A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Deep learning to predict the generation of a wind farm
One of today's greatest technological challenges is adding renewable energies to an electric grid, with the goal being to achieve sustainable and environmentally friendly electricity generation that is also affordable. In order for the incorporation of renewables to be successful, however, predictive tools are required which can be used to determine sufficiently far in advance how much renewable energy will be available to be injected into the grid so that the remaining generation sources, including those based on fossil fuels, can be adjusted in order to fill the demand. This would limit the environmental impact and the dependence on this type of fuel in a foreseeable shortfall scenario. This paper seeks to advance in the creation of these predictive generation models for wind farms using deep learning. We present a predictive model based on a deep, multi-layered neural network that based on a forecast for atmospheric conditions is capable of estimating the generation produced by a wind farm 24 h in advance. These models were trained and validated with data from a wind farm located on the island of Tenerife and show that the best of these predictors is more precise than the reference estimator and the prediction model currently used at the farm. We also note that the problem does not require models based on truly deep neural networks. However, the workflow for correctly developing, training, validating, and tuning these models is greatly enhanced by the advantages that deep learning techniques and tools can offer.
Deep learning to predict the generation of a wind farm
One of today's greatest technological challenges is adding renewable energies to an electric grid, with the goal being to achieve sustainable and environmentally friendly electricity generation that is also affordable. In order for the incorporation of renewables to be successful, however, predictive tools are required which can be used to determine sufficiently far in advance how much renewable energy will be available to be injected into the grid so that the remaining generation sources, including those based on fossil fuels, can be adjusted in order to fill the demand. This would limit the environmental impact and the dependence on this type of fuel in a foreseeable shortfall scenario. This paper seeks to advance in the creation of these predictive generation models for wind farms using deep learning. We present a predictive model based on a deep, multi-layered neural network that based on a forecast for atmospheric conditions is capable of estimating the generation produced by a wind farm 24 h in advance. These models were trained and validated with data from a wind farm located on the island of Tenerife and show that the best of these predictors is more precise than the reference estimator and the prediction model currently used at the farm. We also note that the problem does not require models based on truly deep neural networks. However, the workflow for correctly developing, training, validating, and tuning these models is greatly enhanced by the advantages that deep learning techniques and tools can offer.
Deep learning to predict the generation of a wind farm
Torres, J. M. (author) / Aguilar, R. M. (author) / Zuñiga-Meneses, K. V. (author)
2018-01-01
15 pages
Article (Journal)
Electronic Resource
English
Deep learning for optimal dispatch of automatic generation control in a wind farm
| American Institute of Physics | 2023
Wind in the moorlands: Wind farm piling
British Library Online Contents | 2007
Wind flow deformation inside the wind farm
| Online Contents | 1998