Artificial intelligence boosting the efficiency of electricity network repairs and planning


09.11.2018 at 10:45
Hannu's home experiences short power cuts every couple of hours. The computer starts up by itself at night and turns off right in the middle of work. There is a fault in the network and it is making it hard for Hannu to work. What if power cuts could be forecast precisely and nobody needed to be left without electricity? In the future, artificial intelligence and analytics will make this a reality.

At present, Caruna receives a flood of data on the electricity consumption of almost 700,000 customers and on the condition of an electricity network measuring 85,000 km. The volume of data is enormous, and it is slow to analyse the data manually and prioritise disruptions. Processing this data efficiently calls for artificial intelligence.

Smart forecasting of storms shortens power cuts

Caruna aims to utilise customers' data more smartly than now. When external data such as weather and temperature information is combined with data on the electricity consumed by homes, we can quickly obtain precise information on abnormal consumption changes or disruptions. Automatic analysis would quickly flag up the abnormally frequent power cuts affecting Hannu's home. It would also reveal more accurately where the fault lies – in the external electricity network or the home's internal network. In the light of more accurate information, we would be able to rectify the abnormal disruptions immediately and the electricity supply would be normalised more quickly

We already make use of weather forecasts, which enable us to prepare ourselves for power cuts caused by the weather. However, the estimates are modelled manually, so the forecasts are not as precise as when precipitation and wind data are modelled automatically. Automatic modelling speeds up repair work, as fault management resources can be allocated more effectively. The most common incidents of weather damage are caused by falling trees, snow loads and thunder storms breaking transformers.

Analytic data is also collected for network planning. Our systems are constantly modelling how much electricity will be consumed in the future. This enables us to reinforce the network in the right places and at the right time. 

The electricity network is constantly monitored, but what if the network could automatically communicate its own worries?

Caruna's maintenance team inspects transformers and lines on the electricity network at regular intervals. In practice, this means that an electrician inspects different parts of the network and makes notes on them. The information is entered into the system manually. New technology and new operating models are constantly being introduced to boost the efficiency of network condition monitoring. Some of the inspection work is already conducted by helicopters that image the overhead line network to check for issues such as trees growing too close to overhead lines. Drones are a future technology for this purpose. Caruna is also running a crowdsourcing pilot project with Posti, whereby postmen take images of specified parts of the network along their routes.

Caruna is facing a challenge that every smartphone user will be familiar with – inspections are generating images in such large numbers that they cannot be reviewed quickly enough. In the future, it must be possible to automatically analyse the images and import them onto the maintenance system, which will also prioritise and schedule maintenance jobs. In the future, anybody will be able to use the Caruna+ app to send an image of a part of the network that needs maintenance. Spotting faults could also be gamified, and image reports could accrue points for customers. In the future, instead of hunting Pokémons, people may be searching for Carumons!

Share article
Facebook icon Twitter icon LinkedIn icon