The cool, salty waters off Tasmania’s south-east coast offer the perfect conditions for farming fish. And some of the toughest in which to set up an IoT network.
Out here on the on the estuary, says Sam Boyes, head of IT at Australia’s biggest salmon farming firm Tassal, “it’s not very friendly”.
Until recently, the health of Tassal’s 20,000 Atlantic salmon swimming in pens at around 19 sites was gauged with little more than “gut feel”, Boyes explains. But in this hyper-competitive industry, the intuitive stomachs of even the most experienced aquaculturalists are no longer a precise enough measure.
Improve the data insights and the company can not only avoid the kind of public-image disasters that have recently threatened the brand, but add “hundreds of thousands” of dollars to its bottom line.
But first Boyes must set up a reliable Internet of Things. “With electronics, in a saltwater environment, and out in the sun all day long,” he says. “We’ve had some challenges.”
Gut feel
Successful salmon farming requires striking the right balance between a variety of factors. Some things, such as ocean temperatures, can’t be controlled (when they get too high fish stop growing, or die in large numbers). Others things can be controlled, such as the amount of feed and the density of fish in a pen.
The aim is to create as much biomass as possible at the lowest cost, while ensuring the fish stay healthy and the operation remains sustainable. To achieve this, accurate metrics matter – something Tassal wasn’t getting from its existing system.
“It’s been pretty manual. We had access to information - but would you call it information?” says Boyes. “A lot of [the sensors] were older-style analog sensors. A lot of it was ad hoc and required a person to actually stand there and read the information off the display. They’d go out on a boat and throw a probe in the water.”
The company is now embarking on a project that will put sensors on each pen to monitor water temperature, tidal flow, salt and oxygen levels and cameras above and below the surface to record fish activity and feeding patterns. The data they generate will flow into livestock management and analytics systems so the business can better react to changes, and identify trends.
“Rather than relying on someone to monitor everything and press buttons and slow the system down or speed it up, it can be based on all these metrics coming in,” adds Boyes.
More accurate and timely data will mean the businesses biggest single cost – fish food – can be reduced.
“We're chasing 1 per cent of half a per cent. They have pretty big impacts on the bottom line across the board. A small percentage in saving in feed per annum translates into hundreds of thousands of dollars,” Boyes says.
Salty sea device
Boyes’ team is currently trialling a number of methods for their IoT-at-sea project, working with Tasmania-based Intuit Technologies. One of the hardest elements has been finding devices that can withstand the off-shore environment.
“Do we spend a lot of money on getting a device that is suitable for those environments? Or do we go with a cheaper style device and because the wear and tear will be there anyway, we just throw them out and replace them, and house them in protective enclosures?” he says.
While those devices are stress tested, the team is also tackling the problem of getting the data – some of it bandwidth-heavy video feed – back to dry land.
“In some cases we’ve got our own private wireless networks back to a shore base and then bringing [data] back over normal terrestrial style services. In other cases we’re using the mobile networks with 4g and 3g. In some cases it’s neither of those so we’re looking at bringing wireless communications back to another point, maybe our own tower,” Boyes explains.
In some cases data is being processed on the barge that distributes tonnes of feed to each of the pens, with only key measures being sent back.
Changing tide
While the data is presently sitting in an on-premise data-warehouse, Tassal is seeking to better integrate the feeds into its ERP and livestock management systems. The next step will be automation of the feeding machines.
“Rather than relying on someone to monitor everything and press buttons and slow the system down or speed it up, it can be based on all these metrics coming in,” Boyes says.
“We’ll be getting the livestock management system to do the analytics and tell us how we should be feeding, then actually give the instruction to the auto feed system.”
Those insights that are already available are going “right up the chain” from the workers out on barges to board level, Boyes says.
Although in the early-stages, Boyes expects the IoT project will lead to huge gains for the business.
“[The project] has a bottom line benefit from increased biomass, and anything we can do to understand better the health and well-being of the fish as well, improves things like survivability, reduces stress on the fish that improves growing, conditions within the fish,” says Boyes.
“It’s all up, up, up if we can find out all those little bits and pieces and pull them together.”