Anniina Kittilä, ETH Zürich, Institute of Geophysics, Geothermische Energie u. Geofluide. Sonneggstrasse 5, 8092 Zürich, Switzerland e-mail: anniina.kittila(at)erdw.ethz.ch
The use of weirs and flumes is based on structures with standardized shapes and characteristics. These different structures have specific equations, charts and tables to determine the flow rate applying basic hydraulic mechanisms by forming critical flow conditions. The relationship between depth of water flow in the structure and flow rate is stable and unique for the shape of the structure. These structures can be permanently built on site to have a possibility for continuous monitoring, or portable structures that can be separately installed for individual measurements (Mustonen 1986, Shaw et al. 2011). In general, weirs are structures that change the upstream level and flumes are channel-type structures, but the distinction is not always definitive. However, there are multiple standards, such as ISO 8368:1999, about the water flow measurements in open channels and to determine the right type of the structure. Factors to be considered in choosing right structure include size of the channel and range of flow it is expected to measure. Another important factor is the sediment load.
Weirs are dam-like overflow structures that can be characterized based on the edge or surface over which the water flows, called crest. The most used types are V-notch and rectangular shaped crests (Mustonen 1986, Rickly Hydrological Company 2015). The focus of this article is in V-notched weirs.
Flumes are open channels where the upstream sub-critical flow is constricted by narrowing the channel which causes an increase in velocity and a decrease in the depth of the flow. With sufficiently narrow channel width, i.e. sufficient constriction, the flow becomes critical in the narrowest part of the flume and results in a standing wave further downstream. The flow rate can then be deduced from the water level upstream the flume (Shaw et al. 2011). There are numerous different types of flumes, but in this article only H-flumes, which are a hybrid between flumes and weirs, and Parshall flumes are handled, because they are probably the most commonly used flume types.
Description of the method
V-notch weirs have triangular shaped cross-section, where the angle of the V-notch varies. The most common values of the angle are 90° and 45°, but smaller angles are used for very small flows, and bigger, e.g. 120° angle weirs for higher flow rates (Shaw et al. 2011). Once the weir has been installed or constructed properly according to the standards, the flow rate can be calculated from the measured water level with specific equations. The water level should be measured at a distance 3-4 times the maximum anticipated head on the upstream side of the weir. The water level, or the depth of the water, is measured from the elevation of the lowest point of the crest (the V-notch) to the surface of the water. So the depth of the water is not measured from the bottom of the channel, only the height of the water column passing over the crest matters. Measurements taken too near the weir are influenced by the drawdown of the water flow through the weir, and the flow rate is underestimated (Openchannelflow 2015b).
The design of the H-flume uses features from both weirs and flumes. It has a triangular crest, but in inclined position towards the upstream. Additionally, instead of being a dam-like structure under the crest opening, such as in weirs, H-flumes have flat bottom and they are unobstructed throughout the whole depth of the structure. This allows sediment and debris pass through the opening more freely than in weirs. There are three classes of H-flumes that vary in size and maximum capacity. The geometries of different classes vary, and the dimensions of all of them are based on the flume depth (D). The three classes are:
- HS-flume for low flows with maximum capacity of approximately 22.7 l/s
- H-flume for medium flows, maximum capacity about 2,379 l/s
- HL-flume for high flows, maximum capacity about 3,285 l/s.
The width and the length increase so that the HS-flumes are the narrowest and shortest, and HL-flumes are the widest and the longest. As the geometry changes with different flumes, so does the location of the water level measuring point. In HS-flumes the measuring point is in the distance equal to the flume depth (D) from the end of the flume upstream, in H-flumes it is in the distance of 1.05D from the end of the flume, and in HL-flumes it is in the distance of 1.25D from the end of the flume. The H-flumes in general are suitable in a wide range of applications, such as watershed and drainage studies, or landfill leachate monitoring (Openchannelflow 2015a).
