BRIAN A. SMITH

THESIS TITLE: SUPERCRITICAL FLOW MEASUREMENT UTILIZING A PARSHALL FLUME

ABSTRACT: Developed for open-channel flow measurement, Parshall flumes have been widely used throughout the agricultural industry. Advantages of the Parshall flume include: high accuracy for free-flow conditions, minimal head loss, and the ability to minimize sediment deposition within the structure. However, in order to obtain high accuracy, Parshall flumes must be installed in sections of long, straight, mild-sloping channels with fully-developed, uniformly-distributed approach flow conditions free of curves, projections, and waves. Parshall flumes installed and operated under poor approach conditions provide erroneous discharge estimates when used in conjunction with standard rating equations.

Currently, the National Park Service utilizes a five-foot Parshall flume to provide discharge estimates on a high mountain, steep gradient stream located along the northern boundary of Yellowstone National Park. National Park Service hydrologists have discovered that the rated discharge for given gauge heights within the flume were inaccurate up to 30% when compared to stream-gauge discharge measurements taken over a range of flows. It was hypothesized that flow entering the flume could be in a supercritical state. This resulted in the development of a test program to examine the feasibility of utilizing a Parshall flume to measure discharge under supercritical approach flow conditions. Flume experiments were conducted on a six-inch and nine-inch Parshall flume at the Hydraulics Laboratory of Colorado State University. Six discharges were conveyed over 3 different upstream bed slopes with 3 unique upstream roughness elements to comprise a test matrix of 15 configurations and 82 independent tests.

Analysis of the collected data indicated that the calibration of a Parshall flume to accurately estimate discharge under supercritical approach flow conditions was feasible. Results of a dimensional analysis provided a list of dominant variables used to develop a calibration equation that accounted for variations in upstream bed slope and roughness. Additionally, it was observed that supercritical approach flow conditions resulted in two separate conditions for flow entering the Parshall flume. These two conditions were supercritical entrance flow and subcritical entrance flow. It was determined that subcritical entrance flow conditions were caused by the Parshall flume choking the flow.

Nonlinear regression analysis techniques were utilized to develop predictive equations for both entrance flow conditions. The predictive equation for supercritical entrance flow conditions was accurate to ±5% and the predictive equation for subcritical entrance flow conditions was accurate to ±10%. Due to the possibility of entrance conditions fluctuating between flow regimes, a flow regime predictor was developed so that users could easily determine which entrance equation to utilize in the field.

GRADUATED: Summer 2003

RESUME (pdf)