Engineering tasks have become more and more dependent on supporting software. The amount and complexity of engineering data that CAE systems have to deal with, increases continuously and the management of it has become a tedious task. Traditional database technology is ofthen not suited to handle the complex data types needed for engineering. This has lead to the emerging of a new class of databases called engineering databases (EDBs). They combine techniques developed in research areas, such as traditional database research and object-oriented programming, to support the management of engineering data. This report gives and introduction to EDBs by describing the features that a typicat EDB-system should contain, and presents some existing systems that demonstrate these features. The material has been collected mainly from the latest conferences related to EDBs.

This report shows some promising results in using neural networks for handwritten character recognition in real-time. The characters were input from a digitizer and preprocessed to extract different features. These were taken as input to a backpropagation network which was trained to recognize the 29 lower case letters in the Swedish alphabet. Its configuration was 40 input nodes, 30 hidden nodes and 29 output nodes. It was trained with 4 different versions of each character in 15 000 iterations. After that it was able to recognize 100% of two other separately entered versions of the alphabet. The amount of computation needed for classifying a character with this kind of networks puts no restrictions on real-time performance.

Denna rapport utgör slutrappporteringen till NUTEK för IT4-projekt 2405, Beslutstödsystem för konstruktörer. Projektet har genomförts som ett samarbetsprojekt mellan ABB STAL AB, ABB Carbon AB, Institutionen för Datavetenskap vid Linköpings Universitet och Sveriges Verkstadsindustrier. Det startade hösten 1991 och avslutades kring årsskiftet 1993/94. Ett uppföljningsprojekt kallat Systemutvecklingsteknik för komplexa hybrida system - Konstruktörstödsystem har i skrivandets stund påbörjats. Syftet med detta är att generalisera erfarenheterna. IT4-projektet har resulterat i ett konstruktionsstödsystem för turbin och kraftverktsanläggningar. Det används sedan hösten 93 i skarp drift vid ABB Carbon AB för systemkonstruktion av kraftversanläggningar med trycksatt kolförbränningsteknik. Vid ABB STAL utnyttjas det för konstruktion och utveckling av en ny gasturbin. Under 94 ska det tas i storskalig drift vid ABB STAL, för konstruktion av mera standardiserade ång- och gasturbinanläggningar med en omsättning på mer än tjugo anläggningar om året. Rapporten beskriver hur projektet genomförts, och den objektorienterade systemutvecklingsteknik som utvecklats. Denna uppskattas har minskat utvecklings och underhållskostnaderna mellan fem och tio gånger. I Kapitel 6 redovisas erfarenheter som kan vara till hjälp vid andra konstruktionsorienterade verkstadsindustrier. Inom projektet har ett antal föredrag och kurser tagits fram, som kan ges på förfrågan till CAELAB/IDA vid Linköpings Universitet.

Second generation knowledge-based systems have raised the focus of research from rulebased to model-based systems. Model-based systems allow knowledge to be separated into target domain model knowledge and problem solving knowledge.

This work supports and builds on the hypothesis that fully object-oriented knowledge-bases provide the best properties for managing large amounts of target domain model know ledge. The ease by which object-oriented representations can be mapped to efficient graphical user interfaces is also beneficial for building interactive graphical knowledge acquisition and maintenance tools. These allow experts to incrementally enter and maintain larger quantities of knowledge in knowledge-bases without support from a knowledge engineer.

The thesis points to recent advances in the conception of knowledge-based systems. It shows the need for efficient user interfaces for management of large amounts of heterogeneous knowledge components. It describes a user interface software architecture for implementing interactive graphical knowledge-base browsers and editors for such large knowledge-bases. The architecture has been inspired by object-oriented programming and data-bases, infological theory, cognitive psychology and practical implementation work.

The goal with the user interface software architecture has been to facilitate the implementation of flexible interactive environments that support creative work. Theoretical models of the entire user interaction situation including the knowledge-base, the user interface and the user are described. The models indicate how theoretical comparisons of different user interface designs can be made by using certain suggested measures.

The architecture was developed in the frame of a cooperative project with the Department of Mechanical Engineering on developing a knowledge-based intelligent front end for a computer aided engineering system for damage tolerance design on aircraft structures.

