ENGINEER - Vol. XXXX, No. 04, pp. 129-133, 2007 © The Institution of Engineers,

Flow Modelling of Reach with the Riverbed Intake of Udunuwara-Yatinuwara Water Supply Scheme K. H. S. S. Karunanayake and S. B. Weerakoon

Abstract: The Udunuwara-Yatinuwara water supply scheme of the National Water Supply and Drainage Board abstracts water through a riverbed filtration system located in a bifurcated channel of the Mahaweli River at . The diverted flow through the channel has become inadequate to supply the abstraction demand during the periods of low river flow. The flow pattern in the river and the discharge through the channel are investigated by the application of a two-dimensional depth- averaged flow computational model to a 450 m long river reach including the channel. The model was set up, calibrated and verified for the reach using measurements carried out at the site. The model was applied to ascertain the possibility of increasing the flow through the channel during low flow in the river by changing the river bed level in the reach. Keywords: Flow computation, modelling, water intake

1. Introduction to flow over the filters to enable extraction of the demand even during periods of low flow in the Udunuwara-Yatinuwara water supply scheme is river. There are various options available to one of the major water supply schemes of the solve this problem; some of them are changing National Water Supply and Drainage Board part of the river bed level by armouring and/or (NWS&DB) providing drinking water to desilting, construction of groyne, spur dyke, Udunuwara, Yatinuwara divisional secretarial compound weir, and weir. In this paper, only divisions. At present, it serves a population of the potential of diversion of adequate flow into about 90,000, various institutions and industries the bifurcated channel by changing the river bed in the area. The scheme consists of three water level is investigated using computational intakes located at Alpitiya, Nillambe and hydraulic modelling. However, detail Peradeniya. investigation of each option giving due The present study is focused on the intake at considerations to all issues is required to achieve Peradeniya, sited on the right bank of the an optimum solution for implementation Mahaweli River near the Sarasaviuyana railway purpose. station. The intake has been constructed to Computational models for rivers are essentially 3 extract 4600 m of water daily into two collecting based on the theoretical formulations relating wells on the right bank of the river through two the flow, fluid and geometric parameters river bed rapid sand filters. River flow is derived by the application of conservation laws bifurcated by an artificial island and one of the and some constitutive relationships. It is the bifurcated flow is arranged to flow over the usual practice that varying approximations, river bed sand filters. Filtered water is collected justifiable to the river flow situation in hand, are through perforated pipes 'and arranged to flow made to the model equations in order to obtain into the collecting wells. Though the intake a solution with acceptable level of accuracy. functioned satisfactorily at the initial period Among the computational models available, after construction in 1991, the bifuicated flow one-dimensional river models based on the over the filters is not sufficient to supply the demand during low flow period of every year which is about 30% of the time at present. K.H.S.S.Karunanayake, B.Sc.Eng. (Peradeniya), M.Sc.Eng. (Peradeniya), Engineer, National Water Supply and Drainage Board, Therefore, it is required to find a long term Geteniabe, Peradeniya. Eng. (Dr.) S.B.Weemkoon, B.Sc.Eng. (Peradeniya), M.Eng. (Tokyo), engineering solution which is economical and D.Eng. (Tokyo), C.Eng., MIE(SL), MSLAAS, Senior Lecturer in Civil environmentally sound to divert adequate water Engineering, . Peradeniya

129 ENGINEER Saint Venant equations have been applied to water is assumed to be negligible compared river systems for gross prediction of flood water with the gravitational acceleration yields the levels, discharge distributions and sediment depth-averaged equations applicable to shallow transport along rivers. The applications of three• free surface flows (Vreguendhil and Wijbenga, dimensional computational studies have been [5]). The effective stress tensor components in limited, mostly, to laboratory chann el flows with depth-averaged equations are modelled by a rectangular cross-sections, mainly due to the gradient-diffusion model (Molls and Chaudhry numerical in stabilities en cou ntered in the [4]). The shear stress on the free surface is computations when the irr egular boundary neglect ed and the bott om shear stress is topography inherent to natural rivers is taken modelled by relating it to the local free surface into account. On the other hand, the depth• slope. Then, the local free surface slope is averaged models, which are economical, are computed from Manning's formula. The steady ap pli ed to sh all ow water rivers. The state depth in tegrat ed tw o-dimensional applications of dept h-averaged mod els in mom entum equati on s an d the conti nuity Cartesian coordinates to natural rivers have equation are given in Eqns. (1) to (3). been presented by ASCE [1]. Also, Wijbenga [9],

