EngiMetric Arena / Diagnose the Failure
Why does the pipe bang when the valve closes slowly?
Competency Fault diagnosis primary Quantitative reasoning supporting
What you should take away: Recognise that a slow closure can still produce a significant surge, and distinguish water hammer from cavitation or a mechanical fault.
The situation
What is happening
A 180 m pumping main runs at 2.5 m/s. The pressure-wave speed is 1 200 m/s, so the round-trip time is 0.3 s. The valve takes 1.2 s to close, but operators report a heavy bang and a pressure spike of about 7.5 bar after closure. Flow and pressure are normal before the valve moves. What is the plausible diagnosis?
Given
| Quantity | Value | Unit | Note |
|---|---|---|---|
| Pipe length | 180 | m | Length the pressure wave travels. |
| Initial velocity | 2.5 | m/s | Flow before closure. |
| Wave speed | 1 200 | m/s | Material-dependent input. |
| Closure time | 1.2 | s | Slower than the 0.3 s round trip. |
| Observed surge | 7.5 | bar | Pressure spike after closure. |
Assumptions this challenge makes
- The pipe is a single straight run with a rigid-conduit surge response.
- The wave speed and density are stated inputs, not looked-up material defaults.
- The observed 7.5 bar is treated as the peak rise, not the absolute pressure.
Your call
Your answer
Hints
- Compare the closure time with 2L/a. Which regime does the engine say applies?
- The slow-closure peak is not zero. Compute 2 rho L dV / t and compare it with the observed spike.
- Ask whether the symptom follows the valve action or the pump suction. That separates surge from cavitation.
Your answer Engine result
Where that comes from
Each limit, checked
Why
- The critical closure time is 2 x 180 / 1 200 = 0.3 s. The actual closure is 1.2 s, so the engine selects the slow-closure regime.
- The Joukowsky ceiling would be 1 000 x 1 200 x 2.5 / 100 000 = 30 bar, but that is not the applicable peak for this closure.
- The slow-closure surge is 2 x 1 000 x 180 x 2.5 / 1.2 / 100 000 = 7.5 bar, which matches the observation.
- The diagnosis is therefore water hammer. The closure is slow enough to avoid the full Joukowsky peak, but not slow enough to make the surge negligible.
- A surge vessel, a slower closure, a bypass or a different valve characteristic could reduce the peak. The calculation identifies the mechanism; it does not select the remedy.
What this challenge does not establish
- The model is a single straight run with a rigid conduit; column separation and air pockets are not modelled.
- The wave speed is an input and must be verified for the actual pipe material and restraint.
- The observed 7.5 bar is treated as the peak rise; the absolute pressure and pipe rating still need checking.
- The calculation does not prove that surge is the only problem; it shows that surge explains the measured event.