Sites Inria

Version française


Jean-Michel Prima - 13/06/2017

Leveraging Big Data Analytics to Cut Industry's Power Bill

Véronique Masson et Arnaud Legrand Véronique Masson & Arnaud Legrand

A French software startup specializing in energy analytics for the industry, Energiency helps manufacturers to save big on electricity and gas. The company has recently started a partnership with Inria and Rennes 1 University to stay on the cutting-edge of data science.

50 million euros. That's roughly what some of the leading car manufacturers spend on their global monthly power bill today. And indeed, energy has become a huge budget line item for all kinds of industries. No wonder, then, why the slightest prospect of enhanced energy efficiency is regarded as a gold mine.
Identifying these potential cost savings and slashing the bill by up to 20% is precisely the raison d'être of Energiency a company founded four years ago in Rennes, Brittany, France by Arnaud Legrand, a former energy consultant with Ernst & Young. “In my previous job, I would spend like a year or so working on a particular plant, helping the management to pinpoint new cost-cutting opportunities. It ran the whole gamut from, say, switching off the light when you leave to  rebuilding the plant from scratch. I would spend most of my time collecting the data that was scattered all over ERP databases, spreadsheet applications, photocopies and whatnot. Frustratingly enough, it was only at the very end of this lengthy collection process that I could start working in earnest and bring my value-added expertise. It took so long that some of the data was outdated before we could even act upon it.
In contrast to that slow pace consulting of yore, the Energiency company was started as a software editor meant to leverage three game-changing technologies in order to enable much swifter decision making. Namely: the Internet of Things (IoT), cloud computing and Artificial Intelligence (AI). “Today in smart factories, countless internet-connected sensors collect all sorts of data about the manufacturing process. This information is not stored locally anymore but shipped to repositories in the cloud. Coming on top of that as a web application, our software enables our customers to make sense of the data and carve out a cost-saving plan right away.

Strong Inria DNA

At the crux of this business lies a fledgling yet promising science called datamining. And that's where comes into play the partnership with Lacodam  a Rennes-based Inria research team pursuing novel approaches to data analysis. “Our company actually has a strong Inria DNA, I should say. There are no less than four PhD holders from Inria on our payroll, including our Technical Director and our R&D Director. Having said that, Development keeps us rather busy and we can't allocate as much time as we would like to Research. Therefore, right from the get-go, working with a laboratory was in the picture.

Signed last October, the three-year partnership resulted in the recruitment of PhD student enrolled in the doctoral course in the laboratory. “The collaboration aims at maximizing the automation of knowledge discovery and making the most sense of the data. In that regard, Lacodam has a real expertise. We would be very interested in seeing their technological brick take a crack at solving concrete problems that we have singled out. That's basically the deadlock that we want to overcome together.

Bridging Data Mining and AI Approaches

Getting the chance to handle real-world data sets and test our methods on them is precisely what we are yearning for , says Lacodam scientist  Véronique Masson . Indeed, it brings new problems to the surface. Recent algorithms are capable of extracting regularities ―or irregularities― from huge volumes of data in a rather efficient fashion. But attending data mining conferences these days, one soon realizes that the real difficulty now consists in applying those algorithms to specific cases. The question that still remains is how to make sense of what is being extracted. In other words, how to transform it into domain knowledge and derive sense that will be of value to the end user.
To meet that challenge, scientists are working on a mix of data mining techniques and artificial intelligence approaches. “We are trying to get the best of both worlds. This is still exploratory research, but the start-up turns it meaningful with real use cases that will bring it to the market as fast as possible. And indeed, the first algorithms put forward by our PhD student  have already been passed to the company's engineers for consideration,” Masson remarks.  “With this  research partnership, we are building for the future, Legrand concludes. It will help Energiency to keep one step ahead.

Keywords: Energiency Lacodam Arnaud Legrand Véronique Masson INRIA Rennes - Bretagne Atlantique Energy Big data