SMRT to use data system to predict potential defects, improve rail reliability

SMRT to use data system to predict potential defects, improve rail reliability

Train operator SMRT is testing a range of new systems that utilise data and artificial intelligence to improve rail reliability and better predict waiting time for commuters.

SINGAPORE: Train operator SMRT is testing a range of new systems that utilise data and artificial intelligence to improve rail reliability and better predict waiting time for commuters.

It revealed details of the new technologies it plans to roll out by the end of 2018 at a media briefing at Kim Chuan Depot on Friday (Mar 16).

Among them are sensors that can better spot track defects, a digital platform for maintenance updates and a system that can help predict defects by analysing data collected by the devices.

The soon-to-be-implemented Train and Track Condition Monitoring System (TTCM) uses fibre-optic sensors to detect abnormal vibrations in real time. Such vibrations commonly arise due to either track defects or train wheel degradation.

“We have attached fibre-optic sensors to the axles of passenger trains, so that they can simultaneously detect track degradation as they carry out everyday operation,” said SMRT future systems engineer Ang Zi Keng.

“There is also a separate set of trackside fibre-optic sensors that can corroborate vibration data so that we can locate and identify the respective faults for repair.”

Compared to previous sensors that were less efficient, the new sensors are less prone to human error.

Engineer Dr Tan Chee Keong said: "We also relied on engineers to visually spot track degradation, previously. With the new system, track degradation issues can definitely be detected better.”

The TTCM is currently in the validation phase, where its effectiveness is being compared with current methods. It is expected to be in full operation by end-2018, according to Dr Tan.

Infographic - SMRT track defects sensor

When deterioration is detected, workers will be able to rely on the Track Access Management System (TAMS), which uses artificial intelligence to prioritise the booking of track time for repair works during maintenance hours.

SMRT will also be implementing a paperless Integrated Maintenance Operating System (iMOS) which will let maintenance staff update their works-in-progress with a tablet computer.

PREDICTING ISSUES THROUGH DATA

All the data collected from the various systems will then be fed into the Predictive Decision Support System (PDSS), which will then prioritise certain maintenance activities ahead of others and predict potential issues based on the trends these data suggest.

However, the system will only work accurately when other related data systems are up and running.

“We are currently validating the accuracy of our data trends. Accuracy is key,” said SMRT engineer Ms Sian Pei, who is working on the PDSS. “All the other systems will first need to have an accuracy of about 80 per cent. Only then can the PDSS better predict what it needs via data correlation.”

She added that when operational, the PDSS will double the efficiency of maintenance works.

“We are working on the beta version now and we target to roll this out by this year,” said the SMRT future systems engineer.

The adoption of the various new innovations aims to reduce train delays arising from inefficient practices in the past.

“These technologies will enable our people to work smarter, more productively and effectively,” said SMRT CEO Desmond Kuek. “Through such innovations in our people, systems and technologies, we ultimately aim to deliver safer, more reliable and comfortable journeys for our commuters.”

MORE ACCURATE WAITING TIMES

Also by the end of this year, SMRT is planning to roll out the Advanced Crowd Sensing System (ACSS) that can monitor how crowded train platforms are and tell commuters the waiting time for the train based on where they stand in the queue.

SMRT is currently testing the ACSS, which integrates multiple modes of crowd detection – including through closed-circuit video feeds and fare gate data - to inform passengers of more accurate waiting times using its recently released SMRT Connect app.

It also uses Wi-Fi signals emitted from commuters’ smartphones to triangulate their positions.

SMRT says it is confident that the new technological upgrades will lead to better passenger experience and improved rail reliability. However, the rail operator declined to reveal how much the new systems would cost.

“We evaluated and assessed certain technologies that would deliver reliable systems, which would then lead to enhanced commuter comfort. Only then did SMRT made a decision to invest in (such) technologies,” said SMRT Corporate’s chief technology officer Ng Bor Kiat.

Bookmark