Knowledge based mechanical design with ontological based approach in hybrid knowledge representation

Adam Suchodolski

Abstract

This dissertation is concerned with building Knowledge Based Engineering Applications aimed at Small and Middle Enterprises. It constitutes an overview of current studies, approaches, projects and attempts of frameworks for developing tools for engineering design. In the doctoral thesis an analysis of shortcomings current approaches is presented. Apart from that an innovative point of view at the building of such tools is demonstrated and a general framework for dealing with knowledge is proposed and discussed. Development of a tool to support stamping die design for an automotive supplier as a case study is presented. The main concept assumes an application which consists of two kernels: knowledge base and retrieval system. The retrieval system is a control module which retrieves required information from knowledge base. Knowledge base is a knowledge repository. In proposed solution knowledge is represented with means of two different approaches: ontological knowledge base, which is dedicated for declarative knowledge and procedural knowledge, which is represented in advanced parametrically- driven associative geometrical models. Due to heterogeneous character of knowledge representation the approach is baptized: hybrid knowledge representation. The hybrid knowledge base is readable for both a computer agent and expert. The solution bridges the gap between human oriented knowledge and formal knowledge models and is aimed at solving synthetic and deductive tasks, especially configuration design. Due to applying ontology for knowledge representation as well as common recognized protocol for accessing knowledge base the application meets the requirements of reusability, flexibility and interoperability; human oriented knowledge representation allows partial extending the application by authorized expert without reprogramming. For building the knowledge base CATIA V5 as CAD System and Protégé as ontology editor and knowledge based framework with OKBC (Open Knowledge Base Connectivity) protocol were used.
Diploma typeDoctor of Philosophy
Author Adam Suchodolski (FACME)
Adam Suchodolski,,
- Faculty of Automotive and Construction Machinery Engineering
Title in EnglishKnowledge based mechanical design with ontological based approach in hybrid knowledge representation
Languagede niemiecki
Certifying UnitFaculty of Automotive and Construction Machinery Engineering (FACME)
Disciplinemechanical engineering / (technology domain) / (technological sciences)
Defense Date14-06-2010
Supervisor Margot Ruschitzka (FACME)
Margot Ruschitzka,,
- Faculty of Automotive and Construction Machinery Engineering

Internal reviewers Jerzy Pokojski (FACME / IMDF)
Jerzy Pokojski,,
- Institute of Machine Design Fundamentals
External reviewers Konrad Okulicz
Konrad Okulicz,,
-
Pages171
Keywords in EnglishKnowledge Based Engineering, Knowledge Based System, Hybrid Knowledge Representation, Deductive Reasoning, Computer Aided Design, Ontology, Ontological Knowledge Base, Declarative Knowledge and Procedural Knowledge
Abstract in EnglishThis dissertation is concerned with building Knowledge Based Engineering Applications aimed at Small and Middle Enterprises. It constitutes an overview of current studies, approaches, projects and attempts of frameworks for developing tools for engineering design. In the doctoral thesis an analysis of shortcomings current approaches is presented. Apart from that an innovative point of view at the building of such tools is demonstrated and a general framework for dealing with knowledge is proposed and discussed. Development of a tool to support stamping die design for an automotive supplier as a case study is presented. The main concept assumes an application which consists of two kernels: knowledge base and retrieval system. The retrieval system is a control module which retrieves required information from knowledge base. Knowledge base is a knowledge repository. In proposed solution knowledge is represented with means of two different approaches: ontological knowledge base, which is dedicated for declarative knowledge and procedural knowledge, which is represented in advanced parametrically- driven associative geometrical models. Due to heterogeneous character of knowledge representation the approach is baptized: hybrid knowledge representation. The hybrid knowledge base is readable for both a computer agent and expert. The solution bridges the gap between human oriented knowledge and formal knowledge models and is aimed at solving synthetic and deductive tasks, especially configuration design. Due to applying ontology for knowledge representation as well as common recognized protocol for accessing knowledge base the application meets the requirements of reusability, flexibility and interoperability; human oriented knowledge representation allows partial extending the application by authorized expert without reprogramming. For building the knowledge base CATIA V5 as CAD System and Protégé as ontology editor and knowledge based framework with OKBC (Open Knowledge Base Connectivity) protocol were used.
Thesis file
Suchodolski.pdf 3.06 MB
Citation count*5 (2020-09-18)

Get link to the record

Back
Confirmation
Are you sure?