Parshall flume is open along the whole depth of the structure and consists of three separate sections. The first section upstream constricts the flow to a short and narrow throat section, from where the flow continues to a uniformly diverging downstream section. The upstream section is the longest one with flat floor, the narrow and short throat section has downward sloping floor, and in the downstream section the floor rises again so that the elevation in the end of the flume is slightly lower than in the upstream section (Mustonen 1986, Openchannelflow 2015a). This structure of narrowing the flow with elevation drop makes the flow critical in the throat section and forms a standing wave further downstream. The water level upstream where the flow is sub-critical can then be measured and related to the flow rate (Shaw et al. 2011). The point of water level measurement is 2/3A from the end of the converging section towards upstream, where A is the length of the side wall of converging section. There are 22 different sizes of Parshall flumes, and they are not dimensionless, like H-flumes for example. So although their basic shape is the same, the flow rate measurements are not comparable between different models. The different sizes of the Parshall flumes cover flows from approximately 0.14 l/s to 92,890 l/s, and have proven to be suitable, for example, to dam seepage, mine drainage and tailings water runoff monitoring (Openchannelflow 2015a).
In general, weirs and flumes can be used to measure flow rates with good accuracy, but costs, size of the channel and head losses might restrict their use. The construction of the measuring structures according to the standards might be challenging, especially for flumes, and finding a suitable location for a weir in very gently sloping channels could be difficult. Nevertheless, the structures can be both portable and adjustable, reducing the maintenance compared to permanent structures (Mustonen 1986). The suitability and need for shorter maintenance intervals should, however, be considered particularly in heavily contaminated sites. For example, precipitation of gypsum or iron hydroxide could change the diameter and affect the measurement accuracy over time (Wolkersdorfer 2008).
Restricting fish migration, possibility of clogging of the structure due to high sediment load, and upper and lower flow rate limits are the main disadvantages of weirs (Shaw et al. 2011). Indeed, sedimentation over time and submergence are considerable disadvantages (Wolkersdorfer 2008). If the sediment load in the water is high, a flume may be a more appropriate choice. However, weirs tend to be more accurate in measuring flow rate than flumes, especially the triangular shaped V-notch weir, and for small flows the use of V-notch weirs is relatively inexpensive method (Openchannelflow 2015b).
The H-flumes combine the sensitivity of certain weirs with the self-cleaning properties of flumes (Openchannelflow 2015a). The self-cleaning capability can even keep the H-flume clean from iron precipitates (Wolkersdorfer 2008). Also comparing with other flumes, the H-flumes are cheaper to construct, particularly because of the flat floor, eliminating the need to install it above the channel floor. However, as with other types of flumes also, scouring of earth channels might occur near the entrance and possibly at the exit of the flume too. Scouring could cause flow to bypass the flume or set the flume off its place, so it is good to prevent it by reinforcing the channel (Openchannelflow 2015a).
Parshall flumes have a wide operating range and they are self-cleaning. In addition, the narrowing approach section of the flume followed by a downward sloping floor of the throat, the narrowest section of the flume, keeps the Parshall flume accurate even in partial submergence (Rickly Hydrological Company 2015). In comparison to weirs, Parshall flumes (and flumes in general) are particularly suitable for streams with considerably high sediment load (Shaw et al. 2011), but they need to be constructed carefully in order to meet the standard measurements. Due to the constructive restrictions, for example, Wolkersdorfer (2008) does not recommend using them in mine water applications. However, once the flume has been properly installed and flow conditions have not altered, it does not require additional calibration (Openchannelflow 2015a).
