Proceedings of 1995 American Control Conference - ACC'95
DOI: 10.1109/acc.1995.532753
|View full text |Cite
|
Sign up to set email alerts
|

Using thermodynamic impact for detecting refrigerant leaks in vapor compression equipment

Abstract: A technique for detecting refrigerant leaks by utilizing their impact on the thermodynamic states of the vapor compression cycle is described. Simulation and laboratory experiments were performed to determine which of 7 inexpensive measurements contribute significantly to detection confidence. Experimental results show that suction line superheat and liquid line subcooling are the minimum measurements needed to detect and isolate refrigerant leaks from the other faults considered and provides 99.9% detection c… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2

Citation Types

0
2
0

Publication Types

Select...
1
1

Relationship

0
2

Authors

Journals

citations
Cited by 2 publications
(2 citation statements)
references
References 0 publications
0
2
0
Order By: Relevance
“…Rossi and Braun [5] presented a thermodynamic model-based FDD system for a commercial air-toair rooftop air conditioner. An optimal parametric linear classifier was used for fault detection.…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation
“…Rossi and Braun [5] presented a thermodynamic model-based FDD system for a commercial air-toair rooftop air conditioner. An optimal parametric linear classifier was used for fault detection.…”
Section: Introductionmentioning
confidence: 99%
“…Castro [9] also presented a model-based approach to FDD for a reciprocating chiller and investigated its performance using simulation and experimental test results. The method relies on the steady state system model presented by Rossi and Braun [5] to predict fault-free operating parameters and a two-step FDD approach. In the first step, residuals between model predictions and measured data are classified as normal or faulty using two different classification techniques: a k-nearest-neighbour classifier and a k-nearest-prototype classifier.…”
Section: Introductionmentioning
confidence: 99%