AT-96-06-3 -- Application of Black-Box Models to HVAC Systems for Fault Detection PDF

Describes the application of black-box models for fault detection and diagnosis in HVAC systems. Uses multiple-input/single-output ARX models and artificial neural network (ANN) models. Finds that the use of system models makes it possible to detect faulty behaviour, whereafter component models can be used to locate the defective component. For nonlinear systems, ANN models fit better than ARX models. Notes the need for further development of black-box modelling techniques to develop reliable models that can be used in practical applications.

KEYWORDS: year 1996, Failure, monitoring, calculating