Atmospheric Boundary Layer Wind Profile Estimation Using Neural Networks, Mesoscale Models, and LiDAR Measurements.

Area of Science:

• Atmospheric Science
• Meteorology
• Renewable Energy

Background:

• Accurate wind profile estimation is crucial for wind energy applications and atmospheric modeling.
• Traditional methods often require extensive measurements or complex models.
• Existing solutions may lack the efficiency or adaptability for dynamic wind condition analysis.

Purpose of the Study:

• To introduce a novel methodology for estimating the atmospheric boundary layer (ABL) wind profile.
• To leverage neural networks, mesoscale model predictions, and single near-surface measurements.
• To provide a simplified and potentially time-evolving wind profile estimation technique.

Main Methods:

• Development of a neural network model trained on mesoscale predictions and near-surface data.
• Utilizing LiDAR sensor data from the University of León, Spain.
• Integration of a single near-surface measurement for real-time or post-processed prediction.

Main Results:

• Demonstration of a novel method for estimating wind profiles in the ABL.
• Highlighting the advantage of requiring only near-surface measurements post-training.
• Showing potential for analyzing the temporal dynamics of wind profiles.

Conclusions:

• The proposed methodology offers an efficient approach to wind profile estimation.
• This technique simplifies prediction requirements, relying on trained neural networks and minimal real-time data.
• The findings are valuable for wind energy assessments and computational fluid dynamics (CFD) modeling.