Knowledge Based Systems in Concurrent Engineering & Design

Design, especially concurrent design, is one of the most active areas in the use and

development of expert systems. In terms of computer support for design one of the primary roles is ‘to provide information in a simple and well structured form to support analysis and decision making.’, Miles & Swift (1994). Expert systems have the ability to do this and thus may be a great aid to concurrent engineering.

When it is considered that concurrent engineering or design takes into account considerations within manufacturing, testing, redesign use etc. it can be seen that the information requirements are immense. To this end Bowden & O ’Grady (1989) have identified the principal

requirements of a computerised system to support concurrent engineering:

• It should be flexible enough to allow the design problem to be approached from a variety of viewpoints

• It should allow the designer to design despite the absence of complete information • It should handle the large volume, variety and interdependence of life-cycle information • It should exhibit high performance in terms of speed and reliability

• It should readily interface to database management and CAD systems • It should have a good user interface and be able to explain itself in a manner

comprehensible to humans

• It should support design and environmental audits and be easily updateable as new information becomes available.

An expert system used in concurrent engineering / design should also display these

characteristics. These characteristics will be dictated to a certain extent by the way in which the

knowledge is represented within the system.

6.5.2 Knowledge-Based and Expert Systems for Design Support

A considerable amount of work has been carried out in the area of ES and KBS in design. From the building of such systems, Huang & Brandon (1992), to their use and implementation, Ishii & H ornberger (1992), and their impact on technical development, Chen (1991), many issues have been raised. Part of this research, although taking these issues into account looks mainly at how the use of expert system may structure and accelerate a new and emerging design discipline.

Schiebeler & Ehrlenspiel (1993) define the knowledge based system for design assistance as ‘a tool for the designer which supplies assistance in certain phases of the product development’. This assistance may be in one of many forms as discussed at the beginning of the chapter. There is one very important aspect of ES and KBS in design, that of flexibility. Klein (1993)

argues that the design engineer should have the possibility to ‘extend and modify the knowledge base for special purposes’ thus building in this large degree of flexibility.

In order that this flexibility be achieved there are a number of KBS approaches to engineering

design as discussed by Dym & Levitt (1991). They present their abstraction of a taxonomy of

methods for solving arrangement problems in increasing order of specificity.

• Analogy and mutation - uses case-based and analogical reasoning

• Assembly of solutions from elementary components - uses logical programming techniques, production rules or high level object oriented programming tools.

• Hierarchical generation, testing and evaluation of solutions - uses production rules, or frame-based representation as well as inferencing.

• Prototype selection and refinement - uses rules and frames. • Pure selection - uses heuristics.

The type of approach used will depend mainly on the exact design exercise being carried out. This is supported by Colton & Dascanio (1991) who state ‘Models of design contain

information that describes the various phases of the design process and estimate the sequence of and the interaction between these phases. Models can have many forms and cover any range of the design process depending on the purpose of the study.’ Both Bascaran et al. (1992) and Chung et al. (1993) have looked into the way that different design disciplines and

environments will require different approaches to developing ES and KBS and decision support tools.As a result of the work carried out in this area there are a number of ES which have been developed to support all areas of engineering design.

6.5.2.1 Examples of Expert Systems in Engineering Design

design of material structure controls uses many related factors which can not be systematised

and so is very complicated. By using ES technology Yasuda et al. have developed a

computerised system which accelerates and improve the quality of materials design. The system consists of 300 rules, a large PL/M program and 50 - 100 databases. Yasuda et al. (1992) conclude ‘As processing is made according to the expert designers thinking process, there is no difference in feeling, and it can be said to be a system that is easy to use.’ A new methodology is proposed for enhancing design management and co-operation by Guo et al. (1992). The integration architecture, principles and implementation for Integrated Intelligent Design Environment (IIDE) are presented. The example shown, of engine design process management, consists of databases and a mixture of objects and rules linking everything together. The design results can be represented graphically as fully dimensioned engineering drawings. Using the system for mechanical design They conclude that the integrated system shows great potential to solve complex real design problems.

A KBS called REKK is a design assistance tool that supplies assistance in certain phases of product development developed by Schiebeler & Ehrlenspiel (1993). It consists of several task specific modules as a hybrid system rather than a pure rule based or object oriented system. The system is demonstrated using gearbox design as an example. As well as the model costs of the gearbox, the whole costs of the gearbox and the costs of the entire gearbox

assembly can be calculated. The system also links to CAD and parts may be checked for manufacturability and costs through a direct link to the KBS. In this system three processes (relational database, CAD system and KBS) run simultaneously. They claim that the system supersedes those whose internal data-structure and programming interface cannot cope with highly complex elements. They go on to say that ‘the results of this project help to show how future design-assisting systems may be further developed.’

Other examples of ES in engineering design are: an Integrated Circuit Design Critic, Steele et al. (1992); a KBS for selecting shaft-hub connections, Klein (1992); a KBS for engineering

idealisation, Prabhakar & Sheppard (1992), an ES for the design of components made from

powder ceramic materials, Victor et al. (1993) and an ES for performing techo-economic feasibility studies in the capital equipment industry, Bate et al. (1994)

As this chapter has shown the technologies of Al, ES and KBS have been applied to a large number of different areas one of the most active of which is design. Most of the systems and tools developed have been used in areas of design which involve complex relationships and large amounts of data processing.

Earlier sections of this work have shown how Design for Environment is becoming increasing

important and that environmental concerns are very much on the agenda of every designer. The

process of EIA can be integrated into the design process quite simply. The main causes of concern are the massive amounts of data this involves and the problem of unstructured methods with which to carry out the comparison of different designs, materials or processing options. It is these problems which offer the opportunity for ES and KBS technologies to be used and as a result accelerate and structure the process of environmental design.