NOTE! This site uses cookies and similar technologies.

If you not change browser settings, you agree to it. Learn more

I understand

Learn more about cookies at : http://www.aboutcookies.org/Default.aspx?page=1

May 5, 2020 | New version of LIMA shifts to deep learning

neurones 250The LIMA language analysis engine now has deep learning modules that can analyze 60 languages with state-of-the-art performance.

When a user enters a search term into a search engine, the text is thoroughly analyzed: every word, phrase, and sentence is detected and processed by special software. This analysis must be done before a search can be performed. However, it is also a prerequisite for other applications like automatic summarization and translation. List's language analysis engine, LIMA, is already widely used. Integrating deep learning modules into the software has made it even more powerful.

The latest advances in neural networks and a corpus of annotated texts in different languages provided by the Universal Dependencies cooperative were used to make the software more efficient, expand the number of languages supported, and add three learning modules. The first module segments text into words and sentences; the second performs a morphological, lexical, and syntactic analysis, and the third annotates the named entities identified.[1]

The previous version of LIMA could analyze six languages (English, French, German, Spanish, Portuguese, Chinese, and Arabic). The new version, called Deep LIMA, can analyze more than 60 languages with performance at the state of the art.

[1]An international cooperative project to create treebanks of the world's languages (https://universaldependencies.org/).

Read article at http://www.cea-tech.fr/

 

 

April 30, 2020 | Fiber Bragg grating systems for monitoring in harsh environments

List Reseaux de bragg 250Regenerated fiber Bragg grating systems can withstand the combined effects of high temperatures and gamma and neutron radiation, making them potential candidates for structural health monitoring in harsh environments.

Systems capable of ensuring structural health monitoring (SHM) in harsh environments require special sensors. Fiber Bragg grating (FBG) systems can provide temperature readings inside of complex equipment like nuclear reactors, where the flow of neutrons is intense. However, until recently, there had been little study of FBG sensor behavior when subjected to high temperatures and neutron and gamma radiation. CEA List, a CEA Tech institute, joined forces with other CEA divisions and with SCK in Belgium to test this promising technology in gamma irradiators and research reactors.

Regenerated* fiber Bragg gratings are made from patterns that are laser-inscribed into the cores of optical fibers and heat-treated (using a "regeneration" process, hence the name) to withstand temperatures up to 900 °C. This type of fiber Bragg grating is particularly well-suited to harsh environments. Here, the researchers exposed sensors to 1 MGy of gamma radiation and temperatures of 250 °C. The temperature measurement error after exposure to radiation was less than 2.7 °C, about the same as for a dose ten times smaller at ambient temperature.

Therefore, high temperatures appear to be beneficial to the behavior of regenerated FBG systems subjected to radiation. New tests with in-line monitoring will now be completed with the goal of identifying the optimal parameters for fabricating FBGs for use in research reactors. Ultimately, these advances could lead to in-core measurement, enabling much more detailed monitoring of reactors during operation.

*Fiber Bragg gratings are like mirrors that reflect particular wavelengths called Bragg wavelengths. Variations in the wavelength reflected can be used to determine the local temperature variation, which can thus be measured.

Read article at http://www.cea-tech.fr/

 

 

April 17, 2020 | Covid-19: CEA designs two emergency respiratory assistance systems

EVzoioWWsAMLI3BAs the Covid-19 pandemic continues to spread, hospitals worldwide risk not having enough ventilators. On 13 March the CEA kicked off the CLEAR (CEA List Emergency Assistance for Respiration) project to create additional capacity. Two affordable, high-performance prototypes have now been released as a result of this project. The first, CLEAR-M, is a monitoring system that is used in conjunction with conventional emergency and transport ventilators to make them more suitable for Covid-19 patients. The second, CLEAR-R, is an emergency respiratory assistance system.

CLEAR-M, a monitoring system for emergency and transport ventilators

In France, the number of ventilators already in service turned out to be sufficient to treat Covid-19 cases during the initial peak of the epidemic. However, in some cases, hospital staff did have to use emergency and transport ventilators, which do not have the continuous pressure and flow monitoring required to treat Covid-19 patients optimally. Physicians working in intensive care units felt that these types of ventilators could be upgraded to more effectively treat Covid-19 patients. List, a CEA Tech institute and member of the Carnot Network, partnered with the CEA’s Frédéric Joliot Medical Unit to develop the CLEAR-M monitoring system. On 10 April two tests were completed on a test lung in a hospital setting under the supervision of Dr. Frédéric Minko, who heads the Nord Essonne Hospitals Emergency Department in Orsay, France. The system was then tested at the Raymond Poincaré Hospital in Garches in the intensive care unit, which is headed by Professor Djillali Annane. CLEAR-M was compared with the measurements taken by a reference ventilator and performed well.

