Many people confuse a flume’s resolution with its accuracy and use the terms interchangeably. The two, however, are distinct from each other. A flume may have good resolution, but poor accuracy, and vice versa.
Resolution is a how much change in head is developed in the flume for a given change in the flow rate.
Accuracy is how well the determined flow in the flume reflects the actual flow rate. Secondary devices (flow meters) are rated for accuracy based upon the head measured, producing less error (as a percent of the total head measured) when measuring large heads than when measuring smaller ones.
In general, short throated flumes exhibit similar accuracies, usually +/-3-5%, while long throated flumes are usually more accurate (as are thin-plate weirs) at +/-2%.
Primary Device | Accuracy |
V-Notch Weirs | +/- 2% |
Parshall Flumes | +/- 3-5% |
HS / H / HL Flumes | +/- 3% |
Trapezoidal Flumes | +/- 2-5% |
Cutthroat Flumes | +/- 3% |
Montana Flumes | +/- 3-5% |
RBC Flumes | +/- 2% |
Under actual flow conditions, however, errors associated with approach conditions, flume installation, and measurement error tend to dampen out any differences between devices, with the common assumption that open channel installations will read within +/- 5-10% of the actual flow rate. Accuracy, therefore, tends to be less of a factor when selecting among flumes than resolution does.
Assuming a 10 gallon per minute flow in an open channel the head developed for a variety of flumes would be:
Primary Device | Head |
4-inch Palmer-Bowlus | 0.103' |
50 mm RBC | 0.103' |
1-inch Parshall | 0.17' |
Large 60° V Trapezoidal | 0.19' |
0.4-foot HS | 0.22' |
18-inch L x 1-inch W Cutthroat | 0.235' |
Now using a Hach 950 ultrasonic flow meter, for a level of the accuracy of the meter is +/-0.01-feet would give us a error of:
Primary Device | % Error |
4-inch Palmer-Bowlus | 9.71% |
50 mm RBC | 9.71% |
1-inch Parshall | 5.88% |
Large 60° V Trapezoidal | 5.26% |
0.4-foot HS | 4.54% |
18-inch L x 1-inch W Cutthroat | 4.22% |
Here resolution is a much bigger contributor to the overall system error than flume accuracy is.
in general, and where upstream conditions allow it, when selecting among flumes styles for a given set of flow conditions, the flume that generates the greattest usable head is usually the preferred choice.