Experiment #207

Artificial Intelligence applied for anomalies detection

About 

Early detection of anomalies can bring countless benefits. Therefore, the purpose of experiment #207 is to use Artificial Intelligence to solve all kind of problems possible, focusing on three indicators: the degree of dirt on solar panels, the degree of corrosion in the different parts of windmills and the presence/absence of PPE equipment in the people who are in the enclosure of a building.

Idea

The idea was born due to the need to anticipate problems in the energy sector that generate losses such as the loss of energy absorption caused by the excess of dirt or the enormous amount of money that has to be invested to solve corrosion problems in windmills blades.

Process

We start with a data source as input from video cameras. This data will be captured by a series of Python Scripts, which will be responsible for shelling that information. Then, a number of Azure Custom Vision models will be trained to identify the degree of the problem. Likewise, a series of Artificial Intelligence models will support more customized situations. Finally, all the information generated by our models will be stored thanks to Azure Data Factory in a SQL instance in PaaS, where it will feed our Power BI report for decision making.

Utility 

Anticipating the problems that can arise in these situations has undoubted benefits:

Increasing the energy absorption of solar panels or avoiding a disaster that would have been totally avoidable if they had been wearing PPE, as well as fixing corrosion on the shaft of a windmill before it can even break.

Stay up to date!



Leave a comment