On 16 April, CLEAR-M was tested in the ventilator weaning ward at the Raymond Poincaré Hospital in Garches by Professors Hélène Prigent and Frédéric Lofaso on Covid-19 patients in recovery but still on ventilators. CLEAR-M was also implemented at the Nord Essonne Hospitals Emergency Department in Orsay.

In order to ensure free access to the system, the CEA patented it and made the license available free of charge for anyone wishing to manufacture it. BA-Healthcare reached an agreement with the CEA and will be the first manufacturer to produce several thousand units of the CLEAR-M system.

CLEAR-R emergency ventilators

To meet the needs of countries without enough ventilators, the researchers also developed a new emergency ventilator concept, CLEAR-R. The design combines a robotically-controlled manual resuscitation bag and the CLEAR-M monitoring system.

The proof-of-concept prototype was completed in late March. The prototype and hospital testing protocol were then developed in early April. Healthcare experts validated the prototype and its respiratory assistance capabilities, and List researchers then did their own tests on a simulator and in vivo at the National Veterinary School in Maisons-Alfort. At the same time, manufacturing and assembly instructions were written.

And, to make the new system widely and rapidly available, a test run of five ventilators was manufactured and made available to healthcare facilities and device manufacturers interested in getting the system certified, doing clinical trials, and launching volume manufacturing.

Request information on CLEAR-M or CLEAR-R

Better together

The CLEAR project was made possible by dozens of people at the CEA (through List and the Frédéric Joliot Medical Unit) in partnership with intensive care physicians at the Nord Essonne Hospitals in Orsay and at the Raymond Poincaré Hospital in Garches, as well as at hospitals in Marseille and Corbeil-Essonne, and, finally, with researchers at Inserm.

 

 

 

April 10, 2020 | To combine agility and competitivity in manufacturing industry? The EU DIMOFAC project takes up the challenge!

dimofac illustrationMass customization in a context of mass individualization or reorienting its production capacities to respond to a shortage in times of crisis are real challenges! How to reconfigure production tools while remaining competitive?

The solution is to combine a more agile organization and modular production lines with digital simulation to optimize the industrial manufacturing processes. The EU DIMOFAC project, which was officially launched in November 2019 at the CEA Grenoble premises (France), will develop enabling technologies.

The project aims at providing manufacturers with digital solutions for modeling and reconfiguring the production tool through the digital twin. This tool dynamically reproduces the production line at all points, making it possible to monitor operations, test different scenarios, and even control and reconfigure the physical line.

The ambition of the CEA List and Liten Carnot Institutes teams is to lay the foundations for a standardization of industrial architectures. "The idea is to make this digital twin and production systems modularity replicable, based on the Model Based System Engineering (MBSE) methodology" explains CEA List’s researcher Arnaud Cuccuru.

Demonstrators will be implemented on project partners' production lines, thanks to the tools developed by the CEA, such as CIVA for automated on-line non-destructive testing (NDT) and mostly, Papyrus for the functional digital twin of manufacturing production lines.

Learn more:

 

 

April 9, 2020 | Bringing order to industrial tasks

Factory Lab GECO List 250A software suite to automate the optimization and validation of industrial process task scheduling was developed at FactoryLab.

The automotive and aeronautics industries have particular throughput requirements. To meet these requirements, process tasks must be scheduled in a specific way. Tasks are scheduled to ensure that production is as efficient as possible. And scheduling factors in a multitude of parameters, from the human and material resources available to safety restrictions that would prevent two tasks from being completed simultaneously, for example. Finally, task scheduling, often based on experience, is very complicated to do manually. CEA List, a CEA Tech institute, recently completed an R&D project at FactoryLab with the goal of automating task scheduling.

FactoryLab community members PSA and Safran partnered with List on a joint project called GECO. The project's objectives were to reduce the amount of time it takes to optimize and validate task scheduling on PSA's automotive assembly lines and on Safran's motor maintenance lines. The researchers tailored their Papyrus-based software suite, which creates a digital twin of a production line, to the specific needs of the two cases addressed here.

This digital twin uses models and simulations to generate variants of all parameters and constraints, accurately analyze a complex situation, and predict and resolve any conflicts between tasks. This smart, automated system allocates resources and plans tasks, using heuristics developed with artificial intelligence. Unforeseen events (orders requiring fast turnarounds, a change in human or material resources) can be factored in to rapidly update the task schedule.

About FactoryLab:

FactoryLab, a consortium whose members represent industrial companies and academic research institutions, has the capacity to very rapidly integrate technological solutions and to build prototypes and demonstrator systems to support its members' transformation strategies. The consortium facilitates the adoption of these new solutions and helps shorten time to market.

FactoryLab spans diverse industries and markets, making it a novel space where ideas cross traditional barriers and where members can access shared resources. Whether they are technology providers, integrators, or industrial end-users, FactoryLab positions all of its users to create value.

Website: http://factorylab.fr/

Keep up to date on all of our projects: https://www.youtube.com/channel/UCEG31AgIx0v9CKeBzHci9LA

Read article at http://www.cea-tech.fr/