Md. Akhtaruzzaman Sarker1,
David G. Rhodes2 and
Gregory S. Armstrong3
1Lecturer, Civil
Engineering Department, Bangladesh Institute of
Technology,
Chittagong
4349 Bangladesh
Tel:
880 31 714948 Fax: 880 31 714910
E-mail: masarker@hotmail.com
2Senior
Lecturer, Engineering Systems Department, Cranfield University
Swindon SN6 8LA UK
Tel: 44 1793 785643
Fax: 44 1793 783192
E-mail: rhodes@rmcs.cranfield.ac.uk
3National Fish Pass Officer,
Environment Agency Wales
Llys Afon Hawthorn Rise Haverfordwest
Pembrokeshire SA61 2BQ
Tel: 44 1437 760081 Fax: 44 1437 760881
E-mail: greg.armstrong@environment-agency.gov.uk
Abstract: The low-flow section of a compound Crump weir was modelled at 1/5 scale to investigate the effect of low-cost modifications to the downstream face to improve fish passage. These consisted of baffles with a slot at each wall, and the arrangement was varied in terms of baffle spacing and slot width. Measurements at the 95 percentile low-flow indicated that the modifications, though not yet a satisfactory solution, offer potential that merits further work.
Keywords: fish pass,
hydraulic structure, crump weir
Historically
man-made and natural developments have threatened inland fisheries, but in the
UK the impact of engineering works has been most evident during the past few
centuries of industrialisation. The most damaging effect has been that of
pollution, but structural impacts upon fish habitats have compounded the
problem. That has included the impediment of fish movement by river engineering
works themselves: for example in a recent case (Lucas and Frear 1997), a flat-V
weir on the River Nidd in Yorkshire had such a detrimental effect upon the
populations of barbel and other species that it was decided to remove it.
During the last
century a variety of fish pass designs have been developed and successfully
implemented. Normally, these have been hydraulic structures with a separate
conduit that bypasses the obstruction and provides fish access to upstream
waters. However, there are many thousands of obstructions in the UK rivers and
it would be prohibitively expensive to apply formal technical fish pass
solutions to all of those sites. Because of the scale of the problem, the
Environment Agency in England and Wales has a long term programme of remedial
action, focused on the most seriously affected and valuable sites. In recent
years, purpose built fish passes have been commissioned but, in order to make
significant progress on the broader front, low-cost solutions are frequently the
only realistic means to that end.
This paper reports
on the early stage of an investigation into potentially low-cost solutions for
facilitating fish passage. The site chosen was the compound Crump weir at
Brimpton, on the Enborne in the Thames catchment. The most common fish species
in that locality are chub, barbel, dace, brown trout, grayling and pike, with
roach and carp present in small numbers.
Experimental work
was carried out on a 1/5-scale physical model of the low-flow section of the
Brimpton weir. Froude number equality was chosen for the study with uniform
scaling in all directions. The model fitted full width in a glass lined,
straight, 0.613 m wide × 0.585 m deep × 8.262 m long recirculating flume and
the general arrangement is shown in Figure 1.
Fig. 1 General arrangement of laboratory flume with Crump weir
A 60 l s–1
pump discharged via a butterfly control valve into a 150 mm dia. delivery main,
housing a flow straightener 22 pipe diameters upstream of a 38 mm dia. orifice plate for flow measurement. From
the orifice plate, flow was conveyed into a hood of rectangular section designed
to spread the exit jet across the flume inlet chamber. Following another flow
straightener, a contraction in the inlet chamber walls converged the flow into
the working section in which the model weir was housed. At the downstream end of
the flume, flow depth was controlled by a full width, adjustable weir
discharging into a sump connected to the pump inlet. Longitudinal free surface
profiles were measured to ±0.1 mm by a pointer gauge fixed to a carriage,
positioned manually at the required x-coordinates to an accuracy of ±0.5 mm.
