Comparison of efficacies of current methods for troubleshooting industrial exhaust ventilation systems to a proposed new method.
Seattle, WA: University of Washington, 1996 Feb; :1-84
Industrial exhaust ventilation systems are designed to minimize worker exposure to airborne contam inants. For ventilation systems to operate effectively and to protect workers from harmful exposures, air drawn into the fan must be distributed among the hoods at predetermined (target) airflows. Over time systems age, incurring damage, obstructions, leaks and other alterations that skew the airflow distributions so that some hoods receive less than their target airflows. The ability to swiftly detect alterations ("troubieshooting") and restore systems to their previous working condition is therefore crucial for worker protection. Two methods used to identify alterations that produce shifts in airflow distribution are long-standing, the Industrial Ventilation method (IVM) and the hood static pressure method. In use, it is assumed in both methods that a decrease in hood static pressure (SPH) indicates a new obstruction, but an increase in SPH does not. However, one could broaden the method and assume that an increase also can indicate a new obstruction, thus creating what is called here "One-Sided" and "TwoSide" Hood Static Pressure methods. A third method is proposed by Guffey. It incorporates a variant of the hood static method and introduces two more diagnostic tests: ratios of static pressures (variable called "SPratio") and equivalent loss coefficients ("X-values") which are kinetic power loss coefficients for any volume. Previous laboratory studies have demonstrated the superiority of X-values in locating obstructions that have been deliberately placed in systems. This is important because shifts in airflow are generally due to obstructions. However, to determine values of X requires a timeconsuming velocity pressure traverse. Therefore, it would be convenient to use static pressures as screening tools to reduce the number of cases where X-values must be determined. In this study of an eleven branch ventilation system, static pressures were measured downstream of the hood (SPH), at the end of the branch (SPend), and at a location between the hood and end of the branch (SPmid). Velocity pressures were measured by a Pitot traverse at a convenient location in each branch. Cases where the change in X exceeded a specified threshold were deemed "obstructed." A screening test was deemed "positive" if the change in its variable value(s) exceeded a given threshold. A "true positive" for a method occurred when the value of X for a branch changed by more than a given X-threshold and the value of the method variable changed by more than its threshold. Thresholds for X were tested at values ranging from 0.05 to 0.6 and thresholds for each screening test's variable were varied from 0.0 to 0.6. A family of receiver operating characteristic (ROC) curves were drawn at each threshold for X. Performance for each screening tool was judged in part by area under the ROC curve. The results of this study were that in every case the areas under the ROC curves (indicating superior combinations of sensitivity and specificity) were higher for the SPratio Method and the "Two-Sided SPH" method and were very low for the IVM method and the "One-Sided SPH" method. One reason the IVM method and OneSided SPH method performed poorly was because they ignored obstructions upstream of the SPH measurement location. There was little difference between Two-Sided SPH method and SPratio with high values of X. At moderate changes in X, (e.g., <30%) the SPratio method was clearly superior to all the other methods. Therefore, SPratio method is the best screening tool at finding moderate obstructions. For very substantial obstructions, the Two-Sided SPH method would be adequate.
Ventilation; Industrial-ventilation; Ventilation-systems; Exhaust-ventilation; Air-flow; Air-contamination; Airborne-particles; Equipment-design; Models
Ann Pinsky, University of Washington, School of Public Health and Community Medicine, Department of Environmental Health, Room F226D, Box 357234, Seattle, WA 98195-7234
Empirical determination of the error in the ACGIH method of predicting airflow distribution in two ventilation systems
University of Washington, Department of Environmental Health, Seattle, WA