DESCRIPTION

ESO is collecting large amounts of data about its assets, including asset operational and maintenance data. Such data can be used to move from more time-based and reactive maintenance to proactive and predictive maintenance models.

PROBLEM

ESO engineers use their own models to predict risk in the network, but with increasing amounts of data, such models become more complex and more difficult to analyze.

SOLUTION

Predictive Maintenance pilot was testing predictions on existing data models, using historical data from the distribution network assets. They are used to determine the risk of failure and learn about potential defects.

Pilot was established with two companies: US company GridCure creates a cable line failure prediction model and French company Sensewaves was used for modeling the overhead line, transformer data analysis and fault prediction.

STATUS AND PROGRESS

The project was finalized with the identification of additional data needed for more accurate modeling. The solution should be reevaluated when smart meters data will be available.