Precise
positioning of measuring devices was provided by means of a 3-axis traverse
gear, driven by computer controlled stepper motors. The range of travel was
nominally 1.1 × 0.6 × 0.6 m in the longitudinal x, vertical y and spanwise z directions respectively. The resolution
was 1/4000 mm, though in practice accuracy of
instrument positioning was limited to ±0.05mm by the methodology
adopted. The traverse gear was used to measure the transverse free surface
profile with a pointer gauge and the primary velocity distribution with a pitot
tube and separate static pressure tube. The dynamic pressure was sensed by 5 and
20 mb eddy current diaphragm pressure transducers, with 0–10 V output to the
analogue to digital converter in a PC. Pitot and static pressure tubes of 2.05
mm outside diameter were employed, and in combination gave a response time of
slightly less than 60 s. Water temperature was measured by an electrical sensor
attached to the inner surface of the flume wall.
Fig. 2 General
arrangement of baffles, with and without slots
To date attention has been
focused on the 95 percentile low flow rate, with less detail on higher flow
rates. The modelled flow of 2.97 l s–1, corresponding to the 95
percentile low flow at field scale, was firstly applied to the unmodified model
weir. A tail-water level equivalent to that recorded at the full scale was
established by means of the adjustable exit weir, and sufficient time was given
for the flow to settle to a constant temperature. Using the pointer gauge/
carriage system on the channel centreline, the longitudinal free surface profile
and the weir profile were then measured relative to the weir crest as datum.
Next, the model Crump weir
was modified by fixing, full width of the channel, baffles 30.4 mm high and 1.1
mm thick to the downstream face of the weir (Figure 2). The most upstream baffle
was carefully located downstream of the weir crest, far enough below it to
ensure that there was no effect on the water level upstream of the weir. This
maintained, at the 95 percentile low flow, modular flow conditions for
hydrometric purposes identical to those of the unmodified weir. The dimension
between the first upstream baffle and the weir crest was 170 mm, measured down
the slope. With the most upstream, full width baffle fixed in position, the rest
of the baffles were installed downstream at a uniform spacing, determined by a
trial and error process designed to achieve a thick flow over the baffles and an
adequate gap between them for fish resting. The finally adopted 60 mm spacing,
corresponding to 300 mm gaps at field-scale, gave a near maximum flow thickness
over the 17 baffles.
However, the modelled flow depth above the full-width baffles was barely adequate for larger fish and therefore full-depth slots in the baffles were added in order to provide deeper, straight channels for fish movement. Following trials with slots 12, 22 and 32 mm wide at each side wall, the largest slot width of 32 mm was selected for further study (Figure 2) on the grounds that it provided a flow depth in the slot of almost full baffle height and at field scale a passage width of 160 mm, likely to be adequate for many coarse fish species and individuals.
Fig. 3 95 percentile low-flow over baffles with 60 mm spacing and no slot
The pointer gauge method implemented on the unmodified weir proved to be time consuming for profiling the vigorously fluctuating free surface over the baffles, and so photography was adopted as a supplementary method. The digital camera employed was a FUJIX DS-300, with a shutter speed of 1/125 s, though the flash duration of 1/1000 s actually controlled the exposure time. For each flow, 16 to 19 frames were exposed at random times. In every frame, the pointer gauge was set at a known vertical dimension from the weir face, thus providing a focal plane for the camera and a reference dimension by which the free surface profile and average depth of flow could be measured. Figure 3 shows a typical photograph of one of the baffle arrangements, with the horizontal line of sight transverse to the flow.
For the slotted baffle, the transverse average free surface profile
was obtained from maximum and minimum values of the fluctuating water levels,
using the motorised traverse gear and a pointer gauge. A velocity traverse was
carried out in the plane of the 9th baffle from the weir crest (in the developed
flow). The nose of the pitot tube was located 1 mm upstream of the baffle plane,
with the holes of the static pressure tube at the same streamwise position and
the tube axes 10 mm apart. Slightly more than half of the channel width (340 mm)
was traversed. Velocities above the baffle were taken at z-intervals of 5 mm and y-intervals of 1.5 mm and, in the slot
below the top of the baffles, at 3 mm intervals in the y and z directions.
