Object-oriented paradigms for computer-aided structural engineering

The increase in use of computers in many aspects of engineering constructed facilities has been enormous during the past few decades. Computers are used for analyses of structures, design checks and optimization, cost estimation, construction planning, etc. Using proven algorithms, many of the individual tasks are now highly automated. Engineering, however, involves interaction among many of these tasks and productive structural engineering systems must be an integrated software system. This dissertation deals with the fundamental interface to the resources in a computer system that engineers need to face in developing an integrated engineering system; specifically, the interfaces are a programming paradigm and a data model to represent engineering information in the database.; Object oriented concepts and paradigms have recently emerged as a promising theme in developing large systems. The objectives of the present study are to develop an object oriented software design method and an object oriented data model that are to become fundamental tools in developing large and integrated engineering systems.; The dissertation begins by introducing object oriented concepts. It then evaluates languages, for engineering software development. The commonly used procedural languages such as Fortran and C are discussed. Various object oriented languages are considered and the C++ language is proposed as an appropriate language to develop today's engineering application programs. This is followed by the design and implementation of an object oriented finite element program written in C++. Based on experience in implementing this finite element program, a general guideline for object oriented development of engineering software using C++ is given. This guideline emphasizes levels of abstraction and reusability. A programming paradigm is an essential interface to the resources in a computer; the other interface is a data model that can effectively represent engineering information in the central database of the integrated system. A simple object oriented data model appropriate for engineering information is proposed in this dissertation. The model treats classes as objects and message sending is the only mechanism necessary for communication among data objects, database users and database administrators. This makes the model simple and uniform. The dissertation concludes with a summary of the work and recommendations for future research.