e e a [ Y a +h v° u ... W nka t l. 8], e n d McCorquodale, hu OU + gh 017, .i: o' u +E o' u + gun' (u' +v') -0 (1) [ 2 l OX OY OX p eD 0X ,Y i)y' hy, Weerakoon [6], Weerakoon et.al [7] presented applications in -curvilinear coordinates which hu + + h 17, !!._ o v]+ gvn '(u + v ') � h v� g o _ [ E o' v +E ' ' 0 ... (2) are more suitable for applications to rivers. ax i)y i)y p ,. ox' ,,. i)y' hY,

Objectives of this study are: ...... (3) • to calibrate and validate a depth-averaged computational model to the river reach of the Mahaweli River including the river bed Here, h = the water depth, z = the vertical water intake oft he Udunuw ara-Yatinuwara direction, z b = the bed elevation, z, = zb +h is the water sup ply sch em e, wh ich has been water surface elevation, g = the gravitational plann ed to abstract 4600 m3 / d, acceleration, fl,= the water sur face elevation, p= the density of fluid, h = the local water depth, • predict flow pattern when the river bed level Eif= the eddy viscosity coefficient in direction is changed, and j l ' l ' • to ascertain the changing of river bed level as on su rface , u=,; Judz and v=,; J vdz are the / z r• • one of the potential solution to increase horizontal velocity in the x direction at a point diversion through the chann el. along the vertical coordinate direction and the horizontal velocity in they direction at a point Accordingly, Sur face Water Modelling System along the vertical coordinate direction respectively. (weblink [11]) which is a tw o-dimensional hydrodynamic modelling system using finite• 2.2 D -a ra e Model - Surface Water element analysis is set up to the river reach of epth ve g d Modelli ng System 450 m length. The validated computati onal model is then applied to obtain the velocity The Surface Water Modeling System (SMS), pattern and water depth under different bed which is a com prehensive enviro nment for topography. hydrodynamic modelling includes an interface for the two-dim ensional finit e-elem ent 2. Computational Model hydrodynamic model RMA2 (U.S. Army Corps of Engineers). The RMA2 hydrodynamic model 2.1 Governing equations is based on Eqns. (1) to (3) and iterative solution is computed for velocity and depth at discrete Equations that describe the flow in surface points in a river reach under given boundary w e bo es a e bas o t e cl ss c c c at r di r ed n h a i al on epts cond iti on s. SMS mesh modu le is used to o c s i o a e . f on ervat on f mass nd mom ntum Depth• construct two-dimensional finite element mesh wise integration of the momentum equations of the river reach . The model parameters, and the c i e wi h the ont nuity q uation t sour ce/ sink data, and boundary conditions are a s i tha the vertical accele a ion of s umpt on t r t assigned directly to the nodes, and elements of • ENGINEER 130 the mesh. With a mesh constructed, the RMA2 is the velocities at the grid points when the used to compute flow velocities and water- upstream discharge was 9.7 m3/ s and urface elevations at each mesh node, and the abstraction rate from each filter was 0.027 m3 / s boundary between wet and dry regions in the (Case I). The boundary condit ions for the model. SMS with RMA2 model was setup to computations were the upstream discharge and

