The research on artificial intelligence elaborated by Indra together with the Center for Surveillance and Data Analysis of the Navy (CESADAR) and the University of A Corua was submitted as the best of all articles at the VIII National Congress on Research and Development in Defense and Security.
Of the 140 papers accepted, the one produced by these three partners, due to its level of innovation, attracted the most interest from the members of the Congress Committee, who decided to publish it to share with the scientific community.
Under the title Predictive maintenance of marine engines through automatic learning, the document prepared by the researcher at the University of A Corua, David Novoa, as lead author, together with its directors, Mster Carlos Eiras and Scar Fontenla, with the support of Alfrez de Navo and CESADAR technical officer, Francisco Lamas, and Dr. David Sanz from Indra analyze the use of new artificial intelligence techniques to detect motor malfunctions before they even occur.
This work comes from the SOPRENE R + D + i project, funded by the General Directorate Armament and the General Directorate Planning, Technology and Innovation (DGAM-PLATIN), which Indra carried out together with the University of A Corua and the University of Alcal de Henares in CESADAR, the Spanish Navy’s reference unit in the application of mechanical failure prediction techniques, technically promoted and directed them using the data stored in their Cartagena facilities.
As part of the SOPRENE project, a technological demonstrator that enables predictive ship maintenance was successfully tested. It is a unique solution that combines deep learning, predictive algorithms and diagnostics and is characterized by the use of unsupervised intelligence.
This type of intelligence has great potential because, unlike supervised AI, it uses a strategy that aims to teach the machine to recognize and solve problems without the need to do so in the first place. This makes it easier to use with very different platforms, systems, or devices to identify errors that have never occurred before and for which there is no record, making it different from a traditional AI system.
This latest feature has made SOPRENE the best available solution for servicing next generation vessels that have not yet sailed. In the same way, in combination with the prediction algorithms that have also been developed, it can detect the most serious errors in advance: those that have never occurred, but which in this case could endanger the crew.
The number of applications that this technology offers is enormous. In the military field it could be used to maintain equipment and platforms of one of the three armies, while in the civil field it can improve the maintenance of all types of industrial equipment. In the metal industry, for example, maintenance costs can represent up to 60% of total production costs, which gives an idea of the savings and efficiency improvements such a system can bring.
Copyright © Grupo Edefa SA Reproduction, in whole or in part, of this article is prohibited without the prior authorization of the publisher.