Short Term Wind Energy Prediction for Manisa Soma Region By Using Numerical Weather Prediction Model Results (WRF) and Computational Fluid Dynamics Models (Meteodyn And WindSim)

Short-term wind energy estimations were made using Meteodyn and WindSim computational fluid dynamics models (CFD) with the medium-scale numerical weather forecasting model WRF. It is aimed to create a wind energy forecasting system that estimates the wind energy in Soma Wind Power Plant (WPP), which is located on a complex land within the boundaries of Soma district of Manisa province, at 1 km and 3 km resolution and for 24 and second 24 (48) hours. he modeling results are compared with the turbine observation data to analyze which CFD model performed better. In the comparison of the results, it is observed that the 24 hours results of meso scale model and both CFD models in both resolution scenarios were showing less error compared to 48 hours results. This is explained by the reduction of the prediction ability of the WRF with the prolonged estimation period which caused the high error rates. Analyzing the resolution results, it is observed that 3 km resolution for all models gave less erroneous results than 1 km resolution. This is explained as the 3 km resolution gives less erroneous results than 1 km error due to the better matching with the physical properties of the model. This result is thought to be the damping or wind speed enhancement resulting from the altitude and land-use scale differences between mesoscale model and CFD models. Furthermore, the meso model WRF is affected by the complex field structure. It is expected that micro scale effects will be better solved and the error rates will be reduced with the using of CFD model. The WRF model is better or gives very close results to CFD models. This shows that the height, roughness and land usage data scale difference of the models have a great effect on the results. In seasonal analyzes; it has been observed that all the CFD models, resolution and time conditions has the highest error values in the winter season and the lowest error values in the summer season.