Forum: Deploy tech to develop a more accurate rail reliability model

I disagree with the points in the Forum letter, “Don’t underplay maintenance at the cost of severe rail disruptions” (June 30), and agree with SMRT’s position that there should not be “overmaintenance” with respect to rail reliability (‘We don’t want overmaintenance’: SMRT chairman flags need to balance rail reliability with costs, June 19).

The phenomenon of overmaintenance refers to the need for increased manpower, increased train hardware and spare parts. 

While this would result in fewer train breakdowns and malfunctions, it would invariably mean higher costs that are eventually borne by the public.

Thus, a better and more sustainable approach is for SMRT to deploy technology to develop a more accurate railway reliability model to take preventive actions even before the trains malfunction.  

SMRT should use real-time data, fuzzy logic and artificial intelligence to accurately assess future maintenance decisions, and take pre-emptive measures to minimise the number of train malfunctions.

The writer uses the threat of sea-level rise as an analogy for rail disruptions. But rising sea levels is a known risk that is predictable as it occurs slowly, and pre-emptive measures can be deployed in a calibrated approach. 

On the other hand, railway reliability is not an exact science because there are multiple input variables (for example, passenger travel patterns, railway technology, materials and system usage, and extreme weather), and therein lies the train reliability challenge.

Lee Kek Chin