ProCAD is a product modeling system for power plant system design. It is used for functional design of steam turbine, gas turbine, combined cycle, and pressurised fluid bed coal combustion (PFBC) power plants. The product models can contain descriptions of thousands of functional components: turbines, electrical generators, pipes, pumps, instruments, valves, heat exchangers etc. The power plant system designer can interact with the detailed information in the product model database directly from the drawing environment in the CAD-system. Reports an product documentation is generated from the product model database using 4GL applications with an SQL-interface. The paper describes the system design of ProCAD. An overview of the structure of the product model database is given, and the CAD-client application for drawing process and instrumentation diagrams. The paper provides examples of important generic functionality for multi user product modeling systems, such as configurable project specific attributes in the product model database, assignment of update access rights on distinct substructures in the product model to different engineers, and functions for checking the consistency between file based CAD-drawings and information in the product model database.

Many companies have mature tool support for single engineering disciplines such as system process design, mechanical- and electrical engineering. There is a large potential for reducing development times and increase information quality by integrating these tools. Such integration can benefit from a formalized format for documenting the information models of the different systems, and how information in one model can be transformed to the other. The paper presents an information model for a meta-database for documentation of information systems that have complex information models with overlapping substructures. It describes a prototype application, based on an object-oriented CASE-tool, for editing information model mappings, and reports on early experiences from using the prototype for documenting information systems used in production at the turbine and power plant manufacturer ABB STAL. The results seen so far indicate that this approach is promising for managing the complexity and evolution of bridges and integration mechanisms of engineering tools with large and complex information models.

The complexity of industrial artifacts such as power plants, aircrafts etc., is continuously growing. Typically for such advanced products is that they are hybrids of various technologies and contain several types of engineering models that are related in a complex fashion. For power plant design, there are functional models, mechanical models, electrical models etc. A product modeling system (PMS) is a computer integrated development environment for a specific class of advanced products. A well integrated PMS consists of a central product model database which is interfaced with CAD-applications that support graphical design of various engineering models. This paper describes a successful approach to manage the development of a product modeling system for power plant design. The key idea is to store a high-level PMS design specification in the form of an extended entity relationship model in a data dictionary. Most of the source code for the PMS implementation is then generated automatically, using SQL-based code generators which are easy to develop. Our PMS-development system generates product model database schemas and user interfaces. It also generates high-level database schema related interface modules in the native application development language of a CAD-system. Through these, a CAD application developer has a high-level access to the object structures in the product model database. Using the described approach, we have developed a power plant PMS which has been in production at the turbine manufacturer ABB TAL and the power plant engineering company ABB Carbon for more than half a year now. The PMS database is based on the Sybase SQL Server. AutoCAD is used for design of power plant schematics. Form based user interfaces and report generators are implemented in UNIFACE and Microsoft Access. The data dictionary design and SQL-based code generation technique seems to be generally applicable and has been used for generating source code implementations in C++, LISP, SQL, and various textual form description languages. The architecture of our PMS-development system is described together with the data dictionary schema and examples of generated source code. We estimate that this software engineering approach reduces systems development costs about 5 - 10 times.

Komplexiteten i högteknologiska produkter som kraftverk, flygplan, etc. ökar ständigt. Produktutvecklingstiderna måste minskas samtidigt som design och kvalitet ska förbättras på ett kostnadseffektivt sätt. Ett produktmodelleringssystem (PMS) är en datorintegrerad konstruktionsmiljö för en specifik klass av avancerade produkter. Det finns en stor potential att förbättra produktutvecklingsprocessen med hjälp av modern informationsteknologi. Det krävs dock en långsiktig och genomtänkt satsning för att utveckla ett PMS som är flexibelt nog att snabbt kunna anpassas till nya produktkrav. Systemet måste även kunna porteras till nya generationer av datateknik under det antal decennier som utgör produktens livscykel. Artikeln beskriver ett antal aspekter på mjuvaruutvecklingsmiljön man bör ta i beaktande för att lyckas med en långsiktig satsning på ett PMS. Utvecklingsmiljöns arkitektur och en iterativ utvecklingsmetod beskrivs som borgar för att produktmodelleringssystemet tidigt kan tas i drift och utvecklas mot en funktionalitet som prioriteras av företagsledning och användare. Utvecklingstrender och standarder för kommersiell basmjukvara som är lämplig att använda i ett PMS beskrivs. Artikeln baseras på erfarenheter från en framgångsrik utveckling av ett PMS kallat ProcessCAD. Det används för systemkonstruktion av turbin- och kraftverksanläggningar vid ABB STAL och ABB Carbon. Utvecklingsprojektet startade hösten 91. Efter 10 prototyper med stegvis ökande funktionalitet togs systemet i drift hösten 93. Därefter har fler än tio kraftverksanläggningar konstruerats varav två nu monteras ute hos kunderna. Leveranstiden för de vanligaste anläggningarna är ca 1 år.