Accuracy Talks Straight #3 – Start-up stories
On 21 October 2021, Amiral Technologies announced its first round of fundraising totalling €2.8m. This represents an initial success for the start-up, founded in 2018 in Grenoble on the basis of an observation shared by numerous industrial players: how can we reliably predict breakdowns?
A spin-off of the CNRS, Amiral Technologies is based on almost 10 years of university research in artificial intelligence and automation & control theory. The company has successfully developed disruptive technology: from sensors installed on machines, detecting physical signals such as electric current, vibrations or humidity, algorithms make it possible to generate general health indicators for the equipment. These health indicators are then interpreted by unsupervised machine learning algorithms. They make it possible to identify the causes of breakdowns most likely to take place.
Unlike the majority of other solutions on the market, this solution (named DiagFit), which makes use of machine learning, does not require the history of breakdowns identified on a piece of equipment to be able to use artificial intelligence. Indeed, the algorithm is adapted to a specific use case in order to define a normalised functioning environment for the equipment.
More precise, quicker, and independent of the sensors themselves, the technology is already in use with SMEs and mid-sized businesses, as well as with large industrial groups such as Valéo, Airbus, Daher, Vinci and Thales.
The predictive maintenance market benefits from sustained growth dynamics, driven by an industrial base equipped with more and more sensors, a need to optimise inventories of spare parts and, of course, a greater need to avoid any costly shutdown in the production chain.
Amiral Technologies now aims to become the top supplier for the European market. The fundraising will enable it to strengthen its technical and commercial team, as well as to accelerate the development of DiagFit and its scientific and technological research.