Uncertainties in Metering Stormwater Flows

Abstract
Modern stormwater management often involves the development of numerical hydrologic/hydraulic models in order to provide decision making tools regarding placement of retention facilities, in-system storage, etc. Since many existing combined sewer systems are inadequately characterized and metered, flow meters are often installed in order to provide data sets for calibrating such models. Primary metering devices such as Parshall flumes or venturi meters are often unacceptable because of their high losses and the need to avoid surcharging during high flow conditions. Low loss meters typically utilize velocity measurements to estimate discharge and the accuracy of these types of meters may be less than that associated with primary meters which is reported to be 0.5 to 1% for orifice and venturi meters (Pomroy, 1996) and 5% for Parshall flumes (Parshall, 1926). In a recent hydraulic model development for the city of Detroit, Michigan, the stated goals for the model calibration were to predict the peak discharge to within 20% and the total flow volume to within 10% of the values recorded at select metering stations (Camp, Dresser & McKee, 1993a).
In this particular application, two semi-independent methods were available for estimating the metered discharge. An analysis of the discharges predicted by these two methods indicated significant discrepancies under certain flow conditions, raising questions regarding the ability of the meters to resolve flow rates and volumes to within the stated calibration goals. The subsequent investigation of meter uncertainty is discussed in this chapter. A review of the metering systems indicates the possibilities of considerable differences between the independent discharge measurements; this is verified by a comparison for selected discharge events.
This paper is only available in PDF Format: View full text PDF