As the technical team at Stirling-based Natural Power has been developing new models for more accurately predicting wind farm performance, investment in new high performance computing has enabled the business to produce results from its most advanced computational fluid dynamics (CFD) models at wind farm scale within five to six days, rather than the 60 days previously needed.
Natural Power helps its global wind power clients create and analyse this data. It currently manages 26% of the UK’s installed onshore wind capacity and is steering more than 50 wind farms through construction.
Behind every wind power installation is a complex series of calculations and analysis of potential performance.
When planning new installations, investors, utilities and turbine manufacturers all need an accurate picture of what the energy yield will be, whether it will deliver an attractive return on its financing, and whether it will affect local populations and wildlife
Claude Abiven, Senior Technical Manager, explained: “Wind power has enormous potential, but there are multiple variables involved at each installation that will affect yield and, consequently, the financial viability of both new and existing sites.
“Typically our customers need an accurate picture of the likely output of the wind farms they are planning to build or buy so they can determine projected profit and hence the loan criteria. The more accurate the calculation is, the more attractive the terms for financing become – and the more easily wind power can be integrated into the overall energy mix.”
“The improved speed creates an obvious commercial advantage. No client wants to wait 60 days for modelling outcomes.
“More significantly, it means we can increase certainty in our calculations, which can be directly translated in to cost savings. When the bill for a wind farm runs in to hundreds of millions, that’s an important factor – and a potential game changer for the industry.”
Simple linear wind-flow prediction models can run off a laptop computer, but more computationally intensive models are required when farms are located in complex or forested terrain or in the presence of complex atmospheres such as sea breezes. In these circumstances, linear wind-flow models provide a much less accurate picture.
To overcome the limitations of existing techniques, Natural Power worked in partnership with the Faculty of Engineering at the University of Porto (FE-UP) in Portugal and Renewable Energy Systems (RES) to develop a more advanced CFD. While developing the model, Natural Power realised it would not be able to run effectively on its existing server.
Natural Power has selected a SuperMicro supercomputer powered by Intel Xeon processor E5-2660 v3. Internal assessments have shown that this runs 1.7 times faster than the previous installation.