Talla Reservoir

Client: AECOM
Scale: 1/8

A physical model study was undertaken of the scour pipe to the Talla Reservoir.  The model study was carried out at 1/8th scale and operated on the basis of Froude Law similarity.

It is proposed to extend the existing scour pipe and to install a new diffuser pipe to protect the adjacent supply pipes, supports and walkways due to concerns regarding safe operation of the existing scour facility.

An initial configuration of scour pipe extension and diffuser arrangement was proposed for testing and evaluation.  The proposed arrangement was confined by the requirement to minimise interference with existing pipework, supports and walkways, together with restricted access for installation.

A combined approach of physical and numerical modelling (Computational Fluid Dynamics, CFD) was adopted due to scale considerations in adequately simulating the behaviour of the diffuser port arrangements at reduced size.  Prior to evaluation of the initial proposed arrangement a calibration exercise was undertaken using the physical model output of the scour pipe arrangement to verify the CFD results.

A calibration exercise was undertaken of the numerical model to provide a satisfactory correlation in behaviour between the total pressures of the physical model and the CFD simulation using the scour pipe alone.  This allowed the characteristics of the full scour pipe and diffuser array to be reliably predicted by the CFD simulation.

The arrangement of the scour pipe extension was dictated by the requirement to minimise interference with existing pipework, valve chamber arrangement and walkway support structures.  Rapid changes in the pipe geometry were required and these resulted in the development of significant velocity variation and pressure gradients across pipe sections.

An alternative arrangement of the scour pipe extension was developed that comprised a smoother transition between the 900 mm diameter scour valves and the proposed diffuser pipe at 600 mm diameter.  Testing determined that improved characteristics were achieved with the revised scour pipe and this arrangement was recommended for installation.

An initial configuration of the diffuser array, Diffuser Option C, was proposed that comprised a 600 mm diameter pipe section and a series of slotted ports of varying size arranged along the length of the pipe.  A pipe diameter of 600 mm represented the maximum permissible size of circular section pipe that could be installed within the confines imposed by the site.

The initial configuration of the diffuser did not result in acceptable performance due to unequal flow distribution and excessive outlet velocity.  It was considered that appreciable improvement in the performance of the diffuser arrangement would not be achieved by alteration of the diffuser port sizes and patterns alone.  Developments were therefore first considered to investigate the extent of improvement in performance that could be achieved utilising a 600 mm diameter pipe section for the diffuser.

Diffuser Option 1 comprised internal baffles, set with concentric circular orifices, which reduced in size along the length of the diffuser.  The orifice plates were located at the entry to each of the diffuser sections.  The port configuration of the diffuser sections was modified to a standard pattern with larger ports set upstream and smaller ports downstream.

The Diffuser Option 1 arrangement did produce a significant improvement in the distribution of flow and reduction in slot velocities compared to the original arrangement of Diffuser Option C.  The specified limit of 3 m/s per port was however exceeded at certain slots and the capacity of the pipe was reduced.

Concerns were raised regarding the use of internal baffles which limited the viability of further improvement in the performance of the 600 mm diameter diffuser pipe.  Alternative means were therefore considered to increase the available section of the pipe within the area bounded by the pipework, walkway and supports.  An elliptical pipe section was proposed for investigation.

Diffuser Option 2 comprised an elliptical diffuser pipe section 600 mm high and 1200 mm wide.  The elliptical diffuser pipe comprised seven 1 m long sections, each section housed three sets of ports of varying sizes with four ports per set.  The end plate of the diffuser pipe was set with sixteen 50 mm diameter orifices.

Diffuser Option 2 exhibited an increase in the steady state flow and relatively uniform division of flow was developed through the ports along the length of the diffuser.  The diffuser ports continued to exhibit velocities in excess of 3 m/s downstream of the first diffuser section with velocities significantly higher than 3 m/s evident in the furthest downstream section.

Possible measures were considered to reduce pipe flow and increase head loss down the length of the elliptical diffuser.  Diffuser Option 2a comprised Option 2 plus a 600 mm diameter internal baffle plate upstream of the diffuser entry, together with alteration of the diffuser port configuration.  The end plate of the diffuser was similarly set with 16 ports 50 mm in diameter.

Diffuser Option 2a exhibited a reduction in the capacity of the scour pipe and negative static pressures were developed downstream of the internal baffle.  Net inflow was generated to the first diffuser section, together with low outflow to the ports of the second section.  The diffuser ports continued to exhibit velocities in excess of 3 m/s downstream of the first diffuser section with velocities significantly higher than 3 m/s evident in the furthest downstream section.

Further to review it was considered that the elliptical diffuser pipe option was becoming too complex to allow ready fabrication, a return to the original 600 mm diameter diffuser pipe was therefore proposed.  Diffuser Option 3 was proposed, this comprised 7 sections of 1 m length, each with 3 sets of ports, each set comprising 16 ports, hence 48 ports in total per length of diffuser.  The port sizes reduce from upstream to downstream with equal size ports located within each 1 m section of the diffuser.  The end plate of the 600 mm diameter diffuser pipe was similarly set with 16 ports 50 mm in diameter.

The revised Scour Pipe & Diffuser Option 3 exhibited a steady state flow of 5333 l/s, which was in line with specified requirements, and the flow distribution to the diffuser ports increased steadily down the length of the diffuser.  The diffuser ports continued to exhibit velocities in excess of 3 m/s downstream of the first diffuser section with velocities significantly higher than 3 m/s evident in the furthest downstream section.  Relatively high velocities into the diffuser continued to result in less outflow through the ports in the upstream sections of the diffuser.  As a result, outflow through the ports in the first section of diffuser was approximately one third of the outflow to the ports in the last section of the diffuser.

Development and optimisation of the scour and diffuser pipe assembly determined that the specified requirement for velocity not to exceed 3 m/s at the diffuser outlet ports could not readily be achieved within the confines imposed by the existing valve chambers and support structures.  A velocity in excess of 3 m/s at the diffuser ports was subsequently considered acceptable in view of the relatively low momentum exerted at the small cross sectional area of the individual jets.  It was regarded that these jets would be readily dissipated and diffused a short distance downstream of the ports and at the surrounding secondary baffle plate in the depth of water generated within the tunnel due to backwater effects.

Physical model testing of the tunnel and valve chambers was undertaken following determination of the final proposed scour and diffuser pipe arrangement.  The central valve chamber of the tunnel section is to be modified with a maintenance access through the concrete plug upstream of the shaft.  A review of the hydraulic effects of the new access on the tunnel arrangement was undertaken.

An analysis of the existing tunnel system was undertaken to determine the water level within the tunnel section downstream of the central valve chamber.  The physical model was operated at this downstream water depth and the water levels were determined within the central valve chamber and the upstream section of tunnel.  Conditions were compared with and without the access tunnel through the plug upstream of the central valve chamber.

Testing determined that the proposed access tunnel through the plug upstream of the central valve shaft resulted in a more stable and uniform flow regime within the downstream section of tunnel together with lower velocities within the valve chamber and tunnel.  It was considered that operation without the access tunnel carried a high risk of structural damage within the central shaft and downstream tunnel together with increased risk of blockage.