The world is full of dumb machines, and Brad Parsons wants to drag them whirring, gurgling and buzzing into the Internet of Things era.
Parsons is the founder and CEO of Brisbane company MOVUS, which sells a pay-as-you-go service that he describes as being somewhat akin to a Fitbit for machines: A device that can be easily attached to a piece of non-instrumented industrial equipment and with the aid of some AI smarts give its owner an indication of when things are starting to go south.
Before founding MOVUS, Parsons worked in a range of heavy asset industries such as mining and freight rail. His focus was on strategic projects that blended the worlds of IT and OT, such as mine-site automation and condition monitoring programs.
Parsons found that although a lot of big, heavy industrial assets are instrumented, there’s a large layer of assets — so-called B-class assets — that aren’t.
“It’s the motors and the pumps and the fans – all the stuff that really drives industry but doesn’t get the attention of the big machines,” Parsons told Computerworld. “Everyone wants to see a driverless haul truck with 400 tonnes of dirt in the back,” the CEO said. But people are less excited about a sewage pump, until they're knee-deep trying to fix a failed one.
“There’s hundreds and hundreds of millions of those types of assets. There are about 2.3 billion electric motors on the planet; when you look at the number of humans, it’s about three humans per machine,” Parsons said.
Of those machines, only around 3 per cent are instrumented. The CEO said the key challenge that MOVUS set out to solve is how to instrument the other 97 per cent, allowing their operators to prevent breakdowns or send an employee hundreds of kilometres to service an asset only to have to return a month later.
MOVUS’s answer is a device dubbed the FitMachine, which has a magnet that allows it to be attached to equipment, and some 17,000 lines of on-board code help it monitor vibration, noise and temperature.
Parsons said instead of businesses needing to purchase a sensor, analytics platform, and a gateway and then connect them all together, MOVUS bundles it all together in a single platform centred on the data derived from the FitMachine.
An iOS or Android mobile app is used to register a FitMachine on MOVUS’s platform, which runs on Amazon Web Services. The FitMachine can transmit data either over Wi-Fi if it’s available or a MOVUS-built rugged gateway device designed to survive harsh industrial environments.
In addition to standard cellular connections, the current version of the gateway can leverage Telstra’s Cat M1 network, which is designed for IoT applications (Telstra is one of MOVUS’s investors). Cat M1 can “get into deeper, darker holes” than standard LTE connections, Parsons said.
At the heart of the platform is an AI engine that takes in a continuous stream of data from a registered sensor.
“Once [a sensor is registered], it’s literally a week to two weeks before it’s built a baseline of what’s normal for the machine,” Parsons said. “That normal – what the machine is listening to – is the vibrations, the temperature, and the noise of the machine.”
Once that baseline is established for a particular FitMachine, the AI engine can then issue an alert if a significant anomaly is detected.
“It gives you peace and mind that the thing’s working,” Parsons said.
That means instead of sending out a human once a month, or a quarter, or a year, to spend five to 10 minutes checking that a machine is operating normally, and running the risk of a failure occurring in between checkups, someone can be dispatched when there are signs of impending failure.
MOVUS provides customers with a web-based dashboard that can be accessed from a mobile device or PC. However, its platform is also API-enabled to allow larger businesses to integrate the system into their enterprise applications.
For example a business can link it to an asset management system or feed it into a workforce management system to automatically trigger a maintenance process when necessary, Parsons said.
Because the system is simple (even if the underlying technology isn't), MOVUS’s platform has applicability to “such a diversity” of industries, the CEO said.
“Every week we get a new use case,” he added. Rather than focusing on a particular industry, MOVUS is “focusing on the machines”. It has customers from two dozen or so different industries across Australia, New Zealand and the South Pacific, and has had sales enquiries from around the world — even as far as Iceland.
“It doesn’t really matter where they are on the planet because if they’ve got a Wi-Fi connection then they can just connect the sensors to that and that’s the beauty of the system,” Parsons said.
MOVUS was founded in 2015 and at the start of 2016 launched what Parsons describes as an MVP. The startup spent 2016 honing the FitMachine ahead of its first commercial sale in January 2017.
Development of the device and the software platform is ongoing. The current version is designed for fixed equipment such as fans, motors and pumps, but it makes sense to build a version that can be fitted to mobile assets such as tractors, excavators and bulldozers, Parsons said.
The sixth generation FitMachine is currently in the field, with the seventh expected to launch soon and generation eight currently being designed.
Much of the development has focused on incorporating more sophisticated sensors into the device as well as improving battery life (currently a FitMachine’s battery can last for a couple of years, but some customers are seeking even longer-lived devices).
The company has subscription model with a per-sensor charge, with customers getting access to all software updates as well as hardware updates.
Customer needs and market research are driving development, the CEO said and MOVUS is “constantly innovating”.
“A customer comes in says ‘we want this’ we bundle it up into a release and then we basically push out code every two weeks,” Parsons said.