The Nuclear Engineering and Chemical Engineering Artificial Intelligence Groups at The Ohio State University have developed a diagnostic system for the heavy water production reactors at the Savannah River Site. The diagnostic module of the system uses hybrid hierarchical decomposition methodology to decompose the search space. The knowledge is arranged so that the search space is traversed similarly to how an expert would solve the problem. The system was tested on the SRS development simulator and the results show that the system can properly diagnose all the process water and cooling water malfunctions that are programmed into the simulator. The system was not validated by operators due to hardware unavailability. Since the New Production Reactor development efforts have been halted, the probability for future work on this project is unlikely. The development used a standardized Verification and Validation program to assist in the design and construction of the system. The use of this standardized procedure is referred to as a text book example of designing an expert system in the expectation that its use would provide guidance in future projects. Of the eight phases of the software development lifecycle, five of the phases were completed and documented.