high workload. The task employed was a hovercraft simulation which combined elements of skill-based tracking and rule- and knowledge-based process control. The performance of five separate tasks was compared. In the first, subjects performed missions which presented a relatively low level of workload from throughout while the second presented a high level of workload from throughout. The final three conditions were designed to produce isolated peaks in workload which occurred either early, midway or late in the mission. Results demonstrated that ratings on the three scales increased steadily as the peak approached the end of the mission. Further, the ratings for the early increase were only slightly greater than those reported for the task containing a low level of workload throughout. An analysis of error rates demonstrated that there was evidence for the dissociation of performance and workload which was especially apparent when the increase occurred at the beginning of the mission.