Just as standard methods have been developed for evaluating different building energy simulation programs, there is growing realization of the imperative need to develop methods to evaluate the performance of different fault detection and diagnosis (FDD) methods and tools for different heating, ventilating, air-conditioning, and refrigerating equipment. This paper describes research done in this regard by first reviewing past pertinent literature on FDD evaluation and identifying various elements of a general conceptual evaluation framework. In order to make the tool practical for an impartial evaluation, a small set of quantitative criteria are identified based on which an objective function formulation is described where one minimizes the sum of the costs associated with false positives and those of missed opportunities. Subsequently, we describe how this function can be modified and simplified so as to obtain expressions for normalized rating of any FDD tool in terms of fault detection only as well as in terms of the combined FDD process. These general expressions are then tailored to the evaluation of FDD methods and tools specific to large chillers, and specific numerical values of several of the quantities appearing in the normalized rating expression are proposed based on discussions with chiller manufacturers and service companies as well as analysis of fault-free and faulty chiller performance data gathered in a laboratory from a previous research study. Another paper will describe its application and usefulness to assess four chiller FDD methods customized to the same performance data sets.

Units: Dual