Rising Wet Well After a Storm: Pump Problem, Infiltration, or Something Else?
A duplex submersible lift station serves 4,200 homes. Design average flow is 55 L/s, and each constant-speed pump is rated 80 L/s at 22 m TDH on a 400 mm force main that runs 2.4 km to the plant. In dry weather the station cycles about 6 minutes on, 9 minutes off. After 25 mm of rain in three hours, both pumps run continuously and the wet well - 3.0 m diameter, 4.0 m deep, high-level alarm at 3.2 m, emergency overflow at 4.0 m draining toward a trout stream 600 m away - is rising 0.3 m per hour. Station meters show each pump drawing roughly 15% below nameplate amps, and discharge pressure reads 16 m instead of the usual 21 m. The pump controllers were replaced last month.
1. The pumps are running but clearly not delivering their rated flow. Before you call this infiltration and inflow, what explanations fit the combination of low amps and low discharge head, and how would you check each one?
Read the worked answer
Read the combination, not the individual readings. On a normal constant-speed centrifugal curve, a lower head means higher flow and therefore higher amps. Here head is down 5 m and amps are down 15%. That is contradictory for a healthy pump and points at a loss of hydraulic capability rather than a change in system demand.
Candidate causes and checks:
- Reverse rotation. The controllers were replaced last month. Swapping any two phase leads on a three-phase motor reverses rotation; a centrifugal impeller rotating backward still pumps, but at reduced head and flow with lower amps. Check: bump the pump and watch the rotation arrow / check phase sequence at the starter. This is the highest-value check given the recent work.
- Ragged or partially blocked impeller. A rag or wipes ball reduces head and the absorbed power. Check: pull the pump and inspect, or compare a drawdown test against the factory curve.
- Air entrainment / vortexing. If the pumps are pulling the well down near the inlet, air binding lowers head, flow, and amps. Check: observe the wet well at the low end of the cycle; look for a vortex and for air in the discharge.
- Worn wear rings or impeller clearance. Gradual, but would not appear overnight. Check: compare the last several months of run-time trend and drawdown results.
- Suction-side restriction or a partially closed valve. Check: valve position, suction gauge.
- Current unbalance or a failing phase. Low amps on one leg only. Check: amp reading per phase, not a single averaged reading.
The decisive test is a drawdown (pump-capacity) test: with inflow known or isolated, time the level drop with one pump running and convert to L/s. That gives you an actual operating point to compare with the 22 m / 80 L/s curve and separates a pump problem from a flow problem in one step.
2. Assume the well is at 2.6 m when you arrive and still rising at 0.3 m/h. How long do you have before the emergency overflow discharges, and what is your order of operations?
Read the worked answer
Time to overflow. Wet well cross-sectional area = π × (1.5 m)² = 7.07 m².
Remaining rise to the 4.0 m overflow = 4.0 − 2.6 = 1.4 m.
Storage = 1.4 m × 7.07 m² = 9.9 m³.
Rise rate 0.3 m/h → 0.3 × 7.07 = 2.12 m³/h, so time ≈ 9.9 / 2.12 ≈ 4.7 hours. Treat that as an upper bound: the rise rate is not linear and will accelerate as more of the collection system fills.
Order of operations:
- Notify now, not at the point of spill - supervisor, collection crew, and the regulator's spill line per your SSO procedure, because a discharge is plausible within the shift.
- Mobilize a portable bypass pump and a vacuum truck; stage absorbent boom and sandbags at the overflow outfall and the ditch leading to the stream.
- Cut inflow if you safely can - plug or divert at upstream manholes, and check whether a crew can isolate a contributing area. Do not block flow in a way that pushes the overflow further upstream.
- Conserve the wells - stagger pump starts is not an option here (both must run), but stop any non-essential use and confirm the pumps are not air-binding or tripping.
- Pre-notify the fishery/recreation interest and public health if any discharge reaches the ditch.
- Log everything with times - levels, actions, volumes estimated - because the report and any enforcement discretion depends on it.
The judgement point: 4.7 hours is short enough that the first action is mobilizing equipment, not troubleshooting the pumps. Diagnose in parallel with a second person.
3. What data would you pull to decide whether the excess flow is infiltration/inflow, a downstream blockage, or a mechanical failure - and how do you tell them apart?
Read the worked answer
Pull three independent lines of evidence and see where they agree.
1. Hydraulics at the station. Pump run-time and wet-well level trends versus rainfall for the last 6 months; the force main pressure at both ends. A force main break or a major leak downstream shows as low discharge pressure with normal or high pump output and, critically, as flow arriving at the plant that does not match station throughput. Compare station totalized flow with the plant's influent meter for the same hours - a gap means water is being lost between them.
2. Flow response shape. I/I tracks rainfall with a short lag and peaks then recedes; a blockage or a collapsed pipe shows a step change unrelated to rain and often a rising level in a specific upstream area. Plot wet-weather flow against rainfall depth and against groundwater level (a groundwater-driven response tails for days or weeks after the storm).
3. Source investigation. CCTV and smoke testing of the suspect laterals, dye testing of foundation drains and yard drains, and manhole inspection for inflow from catch basins and roof leaders. Compare dry-weather night flow per kilometre of sewer to flag which sub-basin is contributing.
Rule out the mechanical first because it is cheap and fast: drawdown test, rotation check, amp-per-phase reading, valve position. If the pumps are not delivering their curve, the flow numbers mean nothing until that is fixed. Once the pumps are verified, the rainfall-versus-flow relationship and the plant-meter comparison sort I/I from a downstream loss.