The objective of this research was to develop and evaluate a computer program (expert system) to assist design engineers, fabricators, and field inspectors in determining the extent of damage due to bridge steel fabrication errors and to specify the necessary repair solution. A printout with the situation entered and solution is provided for documentation. If the case situation is beyond the knowledge base, the program will give a "No solution" response instead of an incorrect solution. Testing included 33 actual cases which had not been used to develop the system and 18 hypothetical cases. The program gave correct responses in 34 cases and "No solution" in the other 17. No incorrect responses were given. An effort to augment and enhance the knowledge base of the system using Machine Learning was investigated. Automatically generated knowledge bases comparable to the manually crafted knowledge bases were created. This report discusses the purpose, advantages and shortcomings of inductive learning and the steps of this investigation.