The MRT system's woes can be addressed with solid engineering practices.

Robustness is the engineer's way of measuring how resilient a system is to disturbances. At the most basic level, it starts with having safety margins built into the design.

For example, if you expect a maximum of 100 per cent load, a thing should be designed to handle 120 per cent of that load.

This will reduce the likelihood of failure.

The unexpected is bound to happen at some point, and then something will fail.

When that happens, we need redundancy - or a back-up - in the system.

Think about aircraft design, where every system has a back-up, such as having dual engines when a plane can fly on just one.

Similarly, if there are 10,000 components in the MRT system, there needs to be a failure analysis of every component, along with a system that will be able to maintain a regular schedule in spite of failures.

To create a truly fail-safe system, we also need to remove all single points of failure, so that the failure of one component does not end up affecting a larger part, or whole, of the system.

So the next time something fails, don't ask how to prevent the failure.

Instead, ask how a system can be designed to function normally in spite of failure. Because as we all know, it is not a matter of whether something will fail, it is just a matter of when.

Liang Jing (Dr)