The PI System keeps Montreal’s mass-transit system on track

Société de Transport de Montréal (STM), the public company that manages Montreal’s subway and bus system, wanted to improve service for its customers but was stuck in old ways of managing information. When systems broke down, STM relied on alarms, followed by delays as repairs began. Also, with no centralized way to sort data, maintenance workers would have to run from one building to another to share information on equipment problems. It is the second most utilized transportation company in Canada and the third most utilized in North America. But despite the crucial services it provides to support over 416 million passenger trips a year, the company had limited visibility to its mass transit equipment, using SCADA displays restricted to the control room. Everything changed with the PI System, including PI Vision and Asset Framework—now, maintenance teams can improve customer experience in real time.

Useful data requires a culture change
STM needed to change its way of thinking about data. “The maintenance department had no historical data for fixing equipment such as escalators or elevators,” said Pascal Dubois, process-control engineer at STM. The company’s only indicators of failure were system alarms, which warned that something had already broken down, creating huge delays in maintenance response time, diagnostics, and reports.

It was difficult to make use of the data, and because security requirements demanded that alarm data stay in the command center, analysts had to be there to access the data. Not only that, but maintenance workers also had to be in the escalator mechanical rooms to access asset information.

“It was important to have real-time data from anywhere,” said Dubois. STM began its upgrades by adopting the PI System for its subway system in 2017 as part of the Optimization of Processes and Maintenance Activities (OPALE) project. The goal: improve maintenance on fixed assets through operations data.

Starting from the ground up
STM is now planning to expand its use of the PI System and wants to collect data for its tests of train-wheel integrity, which are assessed at every station. With new modern trains joining its fleet, STM is also looking forward to acquiring equipment with built-in data communication capabilities. “What we would like is for our employees at the stations to have the ability to troubleshoot problems or help customers using a tablet that is connected to the PI System,” said Dubois.

The PI System starts with STM’s data-production network, replicating and sending data from the company’s fixed equipment through a special one-way diode to the corporate network. Because data flows one way from assets to the network, it’s impossible to hack backward into the system. The PI System acts a single source of truth, allowing STM’s engineers, maintenance workers, and programmers to interact directly with critical, real-time information from anywhere.

“The whole [customer] experience is important,” emphasized Dubois, “so we started with the escalators.” The Montreal metro system has nearly 300 escalators, with different kinds of data sensors. The first step for Dubois’s team was to deploy Asset Framework (AF), the contextualization layer of the PI Server, to create a digital twin of the metro’s escalators and other fixed assets. AF transforms strings of data into easy-to-understand information by supplying important metadata, such as equipment location and status, in real time.

The team also adopted PI Vision to monitor operations with web-based displays that can be used anywhere. “If a user selects an escalator and goes to the details of the asset, there are about 40 attributes for that single escalator,” Dubois said. “This detail page is available to everyone but is intended for the maintenance team.” With PI Vision, the team can now prioritize maintenance to improve customer experience in real time. STM is now using its wealth of asset data to set up more effective alarms for breakdowns and anomalies and to create reports that allow engineers to identify the root cause of common escalator issues. “Where are we today after nine months with the PI System?” asked Dubois. Now, he said, the company now has access to data from more than 1,300 pieces of subway equipment, collecting over 50,000 data streams to remotely detect issues before they occur—a far cry from having to run to the next building to look at the data.

For more information about STM and the PI System, watch the full presentation here. (The video recording is in French with English subtitles.)