Weirs are generally used for measuring low and medium flows, and all the structures have upper performance levels for their ability to measure flow rate. If the flow rate increases above the upper level, the structure is drowned out and the unique relationship between the water level and flow rate is lost (Shaw et al. 2011). The limit for appropriate use can be calculated with the relation between the velocity of the discharged water and the depth of the channel. This relation, called the Froude number, should not exceed 0.5 (Wolkersdorfer 2008, Openchannelflow 2015b). Regular maintenance and inspections to make sure the structure is still in standard condition are required. Things to check or verify in the upstream side of the weir are: the approach channel is straight and uniform, approaching flow is tranquil without turbulence, there is no debris or abundant vegetation growth, and the flow velocity is not higher than 0.15 m/s. The inspections and the maintenance relating to the structure itself are: the weir should be centred in the channel, no significant sedimentation has taken place possibly altering the original standard measurements, the weir is in level and in its intended position, and the clean flow of water over the crest is not impeded (by debris or rust, for example). The downstream part should be maintained to avoid scouring and accumulation of debris, both possibly affecting the proper flow of the exiting water. In addition, it is good to make sure that the location of the measuring point has not changed (Openchannelflow 2015b).
Although H-flumes converge towards the bottom so that even at a low flow higher velocity is maintained in order to avoid sedimentation (Marr et al. 2010, Openchannelflow 2015a), the structure is not fully self-cleaning and regular maintenance to avoid excess sedimentation is recommended. The maintenance needs for H-flumes are very similar than those with V-notch weirs and weirs in general, i.e. to make sure initial position and standard dimensions have remained the same, upstream and downstream flow conditions are proper, and the location of the measuring point is correct (Openchannelflow 2015a).
The accuracy of Parshall flumes is 2% and better in laboratory conditions, but lack of maintenance (as with any flume structure) leading to vegetation growth, accumulation of sediment and debris, and deterioration of the flume structure from corrosion, for example, may decrease the accuracy to 5%. In the maintenance of Parshall flumes to the same specifications should be paid attention as in the case of H-flumes. (Openchannelflow 2015a).
Design requirements and installation
Because weirs and flumes are standardized structures that use specific equations to calculate the flow rate from just one water level measurement, their installation should be done with care.
Weirs (V-notched weir)
When installing or constructing a weir the flow in the upstream channel should be uniform, tranquil and steady, also the velocity profile should be well-distributed. Additionally the approach velocity of the flow should not be more than 0.15 m/s. This can be achieved when the approach channel is straight and of the length at least 15 to 20 times the maximum anticipated water head (Hmax). In order to function accurately the weir should be located so that the lowest point of the crest is at least two to three (2-3) times Hmax above the channel floor and the sides of the channel are in the distance of at least two (2) times Hmax from the sides of the crest. It is also preferable to construct the weir so that it can operate under free-flow conditions, i.e. the overflowing water discharges into the air as the downstream water level is a minimum 5 cm lower than the lowest point of the crest. The weir should be positioned perpendicular to the direction of the flow, so that it is level from side-to-side and the upstream face is vertical. There are many materials to choose for a weir, but particularly in permanent structures corrosion and wearing should be taken into account. The channel floor is good to be sufficiently tamped to make sure the weir remains in the intended position and flow does not bypass the structure. (Openchannelflow 2015b)
Probably the most important specification in installing a weir is to size the upstream weir pool properly, because it affects the approaching flow conditions and thus also the accuracy of the flow rate measurements. If it is too narrow, approaching flow velocity is too high and error in flow rate measurements can be significant. (Openchannelflow 2015b)
Considerable research has been done on the calibration of different weirs to get the most suitable coefficients and equations to calculate the flow rate accurately (Shaw et al. 2011). Instructions for these calculations and equations can be found in standard texts on hydrology and hydraulics (Mustonen 1986, Openchannelflow 2015b).