investigate the flow in a 450 m long reach of the the downstream water calibrated for the river Mahaweli River which includes the existing reach to agree the model predicted resultant river bed water intake of the Udunuwara• velocities with the depth-averaged values of the Yatinuwara water supply scheme. measured velocities across a river section. This was achieved by refining the Manning's coefficient ( ) and the 3. Model Calibration and Verification roughness n eddy viscosity () within the ranges typical to the given river reach comprising of highly rough river bed with 3.1 Field Measurements boulders and drops (Chow [2], Wijbenga [9]). The In order to calibrate and verify the model, computed results are found to be acceptable with stream flow measurements were carried out in the measured values for then value of 0.025 and 2 / the river reach using the survey instruments, eddy viscosity of 1.2 m s. Comparison of electromagneti c currentmeters fixed to the measured and computed velocities at the river wading rods, and depth poles while stationed sections A-A and C-C are shown in Figures 2 and 3. Computed velocity vectors for of the reach on a dingy/ boat by trained technical staff. part River morphology and the bed contours in 450 are shown in Figure 4. m long river reach which includes island and --+- observed -- computed the channel are shown in Fig. 1. Measurements under two different flow conditi ons were 0.6 carried out for mod el calibrati on and '""'"' ]_ 0.4 verification. They correspond to river discharges � of 9.7 3 (C I) and 11.7 3 (Case II) from m / s ase m / s � 0.2 .; the upstream respectively, with an abstraction >- rate of 0.027 m3 / s through each river bed filter. The downstream water surface elevations in 0 20 40 60 80 Case I and II were 462 m and 462.11 m MSL Distan ce/(m)

respectively. The velocity measurements were Figure 2: Comparison of Velociti es at carried out at discrete points along cross section Sect ion A-A f or Case I (A-A), (B-B) and (C-C) (Figure 1). using an electro-magnetic currentmeter. The depths and --+- computed --- observed the water surface elevations in these sections and the most downstream section of the reach i:::::- 0 .01.52 were also recorded. .f' 0.1 � 0.05 l � O+- � -,--� � -.-� --,-� � -.-� --,-� --, 3.2 Calibration 0 2 3 4 5 6 Distance/(m) A finite element mesh was generated to cover Figure 3: Comparison of Velocit ies at the reach and the model was applied to compute Sect ion B-B f or Case I

A R

9.75Sm3/s

0.027m3/s 0.027m3/s

Figure 1: The Mahaw eli River Reach of 450m Length, Cont ours are in m MSL 131 EN GIN EER • The independence of the predictions on the grid 4. Model Application size was verified by carrying out the model simulations using a grid with refined mesh size The Mahaweli River flow at Peradeniya shows a under the same parameters and the boundary significant variation due to catchment rainfall v i io , i e se se o ri e w e flo conditions for Case I. No changes were observed ar at n d v r u s f v r at r and w between the results of velocity and water regulation at the upstream (Karunanayake [3]). surface elevati on obtained by the two Among them, the flow regulation by the simulations, and the model results are therefore Kotrnale reservoir with a capacity of 174 MCM i ac s t e Ko a a j independent of the grid used for computation. mpounded ros h tm la Oya, ma or tributary, is pro m inent. The flow in the Mahaweli River during 30% of the year fails to produce adequate flow through the bifurcated channel for the abstraction of 0.054 m3 / s �-1: , , , :-11� . . , through the filters of the Uduuwara-Yatinuwara water supply scheme. For the purpose of present ...... • , .. ... 1. . .. , •• • • analysis the authors consider that the probability of failure of the scheme shouldn ot exceed 5%. The river flow at Peradeniya. The analysis carried out using the daily flows of the Mahaweli River at Polgolla over 16 years from 1989, the river flow Figure 4: Celocity Vect ors in Case I (Scale: 0.25m/ s) at Peradeniya with non exceedance probability 3.3 Verification of 5% is found to be 5 m3/ s.

The calibra ted model was next verified by Modification of the river bed elevation in the applying it to compute flow pattern under the reach by armouring (placement of boulders or measured discharge of 11.7 m3/ s and the water gabions) or removing silt/ gravel deposits in surface elevation of 462.11 ma t the downstream different parts of the river bed where applicable (Case II). The computed velocities are compared is one of the many potential solutions listed in with the measured velocities at the section B-B Secti on 1. The proposed bed elevation is (Figure 5) and model results were found to be depicted in Fig. 6. The mean rive bed elevation acceptable. in the downstream cross section in the reach is kept unalt ered to minimize the river bed changes that might continue downstream. The ---+- computed ---- observed modified bed by stacking the boulders etc. is expected to produce a less roughness to the flow than at present.

