Posters
Come meet the poster presenters to ask them questions and discuss their work
We would like to invite you to come and see the posters at our upcoming conference. The posters will showcase a diverse range of research topics, and will give delegates an opportunity to engage with the authors and learn more about their work. Whether you are a seasoned researcher or simply curious about the latest developments in your field, we believe that the posters will offer something of interest to everyone. So please join us at the conference and take advantage of this opportunity to learn and engage with your peers in industry and the academic community.
PO110: Long-term turbulence intensity reconstruction using hybrid machine learning and physical validation
Pranoy Ray Chowdhury, Energy Yield Analyst, Lightsource bp (BP)
Abstract
Accurate long‑term turbulence intensity (TI) characterisation is critical for time‑series‑based wake loss and energy yield assessment. However, conventional long‑term correction methods such as Measure‑Correlate‑Predict (MCP) are designed for mean wind speed and do not directly address TI, whose variability is strongly influenced by site‑specific roughness, topography, and atmospheric stability. This challenge is amplified when only short‑term on‑site measurements are available. We present a hybrid methodology that combines machine learning (ML) with physical verification to generate long-term hourly TI time series using limited-duration mast measurements and widely available long-term reference datasets. Unlike traditional binning-based extrapolation, the ML (gradient boosting) model directly learns nonlinear TI dependencies on wind speed, direction, hub-height temperature (as a stability proxy), and temporal features (month, hour) from short-term site data. Statistical binning is retained only for validation and bias correction, ensuring percentile alignment and extreme event fidelity. The methodology is applied to five 12‑month onshore datasets in moderately complex terrain and the trained model is applied to long-term corrected hourly data. The ML model captures nonlinear interactions between wind speed, direction, temperature, and seasonality, improving reproduction of TI statistics. This hybrid framework addresses the core limitations of standard MCP for turbulence and provides a replicable pathway for bankable TI reconstructions. The resulting long‑term TI datasets enable more accurate time‑series‑based wake loss assessments and IEC load‑case definitions, while also reducing uncertainty in hybrid plant analyses, ultimately supporting more reliable project design and financing.
No recording available for this poster.
