28.10.2014 Views

Fault Detection and Diagnostics for Rooftop Air Conditioners

Fault Detection and Diagnostics for Rooftop Air Conditioners

Fault Detection and Diagnostics for Rooftop Air Conditioners

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

11<br />

EXECUTIVE SUMMARY<br />

Packaged air conditioning equipment is used extensively throughout small<br />

commercial <strong>and</strong> institutional buildings. However, compared to larger systems, they tend<br />

to be not well maintained. Widespread application of automated fault detection <strong>and</strong><br />

diagnosis (FDD) to packaged equipment will significantly reduce energy use <strong>and</strong> peak<br />

electrical dem<strong>and</strong>, down time <strong>and</strong> maintenance costs. However, techniques <strong>for</strong> online<br />

FDD reported in the literature are expensive to apply because of requirements <strong>for</strong> model<br />

training <strong>and</strong> large computation, have only been tested in the laboratory, <strong>and</strong> can not<br />

h<strong>and</strong>le multiple-simultaneous faults. Furthermore, no one has per<strong>for</strong>med economic<br />

assessments of FDD. Economic assessment is a complicated problem that requires field<br />

evaluations. The primary goals of the research described in this report were to 1) develop<br />

a practical automated FDD technique having low cost, robust per<strong>for</strong>mance, <strong>and</strong> the<br />

capability to h<strong>and</strong>le multiple-simultaneous faults <strong>and</strong> 2) to per<strong>for</strong>m an initial economic<br />

assessment of FDD applied to vapor compression equipment in Cali<strong>for</strong>nia.<br />

The process of developing <strong>and</strong> evaluating the per<strong>for</strong>mance of FDD methods<br />

involved the use of both laboratory <strong>and</strong> field data. Laboratory data from previous studies<br />

were used to provide rigorous per<strong>for</strong>mance evaluations. In addition, a number of field<br />

sites were established in small commercial buildings in Cali<strong>for</strong>nia to allow consideration<br />

of practical issues. An additional field site was set up at Purdue to allow artificial<br />

implementation of faults in order to provide more controlled evaluation of the FDD<br />

techniques under realistic operating conditions.<br />

Two different FDD approaches were developed in this research. First of all, the<br />

statistical rule-based (SRB) FDD method presented by Rossi <strong>and</strong> Braun (1997) was<br />

modified to improve sensitivity <strong>and</strong> robustness <strong>and</strong> reduce computational requirements.<br />

The following components of the FDD method were improved: 1) models <strong>for</strong> predicting

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!