This incredible journey began in the Welsh Agri Hack, run by M-SParc. After a fantastic pitch, they went on to win funding to help get the idea off the ground! Without the money from the Hac, Dewin.Dwr would not have been able to progress further. They were also able to secure funding to hire a new member of staff!
In addition to the Hack funding, Geraint received business support from M-SParc through the first cohort of our Level Up accelerator to help them further develop their business plan and strengthen their ideas.
The funding was given to develop a wireless sensor that would detect, monitor, and alert any managers (in this case normally farmers) on the evidence or potential for water pollution. The device can support farmers to make better decisions on when and how much slurry and fertiliser to apply.
Projects like this require expertise and knowledge and take time, therefore they had a staff member fully funded by M-SParc through the Skills Academy. Dafydd Pierce-Evans is an Electronic Engineering graduate and has supported the work of the project. He has produced the embedded systems software which will run the heart of the device to keep it beating!
As a result of the work done with Dewin.Dwr. Dafydd has successfully secured a full-time job with a local company called Aperito where he is assisting with developing wearable, patient lead medical devices.
One of the most significant developments has been the adoption of Machine Learning algorithms into Dewin.Water’s Cloud-based IoT platform “Pethau”. Dewin.Water utilise the sensor fault reporting and value threshold alerting features that are built into Pethau as standard for all customer modules. The dashboard shows the time series of each probe over different time periods.
Capturing sensor data leads to so-called “Time Series” of data – Long series of historical records and the analysis of these is a niche and complex field of research. Time-series AI/ML is unlike most other tensor/vector based AI applications. Dewin.Water developed necessary components in the platform to perform the “Data Science” elements of AI/ML work – namely exploratory work, training, model selection and hyper-parameter tuning – using live sensor data directly from the system.
Fantastic to see our tenants working together, creating jobs and innovating and their projects taking off!