Bringing together sensor data from different suppliers in salmon farming
Sensor data from different suppliers is brought together by fetching it from each system where it already sits. Then it gets the same names, time and units, and each measurement is linked to the pen, tank or equipment it belongs to. Without that, oxygen, temperature, feeding and power each sit in their own system and cannot be compared. Sensa Unity connects the data sources, quality-checks the data and keeps the connections up to date over time.
By Trond Ove Høie, Co-Founder & VP Customer Success at Sensa.
- 1
Why the data is scattered
A facility usually has sensors, feeding systems, cameras and control systems from several suppliers. Each supplier stores the data in its own system, with its own names, its own units and its own clock. The same oxygen reading can have different names in two systems, and be logged every minute in one and every quarter of an hour in the other. Each system works on its own, but none of them can answer questions that cut across them.
- 2
The ways in
The data is fetched where it already exists, without replacing the equipment. From control systems and PLCs it often goes through standard protocols such as OPC UA and MQTT. From the suppliers’ own systems it goes through an API or a database. Public data, such as lice counts and sea temperature from BarentsWatch, is fetched the same way. Each way in is an integration, and it has to be monitored, because an integration that stops otherwise looks like a reading that has gone still.
- 3
Same names, time and units
Once the data is fetched, it has to be made comparable. Each reading gets a name by the same rule, the timestamps are put on the same clock, and the units are converted. Then each reading is linked to what it describes: the site, the pen or tank and the equipment. This is called metadata, and it is what lets a value be read without knowing the system it came from.
- 4
Check the data before it is used
Sensors drift, drop out and need calibrating. An oxygen sensor with fouling can read wrong, and a gap in the data can look like a quiet night. That is why the data is checked for gaps, frozen values and values outside what is possible before it is used in reports, forecasts or AI. An error found here costs less than a decision made on it.
- 5
What you can do once the data is together
Once the data is in one place and in context, you can see feeding, oxygen and temperature for the same pen in the same chart, compare sites and ask questions across the systems. In a RAS smolt facility, detailed feeding data revealed an old rule in the control system that fed unnecessarily just before midnight, time after time. Unity AI builds on the same data, so the answers come from figures you can trust.
Frequently asked questions
How can a fish farm bring sensor data from different suppliers into one place?
By fetching the data from each system where it sits, through standard protocols such as OPC UA and MQTT, an API or a database, and then giving it the same names, time and units and linking it to the pen or tank it belongs to. The data also has to be checked for gaps and errors before it is used. Sensa Unity does this for the data from the suppliers you already have.
Do we have to replace sensors or suppliers to bring the data together?
No. The data is fetched from the systems you have. What has to be in place is access to the data at each supplier and a shared way of naming it and linking it to the facility.
Book a demo Would you like to see the data from all your suppliers in one place?