Synthesis of heuristic knowledgebase for supporting development of goal-oriented reactor noise analysis programs

Abstract A knowledgebase system which assists a noise analyst to synthesize customized, i.e. goal-oriented, computer programs is proposed in this paper. The knowledgebase stores empirical and heuristic rules for choosing a set of subroutines, selecting control parameters, evaluating validity of the analysis, and for interpretting the outcome of the computation. Various statistical test procedures were utilized extensively in order to extract information to be used as input data for the rules to make decisions. The knowledgebase system is implemented on a personal computer using PROLOG language, while the numerical analysis program is synthesized on a mainframe computer. The subroutines are combined, through consultation with the knowledgebase system, to form a specific-purpose program for given signal set and signal conditions. This approach allow us to improve efficiency of software development for noise analysis, plant monitoring and diagnosis.