Lettres d'information

Choisissez ce que vous souhaitez recevoir :

Merci de renseigner votre adresse de messagerie électronique :

@

Votre adresse email ne sera utilisée que dans le cadre de campagnes d'information ITER Organization auxquelles vous êtes abonné. ITER Organization ne communiquera jamais votre adresse email et autres informations personnelles à quiconque ou dans le cadre d'informations commerciales.

Si vous changez d'avis, il vous est possible de vous désinscrire en cliquant sur le lien 'unsubscribe' visible dans vos emails provenant d'ITER Organization.

Pour plus d'information, veuillez consulter notre Politique de confidentialité.

Actu & Médias


Pour les actualités en français, voir la page Actus & Médias.
Data archiving

Operating in quasi real time

To accommodate the first real-time system integrated with the ITER control system, new components of the data archiving system have been deployed.

Data archiving, which has been in place since the beginning of the project, now operates in ''quasi'' real time to store information coming out of plant systems as they come on line. (Click to view larger version...)
Data archiving, which has been in place since the beginning of the project, now operates in ''quasi'' real time to store information coming out of plant systems as they come on line.
Data archiving, an essential requirement of ITER, is carried out by a collection of integrated software components. The continuous archiving system has been in place to collect and store data since the beginning of the project. To support the first real-time system—the reactive power compensation system—another component was introduced in 2022 to enable archiving of fast data. Fast data is real-time data emitted by even-driven applications and that usually needs to be analyzed quickly. In any case, it needs to be stored as it comes.

The data archiving system now operates in "quasi" real time, meaning it closely approximates real time operation. The quasi real-time data archiving system collects and stores data immediately, and it does so at rates of up to 37 GB per second.

The structure of the data, the amount of data, and the frequency with which data is output varies from one system to another. "We have slow and continuous systems, which produce sparse data, and we have fast systems, which produce dense data," says Lana Abadie, Data Service Coordinating Engineer in the Data, Connectivity and Software Section. "The way of processing the output from the different types of systems is different, with different needs for both storage space and computing power."

Data archiving is important in all phases of the project. Currently, for example, engineering and test data from the commissioning and early operation of plant systems is being stored. This data helps in the evaluation of performance in the short term, and it is stored for the long term to keep a record of how the machine was built for future analysis and to inform future designs.

Data is also used to find tiny problems that might not otherwise manifest themselves for weeks or months. "During commissioning, sometimes you look for defects that are not directly measurable," says Bertrand Bauvir, leader of ITER's Central Control Integration Section. "You operate for a long time, and you archive all the data. Sometimes finding a problem is like looking for a needle in the haystack."

As an example, a motor and a pump might be mechanically misaligned. This might be spotted through higher-than-expected electrical current measurements or features in the data from vibration sensors. Data collected on the electrical current consumption in the motor, the rotation speed, and the pressure inside the compressor can be used by the commissioning engineers to determine whether equipment is operating as expected.

In the cryoplant, where a lot of gaseous helium is stored, a tiny leak cannot be detected immediately. But as the tanks fill up with gas over several months, some of the physics measurements taken inside the tanks will drift slowly, providing a clue. Here too, stored data is valuable in the interpretation of tiny features over very long periods of time.

Once operations begin, diagnostic data will be archived during each pulse to help scientists learn from experiments. The data will be analyzed minutes and days after each pulse—and it will be stored for future analysis and shared with scientists around the world. It will probably be useful for decades to come.

Data will also help operators adjust configurations for subsequent pulses, and scientists and engineers to conduct post-pulse analyses—an important endeavor not only when failure occurs, but under normal circumstances to compare real-life behaviour with that predicted by simulators. Most of all, though, the data will lead to new and surprising discoveries.



return to the latest published articles