Olivier Douillard from AIM45 tells us how Warp 10 helps them analyze ocean races data.
AIM45 develops data measurements and analytics solutions for racing sailboats. The company offers a large service from the measurement of indicators to the data interpretation and the creation of business tools.
With the arrival of flying boats, AIM45 has embarked on a research program about the new data analysis tools of tomorrow. The diversification of sensors and the complexification of on-board measurement systems require a solution adapted to these new issues coming with flying boats technology.
What are the type and the volume of data generated by a racing sailboat? (And what kind of sensors?)
Olivier: A large diversity of sensors going from the wind measurement to the composite deformation of appendages or hulls. We have load sensors for the rig, or for the hydraulic system for appendages trim. As the goal is to go around the world, we are concerned by the energy production (generator, solar, airflow) and so by the complete picture of the different system consumption. There is a big effort to develop an automatic pilot for the single handle sailing for a flying boat. So there is quite a lot of parameters to trig the intelligence.
We have an inertial central at 30 Hz to log the dynamic of the boat and optical fiber at 100 Hz to measure internal deformation and so effort on appendages and hull.
To resume the complete daily log for a maxi trimaran racing sailboat is around 225 million data points for more than 600 sensors or parameters.
|Learn more about Time Series: the future of data|
For what purpose do you use Warp 10 at AIM 45?
Olivier: We use Warp 10 as a time series platform to manipulate large amounts of data. Our field game is large and could represent 80 full days around the world for the next Vendee Globe. So be able to extract valuable information from the real-life is a huge step for the skipper but also for engineers and designers.
Also, this kind of boat has an ongoing process of maintenance and development. So to be able to have quick, easy feedback from specific sessions is crucial. For example, we are using Warp 10 to provide automatic detection of extremum and failure patterns. Another application is to create automatic segmentation of the boat track in accordance with wind stability or specific performance tests.
What future do you see for data analysis for racing sailboats?
Olivier: There were always important developments in racing sailboats. Last years are very exciting with the generalization of foils on any kind of boat. But this big change for us, create a more complex boat with many systems onboard. So data is a real key to the future development of sailing boats to be like a true link between the designer and the offshore experience. Also, the high speed and complexity push skippers to add data to their feeling.
Do you have an example of a success story ?
Olivier: It was for the Ocean Race. We were able to fine-tune the keel angle and that gave a real performance advantage for the boat.
Do you plan to embark Warp 10 for decision support during a race?
Olivier: I think this is a next step to be able to create real-time efficient monitoring of the boat performance especially in rough seas.
What were the decisive criteria that made you choose Warp 10 rather than another solution for AIM45?
Olivier: It is clearly a dedicated solution for our time series data. SenX physical proximity is also important to us.
Discover Amzair, another company using the Warp 10 platform.
Ingesting 300k datapoints per second in the Warp 10 timeseries database on a Raspberry Pi is possible with Warp 10 2.1 new ingestion format. We also benched the PINE64 board!
Data deletion in levelDB can take time. Discover the levelDB extension to speed up deletion within Warp 10 standalone!