Pumpstation Suction Design: Farm Dam Siphon Case Study

Two farms had the same problem. Neither pumpstation could draw enough water from its dam. The obvious solutions involved new pumps, major civil works, or both. We found another way.

At the first farm, a new 135 ha drip irrigation development, the existing Ø200 mm outlet through the dam wall was undersized and heavily corroded.

The second farm had an existing pumphouse in poor condition, supplied by booster pumps on a raft pushing water over the dam wall. A major pumphouse upgrade was already planned, including an additional area that had previously been supplied from elsewhere.

The constraint was not the pumps or the pipework downstream. It was the suction side, before water even reached the pumphouse. Fixing it meant choosing between expensive options, until we looked more carefully at what was actually available.

We assessed each option against capital cost, operating cost, maintenance burden and hydraulic impact on the downstream system.

Option 1: New outlet through the dam wall

Drain the dam, cut a new opening, install a properly sized pipe. Hydraulically the cleanest solution, with positive pressure the entire way. The cost is the dam going offline for a season while the civils work is done.

Option 2: Siphon over the wall

A floating intake draws water from the dam. A pipe runs over the crest of the wall and down the other side. Once primed, it runs continuously while the pumps are operating. No new pumps, no permanent loss of dam storage. The lowest capital cost of the five options, and the fewest moving parts.

Option 3: Over-wall pipe with a submersible booster on a raft

Maintains positive pressure throughout. Also means a pump hanging in the dam that needs servicing. Quoted pump costs at this duty ranged from R370,000 to R963,000 each.

Option 4: Over-wall pipe with a dry booster pump on a raft

Similar cost profile to Option 3. The booster pumps deliver around 2 bar at the required flow, which changes the operating point of every pump downstream. At the new development, that would have triggered a full re-selection of the main irrigation pumps.

Option 5: Over-wall pipe with booster pumps on the dam wall

The same as Option 4 but mounted on the dam wall rather than a raft, which makes them reachable with a bakkie. Same capital and energy cost, same downstream hydraulic impact.

Options 3 to 5 solve the suction problem by adding more pumps and more power draw. Option 1 solves it with civils. Option 2 solves it with engineering.

A siphon over a dam wall tends to get rejected quickly for two reasons. Both are manageable with the right design.

Negative pressure at the crest

The section of pipe that crosses the top of the dam wall runs under negative pressure. If air gets in at that point, the siphon loses prime. We addressed this by specifying continuously welded HDPE pipe throughout, routing the crossing only 200 mm above the water surface to keep the negative pressure as shallow as possible, and sloping the pipe at 1% toward the crest so that any air collects at a single point where a vent valve can release it rather than pocketing along the run. The system was pressure tested to 10 bar before commissioning.

Priming

Rather than adding a dedicated priming pump, we designed a small manually operated bypass around the non-return valve inside the pumpstation. Opening it allows water from the rising main to drain back into the suction line and prime the siphon. One valve, no motor, no control logic.

Suction pipes need to move water within a specific velocity range. Too fast and friction losses erode the available pressure at the pump inlet. Too slow and silt settles in the pipe.

We designed for velocities between 0.7 and 1.0 m/s.

At the first farm, two irrigation sets running simultaneously peak at around 450 m³/h. Three parallel 315 mm pipes carry that flow at 0.7 m/s, with a total head loss of 0.18 m over a 50 m run. There is spare capacity of 230 m³/h before the upper velocity limit is reached.

At the second farm, modelling the full infield system gave a combined drip and micro-irrigation peak of 291 m³/h. Two parallel 315 mm pipes run at 0.67 m/s with a head loss of 0.12 m over 30 m, and 140 m³/h of spare capacity.

Total suction head loss at both sites came in under 0.2 m.

The siphon itself is only as reliable as the components around it.

Intake screen

Standard foot valves were not suitable for the required flow and operating conditions. We specified a purpose-built stainless steel wedgewire screen sized to keep intake velocity below 0.1 m/s. Above that threshold, vegetation is pulled into the screen and head loss increases. The screen is self-cleaning and corrosion-resistant.

Non-return valve

Rather than use expensive Ø300 mm ball-type units, we specified wafer swing check valves that perform the same function at a fraction of the cost.

Floating raft

Specified entirely in HDPE. PVC becomes brittle under prolonged sun exposure. Stainless steel chain and shackles carry the screen load.

Air valve

A single-acting valve at the high point of the siphon, with brass fittings throughout to remove corrosion from the risk register.

Both pumpstations now have the suction capacity they need, without new booster pumps and without draining the dams.

The siphon design delivered the same hydraulic result as the more expensive options at a lower capital cost, with fewer components to maintain and no impact on the downstream pump selection.

Where a permanent outlet through the dam wall is eventually cut at either site, it can be done on a planned shutdown rather than as an emergency response to a failing system.

The trade-off is honest. A siphon depends on sound pipework and a reliable prime. The welding and the air valve have to be right. When they are, it is the simplest, lowest-maintenance answer available for this class of problem.

This approach is particularly relevant where a dam outlet is undersized, corroded or absent, and draining the dam or rebuilding the outlet is impractical. This covers a significant number of older farm storage dams across South Africa and the broader region.

If you have been quoted booster pumps or a new pumpstation to fix a suction problem, get an independent assessment before the order goes in.