“It was fascinating to learn that each machine has its own tonality and timbre. If it malfunctions, it produces a sort of moan, and if we catch it in time, we can make a quick and precise diagnosis and find the best solution to the problem.”
Anticipating the future to reduce costs
We don’t have supernatural powers, but the possibility of predicting the future definitely allows us to act in advance and reduce costs significantly. Thanks to this new technology, we can intervene immediately at the first alert, so it doesn’t affect other parts of the process and become a more complex problem. We will no longer have to stop the entire plant’s production, reducing the inactive times that occur when there is an unexpected malfunction or planned maintenance.
“Predictive technology also allows us to substitute just one part of a machine, reducing costs and intervention times.”
Our technicians will no longer have to physically go to the affected site, because they’ll be able to view and evaluate data directly from their mobile phones with an app. This will also transform the way we work.
For all of these reasons, we’ve decided to change our approach to maintenance, shifting from prevention to prediction.
“To better understand the economic benefits of this application, think about how we behave with our cars. About every 30 thousand km, we bring our cars to a mechanic for a check-up. But as we know from experience, this behaviour doesn’t protect us from unexpected breakdowns.”
Boeing conducted a study in the late '70s on airplane parts. It concluded that more than 50% of the times a plane was subjected to preventative maintenance, it turned out to be useless. For planes, it was a matter of over-maintenance: only 20% of failures can be predicted statistically, while 80% find us unprepared, forcing us into more complex maintenance on the entire vehicle.
Predictive maintenance, based on real data from the machines, reduces the number of interventions, allowing for targeted and precise actions that reduce time and costs.
Our experience at Soverzene
A preliminary test was performed at our Soverzene hydroelectric plant, in the province of Belluno, in Veneto. Here, we installed sensors on 7 machines: 4 turbines and 3 generators. Each system is equipped with a normal microphone, an ultrasound microphone and an on-board computer, designed to sample and communicate acoustic signals.
“The data is then transmitted to a cloud in real time, analysed and classified into different clusters of failures, to be able to activate the best solutions.”
This initial experience on the ground, if it produces the desired results, can be extended to all Enel Green Power plants, even in other technologies, making us a leader in the field of predictive maintenance.