Text: Steve Roman
When businesses look for top-end office rentals in Stockholm, Sweden, they turn to Humlegården Fastigheter, the go-to property owner for space in some of the city’s most exclusive buildings. Humlegården’s clients expect only the best. Needless to say, elevator hassles are not part of the equation.
“This is one of our biggest issues,” says Per Rosén, the company’s head of operations and maintenance. “When an elevator isn’t operating properly, our tenants get really annoyed and it’s a big problem for us. If we have a high-end building and we only have one elevator and that elevator breaks down, you can imagine the difficulty,” he says.
In early 2016, Humlegården approached us with a simple but challenging request, recalls Jaakko Kaivonen, managing director, KONE Scandinavia. “The brief from Humlegården’s side was actually rather clear: ‘We don’t want to have any call-outs or breakdowns ever again. We want to use intelligent technology to predict issues before they happen.”
Using smart technology
It was certainly a tall order, but as it happened, we were in the early stages of developing what would become 24/7 Connected Services, a system that uses the IBM Watson IoT platform to process data from elevator monitoring devices in the cloud, analyzing and predicting equipment problems before they occur. Humlegården, for its part, had already been one of our key clients for over a decade and was a keen adopter of new technologies. Those factors led to what would be one of the first pilots of 24/7 Connected Services, with Humlegården and IBM working together with us to develop this powerful new application for IoT.
Traditional elevator maintenance is calendar based, with various components serviced or parts replaced based on estimated times of wear and tear. Its main drawback is that in the real world, elevators don’t always behave according to estimates. This new approach, by contrast, collects data from these units 24/7 through remote-monitoring devices, analyzes the data to understand the needs of each individual unit and predicts when a problem is approaching to allow all maintenance to be carried out in a pre-planned fashion with minimal disruption.
For the Humlegården project, we began by connecting sensors to dozens of elevators, in each case measuring about 200 points of data, which are sent to the 24/7 Connected Services cloud in real time. It was at this stage, Kaivonen says, the real challenge came. Namely, finding useful patterns in the immense volumes of data being gathered.
IoT to the rescue
“Everyone talks about IoT – it’s a big buzzword. You connect things to a cloud. I’m not sure if that’s exciting. You just get a huge amount of data. When it really gets exciting is when you start to put some intelligence on it, and this is where IBM Watson comes into the picture,” he notes.
IBM’s AI-driven analytics was the key to finding the useful, complex patterns within the data stream, Kaivonen says. For instance, if component A is showing certain vibrations while the temperature rises 0.5 degrees in component B, it is likely that component C will break in about a week. These types of conclusions have allowed us to draw sets of rules on how respond to the IoT data and generate maintenance requests when needed.
The experiences and analyses developed here will naturally be applied to other Connected Services projects as well, and used for the benefit of customers.
Rosén sees these types of self-learning systems as the future of the property management business, particularly in Sweden, where environmental concerns are a priority and thus saving energy by extending the lifetime of installed equipment is a well-understood goal.
Markus Huuskonen, KONE’s head of maintenance processes, notes that the 24/7 Connected Services will become better as more users connect. “We are improving all the time, so the more data we get, the better our service will be. The analytics engine is self-learning, so these kinds of connected services will definitely come into bigger play in the future,” Huuskonen saidd.
For now, though, the Humlegården project is creating the biggest waves among other property managers in Sweden who are intrigued by the prospect of a predictive maintenance scheme. “This idea has actually received quite a lot of interest from our customers,” says Kaivonen. “If you think about it fundamentally, instead of us just selling services, we’re selling safety, availability and uptime. We’re selling people flow. And I think that’s an interesting vision for many.”