For the 95 percentile low flow, the water depth directly above the weir crest was maintained at a constant 14 mm (70 mm at field scale). In the unmodified model, minimum depth on the downstream face was 4.6 mm and maximum depth-averaged velocity 1.06 m s–1. With the same Reynolds number (Re), field-scale depth would be 23 mm and the velocity 2.38 m s–1, and at the field-scale Re the actual depth would be slightly less and the velocity greater. That flow depth would be too small for larger fish and the velocity too great for smaller fish, and therefore the unmodified weir would present a complete barrier to fish passage at that flow rate. The effect of installing baffles full width was to raise the average flow depth above the downstream face of the weir to 48.6 mm, 243 mm at field-scale. The field-scale depth above the baffles was 91 mm, less than the minimum depth of 100 mm recommended for trout (Larinier 1992), but an improvement on the depth of 23 mm on the unmodified weir.
Fig. 4 Transverse distribution of flow depth in the plane of a 32 mm slotted baffle
The presence of the baffles, combined with the undulating free surface (Figure 3), would have still presented a physically demanding ascent path, whereas the two 32 mm wide slots presented straight migratory passages with greater flow depths as well. Figure 4 shows the transverse distribution of flow depth in the plane of a slotted baffle. Above the baffle the depth varied in the range 4.7 to 8.8 mm, and in the slot the minimum depth was 31.2 mm. At field scale the corresponding depths are 23.5 to 44 mm above the baffle and 156 mm in the slot, the latter quite adequate for the larger fish in that locality. Figure 5 shows the transverse velocity distribution in the model, in the plane of a baffle, at y = 30.2 mm. Velocities above the baffle varied in the spanwise direction in the range 0.3 to 0.4 m s–1, ie 0.67 to 0.89 m s–1 at the field-scale. As shown in Figure 6, there was a considerable variation in velocity with y-coordinate along the centreline of the slot. The maximum velocity was 1.22 m s–1, (2.73 m s–1 at field-scale) and actually higher than the maximum depth-averaged velocity on the unmodified weir. This high maximum velocity was due to lateral diversion of flow into the slot in combination with the steep vertical velocity gradients induced by bed shear and lateral shear-layer interaction with the baffle flow. However the velocity was still within the capability of good swimmers such as trout. The slot flows emerged as wall jets downstream of the weir, and at field scale their significantly higher velocities and the associated noise might well serve to attract fish to the passage entrance.
Fig. 5 Transverse distribution of velocity in the plane of a baffle, at y = 30.2 mm
Fig. 6 Variation of primary velocity with y-coordinate on centreline of slot
A model study
has demonstrated that simple, low-cost modifications to a Crump weir, may offer
a means of considerably improving conditions for fish passage while leaving the
hydrometric function unimpaired. Though not as efficient as conventional fish
pass baffles, they can give a substantially increased flow thickness and the
increased heterogeneity of the water velocities is a feature that may also be
exploited by fish. Final conclusions depend upon further work in hand to extend
the range of flow rates and baffle configurations.
Acknowledgements
The work was funded
by the Engineering Systems Dept., Cranfield University while the first author
was on study leave from the Bangladesh Institute of Technology (BIT), Chittagong
.
References
Larinier, M. (1992). Les passes a ralentisseurs. Bull. Fr. Peche Piscic. 326-327, 73-94.
Lucas, M.C. and Frear, P.A. (1997). Effects of a flow gauging weir on the migratory behaviour of adult barbel, a riverine cyprinid. J. Fish. Biol., 50, 382-396
Sarker, M.A. (2000). Application of CFD to modelling local features in a river: free surface flow over broad-crested weir and modification of Crump weir for fish passage. PhD Thesis. Royal Military College of Science, Engineering Systems Dept., Cranfield University.