BHP cuts costs - derails Train

It's easy to take costs out of a business. How do you know that you have considered all the factors and not increasing your risk profile beyond the measured savings?

We recently heard that BHP was forced to derail a run-away train in Western Australia. The train was almost 3km long with its four locomotives and 268 wagons and fully laden with iron-ore train. The train travelled 92km from 4:40am to 5:30am when it was remotely derailed from the remote operations centre in Perth.

Whilst there will be many investigations into the runaway train, it raises many questions:

• Why was there only one person on the train?
• Why isn't there a remote kill switch?
• Why isn't there a 'dead man' switch on the train?
• Why couldn't the train be intercepted?

Whilst there is a procedure to get permission to leave the cab, and set the brake - the investigation will ascertain this. I'm not going to question the procedures, I'm questioning the options available before (and to prevent) the derailment of the train.

Let's look at the impact of the derailment:

• 1.5 kilometres of track was damaged, and the recovery process would take about one week.
• damage to four locomotives and up to 268 wagons
• impact on contractual obligations to supply customers.

Any way you look at it - it's millions of dollars in damages and impact.

One Person on the Train

Someone determined that it was Ok to have one driver on the train. One person, to control the entire train. One person responsible for millions of dollars of rolling stock and product. If something would happen to that one person - alone in the middle of nowhere, in the middle of the night disaster could strike. And disaster struck. Is it really worth only having one person on the train? A second person could have easily prevented this catastrophe. Is the wages of a second engineer / driver worth this? The answer is yes.

Remote kill switch

BHP touts about their AI achievements (rightly so), but before AI and driver-less trains it would seem some simple steps need to be put in place. Sure, data communications can be challenging - but not impossible (remember we have a multi-million dollar train). A remote kill operation would have simplified this operation.

'Dead man' switch on the train

A 'dead man' switch is designed to be activated in the event of that the human operator becomes incapacitated, or in some event is unable to render control of the vehicle. A simple mechanism is to operate a lever or push a button at a pre-determined interval (15 minutes).

In the event the button isn't pressed, the train would come to an orderly stop, preventing this catastrophe.

Could the train be intercepted?

Maybe I've watched too many action movies. The train was travelling at 100kph, and yes and it was night, but would it have been possible to intercept the train and place someone on board to stop the train. In this example, maybe this wasn't an option.

Summary

This was a disaster. It will cost BHP millions of dollars in remediation, loss of confidence and inconvenience. Maybe it could not have been avoided, or at least the high impact mitigated.

When we look at projects and operations, we consider the impact of a disaster (no matter how unlikely), we consider how to prevent and how to minimise the impact.

In this case - there were a number of potential options that would have prevented this catastrophe. What have I missed? What other options should have been considered that would have prevented this train derailment?