An experimental investigation of the impact of aquatic weeds trash racks on water surface profile in open channels

ABSTRACT A considerable quantity of weeds is drafted with flow due to maintenance activities and moving vessels causing serious problem to hydraulic structures and hydropower plants along open channels. Aquatic weed trash screens are normally constructed upstream hydraulic structures to control the drafted weeds and preventing them from passing through the hydraulic structures. The effect of the presence of accumulated aquatic weed upstream trash racks on water surface profile is the main disadvantage of trash racks. In this research, the effect of the presence of trash rack and the accumulated weeds upstream it with different densities on water surface profile is studied experimentally. Ninety runs were applied in the hydraulic lab of Channel Maintenance Research Institute. Equations relating different hydraulic parameters and accumulated weed density to the water surface profile were established. It was concluded that accumulated weed depth is the most effective parameter on heading up, while the blocked area of trash rack screen and extension length of weed upstream the barrier are the most effective parameters on head loss.


Introduction
The drifting aquatic weeds with the water stream adversely affect the efficiency of hydraulic structures in general and hydropower plants in particular. Barriers with trash rack are the main defense line for protecting these structures. Aquatic weed trash rack affects the water surface profile of the open channels where it causes water surface heading up upstream it and head losses downstream it. This effect increases by increasing the amount of weeds and debris accumulated upstream the trash rack. A large body of literature deals with understanding the effect of trash rack on water surface profile and head losses. Kirschmer (1926) studied the effect of trash rack bars, shape, spacing, and inclination on the head losses. He developed the relation, where k is bar shape factor, t is bar thickness, v is the approach velocity, g is gravitational acceleration, and α is the inclination angle with channel bed. Many researchers adopted Kirschmer's formula until today. Levin (1968) and Meusburger, Volkart, and Minor (2001) improved it by taking into account additional affecting factors. They introduced both p as blockage factor, which is the ratio of the area blocked by bars to the gross area of trash rack, and δ as the horizontal flow angle. The resulted equation was, Δh ¼ k 1 þ 0:65 tan δ ð Þ p 1:33 s t À0:43 v 2 2g sin α (2) Orsborn (1968) studied the head losses through rectangular-bar trash rack and baffle. Clark, Tsikata, and Haresign (2010) studied experimentally the energy loss through submerged trash racks. Their adopted equation produced a good result for larger spaces between trash rack bars (s/t > 2). Bradley, Richards, and Bahner (2005) based on laboratory tests stated that the theoretical value of head loss calculated using Kirschmer's formula is underestimated by a factor of 1.75-2.00. This factor is increased greatly when the rack begins to become clogged with debris. It reaches 4.0 with 50% clogging rack. Wahl (1992) introduced an equation to calculate head loss through trash rack based on blockage factor. This formula can be used for any bar shape and angle of inclination. It also can be used for estimating losses for partially clogged trash rack.
where R ¼ 1 À p Department of Engineering and Public Works, USA (2008), adopted an equation for estimating head loss through trash rack. This equation is provided by the United State Army Corps of Engineers (USACE).
This equation can be used for different bar shapes, and bar shape factor for the circular bar is given by Equation (5).
where q is unit discharge. Walczak, Walczak, Hämmerling, Spychala, and Niec (2016) studied and analyzed how the bar shape and the orientation angle of the trash rack as well as the accumulated plant debris affect head losses. They suggested using a trash rack with cylindrical bars inclined by an angle of 80°to have the most beneficial and preferred solution.
Eraky (2016) studied how to maximize the hydraulic efficiency of aquatic weed barriers in open channels. She concluded that keeping both relative weed length (accumulated weed length/top width of trash rack screen) and relative weed depth (depth of accumulated weed/upstream water depth) below 0.4 is recommended for acceptable heading up and flow velocity downstream barriers.
Zayed, El Molla, and Sallah (2018) investigated experimentally head losses through a triangular V-shaped screen. The results indicated that a low screen angle leads to a low screen head loss coefficient, whereas high blockage ratios will decrease the effect of the screen angle.
Flow through trash screens has been investigated by various researchers; Taylor and Batchelor (1949) studied the effect of wire gauze on small disturbances in a uniform stream, Elder (1959) studied the steady flow through nonuniform gauzes of arbitrary shape, Stefan and Fu (1978) studied the head loss characteristics of six profile-wire screen panels, Yeh and Shrestha (1989) studied the free-surface flow through screen, Tsikata, Katopodis, and Tachie (2009) studied experimentally the turbulent flow near model trash racks, and also Raynal, Courret, Chatellier, Larinier, and David (2013) had an experimental study on fish-friendly trash racks.
Most of the trash rack studies are interested in energy and head losses through it and express the resistance of accumulated debris and weeds by blockage percentage. In the present study, the effect of accumulated weed depth and extension length upstream trash rack in addition to the blockage area on both heading up and head loss was investigated. New precise equations relating head loss through trash rack and heading up upstream it with the characteristics of accumulated weeds were deduced.
An extensive experimental measurement was applied on trash rack model built in the hydraulic lab of the Channel Maintenance Research Institute. The various hydraulic parameters that effect on water surface profile were measured and recorded for different weed infestation length and depth. The effect of accumulated weeds (extension length-depth) upstream trash racks on water surface profile was investigated.

