The Acquisition and Codification of Knowledge for Knowledge-Based Engineering
MetadataShow full item record
Knowledge is the backbone of any engineering challenge. Through this knowledge, man has been able to transform the elements of our planet into usable structures that serve a purpose in our modern world. By utilizing ever-increasing computing power, this knowledge can be transferred to computers to help engineers solve increasingly complex challenges. Knowledge capture and transfer is at the heart of the field of Knowledge-based Engineering or KBE. KBE is a discipline that has seen mixed success. On one hand, by automating routine engineering work, companies have been able to reduce the time of some bottlenecked design activities by up to 97% compared to pre-KBE implementation. On the other hand, KBE has seen a limited breakthrough compared to other technologies of similar age such as CAD-systems (Computer Aided Design). The backbone of any KBE-system is the knowledge on which it is built. Knowledge has to be acquired, before being coded into the system itself. This PhD-research project is focused on finding areas where KBE has seen limited success, exploring ways to acquire and codify knowledge in this area, and implementing this in a pilot KBE-system. One of the main challenges in acquiring knowledge is the massive amounts of knowledge available. Successful KBE often is applied to complex engineering challenges, and becoming an expert in even a small part of these domains often requires years of training and experience. Becoming a successful KBE-developer would therefore be impossible if you have to become an expert in every field for which you are developing an application. Due to the years of training required to become a domain expert, effective KBE solutions require input from established experts in their field, but their time is limited, and they cannot be expected to tell you everything they know. The technique is to acquire an overview of the experts’ domain, so that you can talk to them in their language. Then the sharing of knowledge will be more effective. To address this problem, this research project proposes the use of Knowledge-briefs (K-briefs), which condenses knowledge about a particular subject into a semi-structured A3 format. Although time consuming to create, K-briefs were found to be an excellent tool in communication with domain experts, even bringing out some knowledge that was hard to explicate in the first place. To test this way of acquiring knowledge, a test-case within the oil and gas industry was found, namely layout engineering. By combining K-briefs with rapid prototyping of new solutions, a pilot KBE layout application was developed. Although successful in implementing the pre-set requirements, the pilot application also identified many additional research possibilities within KBE.