As with weirs too, the three main areas of concern in installing H-flumes are upstream conditions, flume positioning, and downstream channel conditions (Global water 2015). Although the channel upstream does not need to be straight from as long distance as for weirs, the channel should be uniform immediately upstream from the flume (Openchannelflow 2015a), preferably for the length five (5) times the height of the flume (Global Water 2015). The approaching velocity distribution should also be tranquil and uniform and the flume itself should be positioned to the centre of the channel (Openchannelflow 2015a) so that the channel is equally wide as the flume (Global Water 2015). H-flumes should be installed on a flat and level floor so that the floor of the flume is level in both longitudinal and transverse directions, and the narrow V-shaped section pointing downstream. The flow should also be able to exit the flume freely, so that the structure does not operate submerged (Marr et al. 2010, Global Water 2015, Openchannelflow 2015a). Making the downstream part have a large fall or sloping steeply can prevent the possibility of submerged flow conditions (Global Water 2015). In installation it is also good to make sure that dimensions of the flume do not change and the structure does not distort, or that the flume does not float out of its intended position, particularly when the flume is installed with concrete (Global Water 2015, Openchannelflow 2015a).
Proper and successful installation of Parshall flumes requires many of the same specifications as the H-flumes. The upstream flow should be tranquil and uniformly distributed, which can be achieved by straight upstream channel that is approximately 10 to 20 times the length of the Parshall flumes throat width. The width of the throat section should be about 1/3 or half (1/2) of the width of the upstream water surface, and the flume should be in level in the center of the channel. The downstream channel should be straight for five (5) to 20 times the width of the throat section to ensure free exit of the water. (Openchannelflow 2015a)
In order to make sure that free-flow condition exist in the exit of the flume, it usually needs to be installed above the channel floor (based on specific calculations), which might result in additional sedimentation upstream. Submerged flume might solve this problem, but in these conditions the accuracy of the structure is slightly compromised. Free flow occurs when the ratio of the lower water level Hb (located in the throat of the Parshall flume) to the higher water level Ha (the measurement point for flow rate), Hb/Ha ratio, is less than a certain ratio determined separately for different flume sizes. The bigger the flume, the higher submergence transition, and at this ratio and above it, the certain flume operates in submerged conditions (Openchannelflow 2015a). In submerged conditions reduction adjustment to the measured water level rating is needed, which makes the flume less accurate (Rickly Hydrological Company 2015). The submergence ratio of approximately 0.9 to 0.95 is in the upper limit of correctable flow. Same equations for calculating the flow rate cannot be used for both free-flowing and submerged Parshall flumes, so during the maintenance of flumes initially installed to free-flow conditions it should be checked that downstream sedimentation or vegetation growth have not increased the flow resistance too much resulting in submerged conditions. As mentioned above, submerged conditions require the installation of two measuring points, Ha and Hb. Because Hb is located in the throat section, where the flow is turbulent causing water level fluctuations, it might be necessary to build a stilling well for measuring the water level in Hb. This, however, could cause additional clogging problems in addition to higher construction costs (Openchannelflow 2015a).
Global Wate. 2015. H flume installation guidelines. Site visited 13.4.2015. http://www.globalw.com/support/hinstall.html
ISO 8368: 1999. Hydrometric determinations – Flow measurements in open channels using structures – Guidelines for selection of structure. International Organization for Standardization, 10 p.
Marr, J., Johnson, S. & Busch, D. 2010. Performance assessment of H flumes under extreme approach flow conditions. University of Minnesota, Project Report No. 538, 23 p.
Mustonen, S. 1986. Sovellettu hydrologia. Vesiyhdistys r.y., Helsinki, 436 p. (In Finnish)
Openchannelflow 2015a. Flumes. Site visited 12.4.2015. http://www.openchannelflow.com/products/flumes/
Openchannelflow 2015b. Weirs. Site visited 12.4.2015. http://www.openchannelflow.com/products/weirs
Rickly Hydrological Company 2015. Weirs and flumes. Site visited 12.4.2015. http://www.rickly.com/sm/WeirsandFlumes.htm
Shaw, E.M., Beven, K.J., Chappell, N.A. & Lamb, R. 2011. Hydrology in Practice. 4th edition, Spon Press, New York, 546 p.
Wolkersdorfer, C. 2008. Water Management at Abandoned Flooded Underground Mines. Springer, 465 p.