M.J.S Vancoille

Practical Building of Expert Systems for Materials Selection and Corrosion Problems

This paper will illustrate some of the principles of building expert systems. As an example, the expert system PRIME which is one of the most advanced realizations in its field, will be discussed. This expert system has been developed at the University of Leuven and deals with material selection and corrosion problems in the chemical process industry. In its implementation, many of the characteristics of second generation expert systems can be found such as multiple reasoning models, and multiple reasoning strategies. Furthermore, PRIME has a highly advanced, graphics-based and user-friendly interface which takes man machine interfaces one step further than most present day developments.

Delivering materials engineering information using hypermedia systems

The availability of reliable and detailed information on the properties, behavior, and characteristics of materials is of crucial importance to engineers during materials engineering processes, that is, designing, manufacturing, installing, and maintaining industrial components and installations. The challenges facing computerized information tools that must provide effective support for these materials engineering processes are threefold: to develop materials information systems for use by engineers that allows structured, standardized access to diverse, highly interconnected information on materials. In this paper, we discuss how the use of hypermedia tools and techniques allows us to address these three challenges. We contend that the rich diversity and interconnectedness of materials information can only be adequately captured using the hypermedia data model. We also discuss how, using the interface functionals of a hypermedia tool, we can tailor the interface of a materials information system to the task-related needs of the corrosion engineer. We illustrate this by examining two hypermedia materials information systems that have been developed in cooperation with our research group: the Active Library® on Corrosion (ALC) and CONNAISSANCE. The ALC is a CD-ROM based hypermedia system with a large but fixed collection of corrosion engineering information that can be consulted on the engineers personal computer. CONNAISSANCE is a unique hypertext system with continuously updated materials engineering information that can be accessed and consulted on-line.

A.I.-Based Corrosion Risk Analysis in Oil Refinery Operation

The assessment of corrosion and related risk behavior in oil refinery operation is both based on practical experience and basic understanding of the prevailing corrosion mechanisms. It was thought necessary to automate to a larger extent the process of corrosion risk analysis. This automated process would enable corrosion and material experts to give a faster and more uniform evaluation of corrosion risks. The usefulness of artificial intelligence techniques in this domain was investigated. Some preliminary studies have been made and prototypes built to demonstrate what can be done with state of the art technology. Right from the conception of the project, much attention was given to produce a useful working product.