RES gets ‘seven times more’ HPC compute power with new cluster

RES is enabling more accurate decisions about wind turbine placements since boosting its HPC with new OCF cluster.


Renewable Energy Systems (RES) can now make more accurate decisions about where to build wind turbines since boosting its high-performance computer (HPC) with a a cluster from integrator OCF.

The UK-based renewable energy project developer uses its HPC to process historical data about wind in different locations. It then uses this information to generate wind maps for internal project teams and customers developing wind farms worldwide.

However, before it implemented a 128 terabyte OCF cluster supported by Cray Intel servers, the HPC did not have enough computing power to process more than a year’s worth of wind data for a specific location. Now it has enough power to process 10 years’ worth of information.

Clément Bouscasse, RES’ forecasting and flow modelling manager, told ComputerworldUK: “More accurate historical data has huge implications for us because it means you can do a lot more analysis the fact that we have time series data available we can look back at historical trends, you can do some correlation in a more efficient way…We can do a lot more than we could before with better wind maps and more things to offer to our customers internally.”

“Now we are struggling to keep up with the cluster,” he added. “We have seven times more computing power available and we need continuously find ideas and ways to keep it running. It is quite a challenging thing but very exciting for us.”

Sharing the HPC

With the computing boost, the wind flow team can dedicate some of its HPC power to analysts across global departments in the company.

RES implemented a range of applications on top of the computer to expand its capability, including installing Python programming language on the cluster.

Python has a number of scientific and numerical number crunching packages and is the language used internally by the technical team, including analysts. Previously developers could only run it on their desktop PCs.

RES used Python to design its new automated system that interfaces with its internal software across the global company, which allows other analysts to take advantage of the HPC.

As a result, the number of users has grown from a team of four in the wind flow team, to 40 analysts around the world using the wind flow calculations. This opens new renewable project possibilities, Bouscasse said.

“The part of our company looking at offshore wind farm development could be able to use it to do some meteorological ocean analysis in the near future,” he said.

Another weather-based organisation using supercomputers to crunch their data is the Met Office