Key
2470
Conference Year
2012
Keywords
Fault Detection and Diagnostics, FDD, Evaluation
Abstract
Fault detection and diagnostics (FDD) tools are being increasingly applied in air-conditioning systems. There are many different protocols used in these FDD tools, so an important question to ask is: how well do the protocols work? This paper describes the ongoing development of the first standardized method of evaluation for FDD protocols applied to air-cooled vapor compression air-conditioning systems. The general approach is to feed a library of data – including temperatures, pressures, and humidity – to a candidate FDD protocol, and observe the responses of the protocol. The library includes data from units operating with no fault or with single faults. The faults include undercharge or overcharge of refrigerant; impaired heat transfer in the condenser or evaporator; liquid line flow restriction; presence of non-condensable gas in the refrigerant; and compressor valve leakage. The outputs from the test are categorized, and statistics and figures are generated. A preliminary evaluation of RCA, a public domain FDD protocol, has been conducted using laboratory measurement data from 13 air conditioning systems. The rates of (a) no response, (b) correct, (c) false alarm, (d) missed detection, and (e) misdiagnosis show that the protocol performed poorly. This underscores the need for standardized FDD evaluation methods.