Intake Pump Cavitating After Demand Growth
A horizontal split-case intake pump at River Station is rated 1,400 gpm at 95 ft TDH, driven by a 50-hp, 460-V motor (62 A nameplate, 1.15 service factor). Site elevation is about 1,000 ft and summer water temperature is 15 °C (vapor pressure about 0.25 psi). Static suction lift is 9 ft through an 8-inch suction line that drops 4 ft of head at 1,400 gpm. A new subdivision has pushed station flow to about 1,750 gpm at 68 ft of head with the VFD at 100%. The motor now draws 68 A, bearing temperature is climbing, and the pump makes a gravel-in-the-casing noise that worsens as the wet well level drops. Yesterday a mechanic suggested simply raising the overload setting.
1. Calculate the available NPSH at the current operating point and explain what the noise means.
Read the worked answer
Convert everything to feet of head:
- Atmospheric pressure at ~1,000 ft elevation ≈ 14.2 psia ≈ 32.8 ft.
- Vapor pressure at 15 °C ≈ 0.25 psi ≈ 0.6 ft.
- Static suction lift = 9 ft (subtracted).
- Suction friction at 1,750 gpm scales roughly with Q^1.85: 4 ft × (1750/1400)^1.85 ≈ 4 × 1.51 ≈ 6.0 ft (subtracted).
NPSHa = 32.8 − 0.6 − 9 − 6.0 ≈ 17.2 ft.
Compare that to the pump's required NPSH. At 1,400 gpm the pump needs something like 12 ft, but NPSHr climbs steeply toward runout - call it 19–20 ft at 1,750 gpm. NPSHa is now below NPSHr. Cavitation is not a possibility; it is happening.
What the noise means: the gravel sound is vapor bubbles collapsing against the impeller. It is already damaging the impeller, wearing rings, bearings, and mechanical seal, and each hour of running makes the rebuild more expensive. Vibration and bearing temperature rising are the same failure in a different unit.
Fixes, in order of cost:
- Move the operating point back left - throttle the discharge or limit the VFD speed, which lowers flow and suction friction and raises NPSHa.
- Keep the wet well level up; the noise worsening as level drops is your confirmation that lift is the lever.
- Reduce suction friction: replace the 8-inch suction with 10-inch, shorten the run, clean the strainer and foot valve.
- Reduce lift physically - lower the pump or flood the suction.
Design margin should be NPSHa ≥ NPSHr + 3 to 5 ft at the worst-case (maximum flow, minimum wet well level, highest water temperature) condition, not at design point.
2. The motor is drawing 68 A against a 62 A nameplate with a 1.15 service factor. Is that acceptable, and what do you do about it?
Read the worked answer
No - not as a steady condition.
68 A is 110% of nameplate. It sits inside the 1.15 service factor (about 71 A), but a service factor is a margin for short-term or occasional conditions, not a continuous operating point. Running continuously into the service factor raises winding temperature, ages insulation faster, and gives you nowhere to go when summer water temperature or wet well level makes it worse.
Why it is happening: the pump has run out on its curve. It was selected for 1,400 gpm at 95 ft; it is now delivering 1,750 gpm at 68 ft because demand grew and the VFD is pinned at 100%. Brake horsepower rises with flow, so the motor is following the pump out past its design point. Cavitation also degrades efficiency, which adds amps with no useful work.
Do not raise the overload setting. That removes the only protection the motor has and moves the failure from the overload heater to the winding.
Options, in order of immediacy:
- Right now: throttle the discharge valve or set a speed/flow limit to bring flow back toward 1,400–1,500 gpm. You lose a little energy efficiency and you protect the pump and motor tonight.
- Short term: trim the impeller, which shifts the whole curve left and down, or reprogram the VFD with a current limit.
- Medium term: add a second pump running in parallel at reduced speed so each unit sits near its best efficiency point.
- Verify: that the overload is set per the motor nameplate and the actual supply voltage, and confirm phase balance and voltage at the motor terminals - a low-voltage condition will also push amps up.
Log amps, flow, suction and discharge pressure, vibration, and bearing temperature before and after every change so you can prove the fix worked.
3. What long-term changes and records would you recommend to keep this from recurring?
Read the worked answer
Engineering
- Redo the hydraulic profile for peak-day demand plus fire flow, then select the pump near its best efficiency point at that condition. A pump chosen on average-day demand will always sit at the wrong place on the curve during peak.
- Set the suction-side design at the worst case: maximum flow, minimum wet well level, highest water temperature. Build in 3–5 ft of NPSH margin at runout, not just at design point.
- Upsize the suction piping and shorten it. Suction friction is the cheapest NPSH you can buy.
Controls and protection
- Program a flow or current limit into the VFD so the pump cannot be driven past its allowable operating range by demand alone.
- Confirm the overload is correctly sized to nameplate, and add vibration and bearing temperature trending if you do not have it.
Records that protect you
- A written NPSH calculation with elevation, vapor pressure, lift, and friction at the actual operating flow, filed with the pump's O&M data and used to justify the capital request.
- Operating logs showing amps, flow, suction and discharge pressure, wet well level, vibration, and bearing temperature - before and after each change.
- Vibration and seal/bearing maintenance history. If a rebuild is needed, the trend record is what distinguishes wear from misoperation.
- A note in the O&M manual documenting the allowable operating range and the reason the discharge is currently throttled, so the next operator does not 'fix' it by opening the valve.