A Wind Power Prediction Method Based on Deep Convolutional Network with Multiple Features

As a clean and renewable energy, wind power plays an increasingly significant role in the power system. And wind power prediction is crucial for the operation planning and cost control of power plants. In wind power prediction, the wind-related information such as wind speed, wind direction, air pressure and temperature will affect the accuracy of prediction. However, most of the existing models either use only one kind of information or fail to effectively integrate a variety of information. Considering these problems, a new convolutional neural network model is proposed by integrating multiple information based on spatio-temporal features, called FB-CNN (Feature Block CNN). Obtain the output of each convolution layers in the whole neural network, then integrate these outputs, the prediction results are obtained through the full connection layers in the end. Compared with the current convolutional network with the highest accuracy, the proposed FB-CNN combining various features can excellently fit the actual change of wind power data. And the Mean Square Errors (MSE) on two data sets are reduced by 9.53% and 7.13% respectively.

[1]  Jie Zhang,et al.  A data-driven multi-model methodology with deep feature selection for short-term wind forecasting , 2017 .

[2]  Bin Li,et al.  Scene Learning: Deep Convolutional Networks For Wind Power Prediction by Embedding Turbines into Grid Space , 2018, Applied Energy.

[3]  Oliver Kramer,et al.  Enhanced SVR ensembles for wind power prediction , 2016, 2016 International Joint Conference on Neural Networks (IJCNN).

[4]  Zongxiang Lu,et al.  Probabilistic short-term wind power forecasting based on deep neural networks , 2016, 2016 International Conference on Probabilistic Methods Applied to Power Systems (PMAPS).

[5]  Na Zhang,et al.  A novel wind speed forecasting method based on ensemble empirical mode decomposition and GA-BP neural network , 2013, 2013 IEEE Power & Energy Society General Meeting.

[6]  Meysam Qadrdan,et al.  Probabilistic wind power forecasting and its application in the scheduling of gas-fired generators , 2017 .

[7]  Mengxuan Song,et al.  A combined AR-kNN model for short-term wind speed forecasting , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[8]  Florian Ziel,et al.  Forecasting wind power – Modeling periodic and non-linear effects under conditional heteroscedasticity , 2016, 1606.00546.

[9]  Hamidreza Zareipour,et al.  Wind power forecast using wavelet neural network trained by improved Clonal selection algorithm , 2015 .

[10]  Oliver Kramer,et al.  Wind Power Prediction with Machine Learning , 2016, Computational Sustainability.

[11]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  Robert P. Broadwater,et al.  Current status and future advances for wind speed and power forecasting , 2014 .