Can ProRail grind turnouts even better?

The key question is how often a turnout should undergo preventive grinding, referred to by ProRail as cyclical grinding because it forms part of an annual planning cycle. After all, the more material is removed from a rail, the shorter its service life becomes. Another question is whether all turnouts should be ground. ProRail formalised its approach to these issues in a grinding policy introduced in 2017.

Shevtsov explains:

"We inspect all centrally controlled turnouts at least once a year using ultrasonic testing. Of the 6,000 turnouts across the network, 4,500 are centrally controlled, meaning they form part of the operational timetable and are therefore used frequently. The most heavily loaded turnouts are inspected more often.

Of these 4,500 centrally controlled turnouts, 160 are included in our preventive grinding programme, typically on an annual or biennial basis. These are turnouts with a crossing angle ratio of 1:18 or higher, which are generally more important to rail operations than turnouts with lower ratios. Components for high-angle turnouts are also expensive, and replacing a turnout component takes longer than the standard four-hour overnight possession window.

For both economic reasons and to minimise disruption to passengers, it is therefore important to prevent defects in these turnouts wherever possible. Turnouts with lower angle ratios, by contrast, can be repaired much more quickly and at a lower cost, meaning preventive grinding is usually not cost-effective for them."

However, the prediction was not the most important outcome of the model at this stage, explains data scientist Rutger: 'We primarily wanted to know which parameters contribute significantly to the likelihood of defects and what relationship this has with whether or not a switch is ground. In addition, together with ProRail, we performed a cost calculation for two types of switches to determine the impact of grinding on the total cost structure surrounding a switch.' CQM started the project in the spring of 2024; in October, Minou and Rutger delivered their evaluation report with recommendations. Minou: 'The main question was: is the policy of preventively grinding 160 switches with a high angle ratio still the right choice? It turns out it still is. These switches are indeed prone to defects, and grinding helps to prevent defects. Moreover, the cost calculation shows that preventively grinding these switches is cost-efficient. However, the latter does not apply to the switches with a lower angle ratio.' While preventative grinding of these switches with the grinding train helps to prevent defects, it does not lead to lower costs. In short: the current grinding policy is therefore still adequate, even according to our latest model. This may not be the most spectacular conclusion, but both we and ProRail were pleased with it. We therefore made the right choices at the time. Moreover, we can now substantiate it even better and with more data.

Minou and Rutger are currently working on a smart, balanced pilot design that will allow ProRail to conduct targeted experiments with the recommendations over the coming years. The projects for ProRail are once again an example of the type of work that CQM usually does behind the scenes, but which has a major impact on society. Minou: 'When you are just sitting on the train, you have no idea what is needed to keep that train running. The fact that the experts at ProRail were willing to share their vast knowledge with Rutger and me has been incredibly valuable for this project. Riding along on the grinding train and simply seeing a switch up close is also of great value.' Shevtsov: 'CQM mastered the complex issues very quickly. The data analysis was also done very quickly and with dedication. Truly a top team.'