Experimental facilities
The experiment was carried out in the hydraulic lab of the Channel Maintenance Research Institute. The used flume is of a recirculating type of length 22.10 m. An underground reservoir of dimensions 24.10 m long, 1.75 m wide, and 1.50 m depth supplies the flume with water. The inlet part of the flume has the dimensions of 4.52 m length, 1.63 m wide, and 1.16 m height. Two vertical reinforced concrete walls are used to dissipate any excess energy in the possible shortest distance. The inlet bed that has a ramp with slope 3:1 is located downstream of the two vertical walls. The main part of the flume is the trapezoidal artificial channel of length 16.22 m, bed width 0.60 m, maximum water depth 0.42 m, and side slope of 1:1. A tilted tailgate is located at the end of the channel to control water depth. The outlet basin's dimensions are 0.96 m long, 1.63 m wide, and 1.21 m height. It drains water through (2−8 inches) diameter pipes as shown in Figure 1. An aquatic weed barrier is simulated using trapezoidal mesh of steel ribs. The vertical ribs are 2.0 cm apart and the horizontal ribs are 8.0 cm apart. All ribs are of 0.2 cm diameter as shown in Figure 2. The weed blockage of both floating and drifted cut submerged weeds was simulated using a material exerted from Palm Park called "Karena" (Photo 1). A wire mesh box with different dimensions (width, depth, and length) filled with Karena is used to simulate the different cases of weed density. The mesh opening was of 5.0 cm diameter to minimize its effect on the flow through the simulated weed blockage and to prevent Karena material from drifting with the flow. Seventy-five runs were carried out in addition to 15 runs for the no weed case. Through the applied runs, five cases of weed density were tested. In the first and second cases, the free water depth under the barrier is 15.0 cm. The depth of weed blockage (D w = water depth −15), while weed lengths (L w ) are 25.0 and 50.0 cm, respectively. In the third and fourth cases, the free water depth under the barrier is 10.0 cm. The depth of weed blockage (D w = water depth −10), while weed lengths are 25.0 and 50.0 cm, respectively. In the fifth case, the free water depth under the barrier is 10.0 cm. The depth of weed blockage (D w = water depth −10) and weed length is 50.0 cm, and the screen is inclined by an angle 12.5°to the vertical direction on the flow. Five discharges (28,31,34,37, and 40 l/s) and three tailgate openings (33.0, 34.0, and 35.0. cm) were used. Table 1 shows the design of

Theoretical approach
Dimensional analysis is used in the derivation of a general equation to relate both of water level heading up upstream the trash rack and head loss downstream it to different hydraulic parameters and the characteristics of the accumulated weeds upstream the trash rack. This equation may be written in the following form: ; Fr u ; Re u Þ ¼ 0 (7) where (H up %) is water surface heading up percentage = (H up /water depth in no blockage case) x 100%, (H los %) is water surface head loss percentage = (H los /water depth in no blockage case) x 100%, (V d ) is flow mean velocity downstream the barrier, (V u ) is flow mean velocity upstream the barrier, (L w ) is the length of gathered weeds upstream of the barrier, (D w ) is the depth of weeds upstream barrier, (A w ) is the area of the trash rack which blocked by weeds, (Y u ) is the water depth upstream the barrier, (Fr u ) is Froud number of the flow upstream the barrier, and (Re u ) is Reynold's number of the flow upstream the barrier.

