Tasmanian Atlantic salmon has long been held in high regard in culinary circles around the world, and for one major farmer, the introduction of Internet of Things technologies to its operations is set to bring greater efficiencies and new data insights.

From its humble beginnings in 1986, Tassal has now moved into the world of IoT together with its longstanding digital partner and fellow Tasmanian business, Intuit Technologies.

For Sam Boyes, Tassal’s ICT senior manager, the decision to use IoT was borne from a project initiated four years ago, whose objective was to get more accurate information from all parts of the business to the company’s stakeholders in a more timely manner.

“Most of our processing facilities had access to equipment, sensors, and so forth that we’ve had in place for a long time, but the marine farm side of our business really lacked in any sort of technical sophistication, so we placed our focus on that part of it,” he told IoT Hub.

“There are obviously a lot of challenges being in a marine environment. In most cases, the farms are fairly remote, which introduces challenges around power and connectivity.

“Also, these farms were very manual, relying on a lot of field decisions made by people out on the water, and not really being able to provide any sort of meaningful trends or analytics with the information gathered over time.”

What was the scale of the problem?

Tassal and Intuit wanted to devise a strategy to modernise their feed operations across the 18 lease areas the salmon producer utilises, with each area containing on average 20 salmon holding pens.

Each pen holds approximately 20,000 harvest-sized salmon, and each lease area contains a single barge that holds multiple tonnes of feed for the salmon that mechanically feeds the fish contained in those pens.

All of the data collected in relation to salmon and feed counts, weight checks and other metrics would then be manually input into the company’s livestock management system.

“Also, although we were feeding the fish mechanically, we weren’t using any sort of insight from the environment around us; things were instead done via ‘gut feel’ by the guys in charge of feeding the fish,” Boyes added.

“At times, you might have different water temperatures, currents, weather patterns coming through that affect those operations, hindering our ability to identify how we could best deal with those situations, and whether or not there were better times of day that we could feed the fish and get better outcomes, for example.”

What does the solution look like?

Tassal already had a number of analogue sensors that it had been using, so the first step was to replace them with IP-based sensor platforms.

“We focused on deploying some key environmental sensors. The majority of the sensors we deployed are in-water, measuring things like water temperature, tidal flow, salinity, dissolved oxygen, and so on,” Boyes said.

“These are the sorts of metrics that have potential impacts on not only the health and well-being of the fish, but also tells the story on how well they’re feeding, or if the environment is or isn’t conducive to getting the best outcome from feed events.”

Tassal is also in the process of deploying a video capture system which provides information of the behaviour of the fish and feed supplied to them, both above the water and below.

“One of the plans we have for that video system is to look into providing video analytics over the top to identify the feed pellets in the water, and monitoring fish behaviour during feed times,” he said.

These sensors and video feeds then send the data back to Tassal’s core infrastructure in a number of ways, including wireless back-to-shore networks, and 3G/4G connectivity.

Early hesitation turned to excitement

To date, the IoT system has been deployed to approximately one quarter of Tassal’s salmon farms.

Boyes said that while the senior management were excited at the prospect of new data streams being provided, which gave them more information to make better business-wide decisions, the farmers themselves at the operational layer were hesitant, at first.

“Initially, you had some that were about nervous about what the new technology meant, with a belief that ‘a computer can’t tell me how the fish are reacting’,” he explained.

“However, I think we’ve found that if most of them weren’t on board at the start or excited by it, they certainly are now.

“We’ve almost got a challenge now of being able to roll it out quick enough now, because everybody is wanting the get their hands on the technology to do their jobs more efficiently and accurately.”

What’s next?

Boyes is hopeful the remainder of Tassal’s farms will receive their IoT deployments over the next three years.

In the meantime, a more pressing challenge for the project is to improve communications to the remote locations to ensure uninterrupted transmission of the exponentially greater amounts of data.

“We’re already looking at ways we can improve comms, whether we create bigger communication backhaul that we invest in ourselves, or partner with someone to do that and pay for it over an amortised timeline,” he said.

“We could also do something a little bit different there and do some edge processing on-site and then send back only the critical data back to our central infrastructure.”

Boyes said that while existing data transmissions used are coping with the added load, he’s finding that the upper limits of these networks are being reached.

“The next iteration [for the project] is strengthening the backhaul along with some data cleansing to only bring back key data points, rather than trying to bring back every single data point.”

Boyes is also hopeful that the data that Tassal collects from its farms would be useful in a research capacity, combining with other data sets to help organisations like the CSIRO further their research.

“The next phase of our deployment could be specifically around trying to overlay the data we’ve got on our farm sites and blending that with other sources of data like CSIRO environment data,” he said.

“That could then start to uncover patterns and stories that can give us more direction and predictable outcomes.”