A Hierarchical Heuristic Engineering Design Method for Nuclear Systems Based On the Progressive Definition of Constraints

Abstract

A novel hierarchical heuristic engineering design method is developed that is based on levels of abstraction. This was created in response to the dearth of engineering design methods geared for nuclear engineering and the inability of existing engineering design methodologies to address the complexity of the nuclear system. Constraint selection is regarded as a more critical decision in nuclear system design than optimization of the design problem. The nuclear system is defined as a nested hierarchy of systems that is decomposed from the most abstract to the most concrete. Objectives and constraints for each level are created that bring further specificity to the objectives and constraints of the preceding (more abstract) level. The design or solution of the more abstract level is the basis for the more concrete level of abstraction. A sequence of levels of abstraction is presented that should be sufficient for the majority of nuclear systems but this sequence is not mandated. The engineer is encouraged to adapt his levels of abstraction to his problem’s design space. A comprehensive set of heuristics is provided for the method. The developed hierarchical design method is agnostic to the optimization method within the level of abstraction although heuristics are provided to select the one best suited to the problem. The method is illustrated in the design of a materials testing fast spectrum reactor. It is used to optimize both global reactor systems and smaller problems dealing with a specific aspect of the system. The method was compared against Axiomatic Design using the materials testing reactor and nuclear system example problems. The method possessed greater flexibility than Axiomatic Design. Traditional engineering design methodologies and their variants do not provide enough emphasis on constraint development and were not examined as in-depth.