An Evaluation of Industrial Ventilation Troubleshooting Methods in Experimental Systems

• Published in: AIHAJ - American Industrial Hygiene Association Vol. 62; no. 6; pp. 669 - 679
• Main Authors: Guffey, Steven E., Booth, Derrick W.
• Format: Journal Article
• Language: English
• Published: United States AIHAJ 01.11.2001
• Subjects: direct pressure comparison method; Evaluation Studies as Topic; Humans; Industry; Occupational Health; Pressure; pressure ratio method; ROC Curve; ventilation; Ventilation - instrumentation
• ISSN: 1529-8663; 2163-3711
• DOI: 10.1080/15298660108984673

This study determined the efficacy of specific methods of identifying and locating obstructions and alterations to industrial exhaust ventilation systems under challenging conditions when measurement errors were minimized. Two traditional screening methods were evaluated: (1) two variations of the hood static pressure method and (2) a severely modified version of the "Check-out" method. Three proposed pressure ratio methods also were evaluated and compared with the traditional methods. Two full-sized experimental ventilation systems in two ventilation laboratories were tested. One system had five branch ducts, the other had eight, with branch duct diameters ranging from 4 to 7 inches. To create challenge, each system received multiple alterations and, in some cases, the airflow level was changed throughout the system. For each round of measurements (1) different combinations of alterations were made to some ducts; (2) on a given system, relevant pressures and flows were determined for each duct using calibrated pressure sensors and standard pitot tubes held in a traversing device; and (3) the numbers of true and false positives and negatives for each screening method were computed for a broad range of threshold values. Sensitivities were plotted against the false positive rates for all thresholds for each method. The area (A ROC ) under the resulting "receiver operating characteristic curves" was computed for each method. Variability was simulated using bootstrap methods to determine significance of differences. In addition, the thresholds that would achieve 10 and 20% false positive rates were determined for each method and the accompanying sensitivities compared. The pressure ratio methods detected nearly all nontrivial obstructions with nearly zero false positives (A ROC =1). The direct pressure comparison methods showed substantially inferior performance for the substantial challenges presented in these tests. The latter may be useful under less challenging conditions but were of dubious utility in locating obstructions under the ranges of conditions tested.