The calib a ed a d ve ified 20 40 60 80 r t n r depth-averaged Dist ance across the river /(m) computational model was applied to estimate the velocity pattern for the discharge of 5 m3/ s Figure 5: Comparison o Velociti es at f under the proposed river bed (Case III). Section A-A f or Case II Manning roughness coefficient of 0.021 and eddy viscosity coefficient of 1.2 m2 / s were used. The downstream water surface elevation of

·"------• 461 - - - - �

461

Figure 6: The Mahaw eli River Reach of 450m Length wit h the Modified Riverbed Cont ours are in m MSL • ENGINEER 132 461_m was based on the measur ements at low scheme. However, detailed investigation of each flow at present. Since the mean bed level of the option giving due considerations to all issues is downstream section of the reach is unaltered required to achieve an optimum solution before this measu red value is con sidered to be a implementation. reasonable estimate. References Computed velocity vectors for the reach are show n in Fig 7. The disch arge through the 1. ASCETC (ASCE Task Commi ttee on Turbulence bifurcated chann el becomes as 0.3 m3 / s with the models in Hydraulic Computations): Turbulence modified bed, whereas that at the present bed is modeling of sur face water flow and transport. J H ydr. Engrg. AS CE, 114(9), 1989, pp. 970-1064. only a tiny fraction of the demand( < 0.01 m3/ s). 2. Chow, V. T., Maidment, D.R. , and Mays, L. W.,. "App li ed Hydr ology", McGrew -H ill Book ------Company, New York, 1988. ------3. Karu nanayake, K.H.S.S.: "Modelling of flow in ------the Mah aweli Riv er near the in take of ------Udunuwara-Yatinuwara water supply scheme", ------MScEng Thesis, Dept. of Civil Engin eering, ------of P Sri Lanka, 2007. ------University eradeniya, ------3. Molls T. and Chaudhry:W.H .: Depth-averaged

open-chann el flow model. J of H ydraulic Engrg., - - - .,. "'#' .. ------AS CE, 121(6), 1995, pp. 453-465. 4. Vreugden hil C. B. and Wijb en ga J. H. A. Figure 7: Velocity Vectors in Case III (Scale: O.Sm/ s) Computation offl ow patt erns in riuers. J H ydr. Div. AS CE, 1982, 108(11), pp. 1296-1310. However detailed investigations of altern ative 6. Weerakoon S.B .. ." Ap p licati on of a two• l 1 options isted in Section need to be carried out dimensional depth-averaged flow computational considering overall im pacts. This study is model for aquatic habitat simulation", Proc. Tenth limited to ascertain the riverbed modification as Asian Congress of Fluid M echanics, Sri Lanka, May 2004. one of the potential options. Furthermore, even for this option the most suitable bed profile, the 7. Weerakoon S.B., N. Tam ai and Y. Kawahara : st able armou r size need to be selected "Flow computation at a river confluences", J of W ater and M aritime Engineering, Inst. of Civil considering flood disch arge and sediment Engineers, United Kingdom, Vol I, 2003. transport in the reach. 8. Wenka Th, Rodi W. and Nestman F. : "Depth• averaged calculation of flow in river reaches with 5. Concluding Remarks flood con tro l an d sedim ent regulat ion structures". Proc. 25th Congress of ! AH R, Tokyo, A two-dimen sional depth averaged A(l ), pp. 1993, 41-48. computational model was set up to the river 9. Wijben ga J. H. A.: "Determin ation s of flow reac o t e M a eli Riv , w ic incl e t e h f h ah w er h h ud s h patterns in rivers with curvilinear coordinates". river bed intake of the Udunuwara-Yatinuwara Proc. 21st ! AH R Congress, M elbourne, 2, 1985, water supply scheme. The model was calibrated pp.132-138. a d valida e t e· fiel measu emen s. n t d using h d r t 10. Ye J. and McCorquodale J.A. : "Depth-averaged The model can be used to investigate flow hydrodynamic model in curvilinear collocated pattern under different river flows. grid". J ef H ydraulic Engrg., AS CE, 123(5), 1997, pp. 380-388. The model was applied to compute the velocity 11. http:/ / w ww .scien tificsoftwaregro u p .com / patt ern in the river reach with mod ified pages visited June 2006. riverbed profile in the reach under the river flow of non exceeda nce pro babilit y of 5%. It is ascertained that bed level modification in the reach is one of the potential options to divert adequate discharge into the bifurcated chann el with riverbed filters to supply the demand for the Udunuw ara-Yatinu w ara wat er su p ply

133 ENGINEER I