This article is more than 1 year old

Mining for insight in the economy of things? Check your toolkit

Cars, conference rooms and tractors would like to play too, please

The Internet of Things (IoT) has given us watches that talk to our phones, phones that talk to our cars, and robot spies that talk to Amazon. But can it revolutionise industries with lots of unstructured information that hasn’t yet been collected?

The IoT concept is a bit like soap in the bath. You think you’ve got a grip on it, but then you squeeze a bit too hard, and – whoops – it’s gone. Conventional wisdom says that it’s a network of connected everyday devices that aren’t PCs. But we’ve had that for a long time, since Supervisory Control and Data Acquisition (SCADA) first evolved, in fact. So, what’s the difference?

One is an evolution of the other. The beginnings of the IoT came from hard, dirty industry – the sort of industry that you wouldn’t necessarily think would have conceived a global constellation of shiny, smart objects, all talking to each other and automatically ordering your washing powder for you. These are industries that pull hydrocarbons out of the ground, and move them vast distances under intense pressure. Industries that politely flush away your wastewater so that you don’t have to think about it anymore, and then do their best to sanitise it.

What’s old is new again

These industries have had machines talking to machines for years. Pumps, often. And generators, pressure regulators, and so forth. Software causes them to connect with each other, relay data, and often react to what each other is doing. In that sense, SCADA was a precursor to the emerging IoT.

“In the mid-1990s, several things converged,” explained Kevin Ashton, co-founder of the Auto-ID Center at MIT. “One was incredibly low cost, low powered electronics; another was big advances in radio communications; a third was the arrival of networked communications on a massive scale.”

This created an opportunity to make electronics work together on a far larger scale than it had before, he added. “So, for example, rather than limit control systems to something like a KUKA arm or a CNC lathe, operating in a highly regulated, standardised environment, you could have a system that was as ad hoc as the real world, with sensor data streaming into the network from multiple sources, and that this would be incredibly, profoundly valuable.”

Consequently, the IoT has moved full bore into all kinds of other applications, many of which are focused on service-based businesses dealing with information flow, and with instruments far closer to consumers than oil pipelines or sewage treatment plants.

These systems are often more modern than original SCADA equipment, which typically has long refresh cycles and may have been designed with rigid, preset communications models in mind. The idea of an IoT that adapts to fit an emerging, dynamically changing set of applications and communications models may be at odds with these original functions.

So what opportunities exist to drive the IoT into a space that may have rigid machine-to-machine communications – or none at all? And what kinds of industries might those be? What would this concept look like, and how quickly is it moving?

Bit Stew is one company that is trying to marry modern IoT technologies with legacy SCADA systems in utility companies and further upstream in energy production. The Canadian firm has focused on helping energy firms like British Columbia’s BC Hydro with their smart meter rollouts, and is now looking at moving into the oil and gas business.

“In the oil and gas market, the data off their SCADA systems is there today. It doesn’t require a refresh of control systems,” said VP of marketing Franco Castaldini. “What we do is unlock that data, put it into context with other sources and provide new value.”

One example is leak detection systems for pipelines. There’s no way for an operator to differentiate a SCADA alarm from a fault on the communications network that is being sent to them. “It all gets muddled up,” he said, which can lead operators to ignore or prematurely investigate the alarms.

Using analytics at the edge, rather than waiting for indeterminate signals to reach the centre, makes it easier to understand what’s going on in the network, he added.

Digitally monitoring physical assets

The data may already be there in some industries, and may simply need massaging. There are many industries that are relatively unstructured, though, with large amounts of information that is poorly collected, or not at all. These industries often have physical assets that are underused because we haven’t have very good ways of monitoring them and quantifying their usage.

Take farming, for example. Not traditionally at the forefront of technological change, this industry has a lot of headroom for efficiencies using sensors and modern, two-way control systems. Using field sensors to evaluate the moisture or chemical makeup of soil could help farmers to regulate their flow of fertilizer or their irrigation, for example. Having these sensors communicate with flow controllers in spraying and irrigation equipment could help to automate the process and drive efficiencies into the system, potentially increasing yields while minimising material costs.

In the manufacturing business, the use of IoT technologies could create opportunities to price assets in different ways. One of the most famous examples is Rolls Royce, which began charging for engine thrust by the amount of time used in a ‘power by the hour’ model. This changes the proposition value for customers, and enables them to drive efficiencies into their own operating budgets.

Earlier this year, Tata published a report exploring the use of IoT technologies among large companies. It found that industrial manufacturers using IoT gained 29 per cent in revenues during 2014, far exceeding the 12 other sectors that it surveyed. One of the key applications in this sector was the use of sensors to monitor the equipment that they sold to customers - 40 per cent of industrial manufacturers were doing this.

These IoT technologies are still expensive. The study found that products were far more likely to be outfitted with sensors tracking their performance when they were higher-priced. The average price of a tracked product was $948, but some of them were far pricier: 54 per cent of companies placed sensors in products costing $1m-$10m, vs 20 per cent for products costing $100-$500.

More about

More about

More about

TIP US OFF

Send us news


Other stories you might like