Results and discussion
Effect of the presence of aquatic weed barrier on water surface profile The presence of aquatic weeds upstream trash rack has a remarkable effect on the water surface profile. It causes the water surface level to higher upstream it (heading up) and to lower downstream it (head loss). The values of both heading up and head loss are related to both accumulated weed depth and extension length. Figure 3 shows a sample of the results of water surface profile measurements.
• Effect of change in accumulated weed characteristics on heading up. • Effect of change in accumulated weed characteristics on heading up.
Generally heading up increases by increasing accumulated weed length or depth. In case 4, heading up is maximum for all discharges as it has the largest values of accumulated weed length and depth. Increasing accumulated weed length is more effective on heading up for  deeper accumulated weed depths (case 3 and case 4) than smaller depths (case 1 and case 2). Rotating trash screen to an angle of 12.5°perpenicular to the flow direction decreases the heading up by an average value 20.7% for the same accumulated weed length and depth for all applied runs (case 5). Figure 4 shows the heading up percentage values for different cases of weed accumulation upstream trash rack.
• Effect of change in accumulated weed characteristics on the heading loss.
Head loss downstream trash rack increases by increasing accumulated weed length or depth. The change of accumulated weed depth has a greater effect on the head loss compared with accumulated weed length. Rotating trash screen to an angle of 12.5°perpendicular to the flow direction decreases the head loss by an average value 12.6% for the same accumulated weed length and depth for all applied runs (case 5). Figure 5 shows the head loss percentage values for different cases of weed accumulation upstream trash rack.
Effect of accumulated weed characteristics and hydraulic parameters on water surface heading up upstream barriers Linear stepwise regression analysis was used to establish an equation relating water surface heading up upstream weed barrier with open channel hydraulic Figure 5. H los % for different cases of weed accumulation upstream barriers.
parameters and the characteristics of accumulated weed upstream the barrier. In linear stepwise regression, the most effective variables are entered and the less effective variables are excluded.  Figure 6. The resulted trend line equation is It means that the predicted values for heading up from Equation (8) are almost equal to that measured in the flume, and it could be trusted to evaluate the values of heading up upstream barriers.
Effect of accumulated weed characteristics and hydraulic parameters on water surface head loss downstream barriers Using linear stepwise regression analysis, the equation relating water surface head loss downstream weed barrier with open channel hydraulic parameters and the characteristics of accumulated weed upstream the barrier is in the form The most effective variables on head loss are flow velocity, the blocked area of the screen barrier, and extension length of weed upstream the barrier.  values of measured head loss and the predicted values from the equation is shown in Figure 7. The resulted trend line equation is: It means that the predicted values for head loss from Equation (10) are almost equal to that measured in the flume, and it could be trusted to evaluate the values of head loss downstream trash racks.

Conclusions and recommendations
Based on the result of the analysis of the experimental measurements of the study, it was concluded that: (1) Aquatic weed barriers with trash rack effect on water surface profile. It increases the water level upstream it (heading up) and decrease the water level downstream it (head loss).
(2) The values of both heading up and head loss are directly proportional to both length and depth of accumulated aquatic weed upstream trash rack. (3) Rotating trash screen to an angle of 12.5°perpendicular to the flow direction decreases heading up and head loss by an average value 20.7% and 12.6%, respectively, for the same accumulated weed length and depth. (4) The most effective variables on heading up are flow velocity and accumulated weed depth. Heading up may be estimated using Equation (8). (5) The most effective variables on head loss are flow velocity, the blocked area of the screen barrier, and extension length of weed upstream the barrier. Head loss may be estimated using Equation (10).
It is also recommended that periodic maintenance program must be applied to maintain trash racks. The time schedule of this program is designed according to the quantity of weeds and debris that drifted by the flow.  Figure 7. Comparison between measured and predicted H los %.