Commissioning – Allowing for an unlikely scenario

The task was to commission a 600 mm diameter and 500 metre long rising main process water line and associated supply tank, pump and control valves.

The main purpose of the commissioning test was to locate water leaks from the pipeline bolted flanges.

As this was first time commissioning a detailed “What If” Risk assessment was done.

One of the more unlikely safety risks was structural failure as the pipeline and supporting structure was being dynamically loaded and operated for the first time.

It was assumed that the system had been subject to a design HAZOP – but this was never determined prior to the commissioning test.

The supply pump was variable speed and was fed via an offset intake line and header tank connected to the pump intake.

The pump , air operated shut off valve and supply tank were supported off a steel frame – consisting of three steel columns attached to the header tank and bolted in place to the concrete foundation plinths.

The pump was started on low speed and water was pumped into the 500 metre long, 600 mm diameter rising main pipeline for several minutes.

As expected, several large water leaks were observed and noted for repair once the test had been completed.

One of the hazard controls was to prevent entry of personnel to the test area by barricading the area adjacent to and underneath the pipeline which was located above head height.

The other control was to barricade the area around the header tank and nearby pump and shut off valve at a distance of 15 metres from the structure.

After several minutes the test was completed and the pump shut down.

Then life became very interesting.

About 8 seconds later there was a tremendous noise with the header tank being spun on its axis with such force that nearly all the column base-plate nuts were ripped off their bolt threads.

No one was hurt.

What happened?

Once the pump stopped – the shut off knife gate valve was automatically activated and slammed shut about 8 seconds later.

During the 8 second delay, a reverse flow situation developed.

Water in the rising main was flowing backwards towards the header tank through the knife gate shut-off valve and pump.

When the valve slammed shut – it was estimated that about 50 tonnes of water was flowing backwards perhaps at a velocity > 2 metres per second to impact the knife gate valve.

The force was translated through the valve to the intake pipe work and through to the header tank which caused the header tank to violently rotate in the horizontal plane.

Key issues identified:

  • The pipework design did not include a non-return valve. This should have been determined in the HAZOP study.
  • The time delay in shutting the knife gate valve was excessively long – It had not been checked and set correctly for the commissioning test.