Weather image-based short-term dense wind speed forecast with a ConvLSTM-LSTM deep learning model

Abstract

<p>Short-term <a href="https://www.sciencedirect.com/topics/earth-and-planetary-sciences/wind-velocity" title="Learn more about wind speed from ScienceDirect's AI-generated Topic Pages">wind speed</a> predication is of great significance for scholars (e.g., understanding wind profiles), practitioners (e.g., building energy management), regulators (e.g., urban <a href="https://www.sciencedirect.com/topics/engineering/microclimate" title="Learn more about microclimate from ScienceDirect's AI-generated Topic Pages">microclimate</a> regulation), and even the general public. Current wind speed forecasting methods either generate sparse predictions or occur high cost. This paper reports a novel, inexpensive framework to forecast urban local dense wind speed. The central tenet is a convolutional long short-term memory (ConvLSTM) and LSTM combinatorial <a href="https://www.sciencedirect.com/topics/engineering/deep-learning" title="Learn more about deep learning from ScienceDirect's AI-generated Topic Pages">deep learning</a> model to learn the features of input historical weather image series coupled with spatial-temporal correlations. The model was trained and tested using Hong Kong datasets. The feasibility and effectiveness of the proposed model are verified and compared with parallel models under different criteria, including <a href="https://www.sciencedirect.com/topics/engineering/mean-absolute-error" title="Learn more about mean absolute error from ScienceDirect's AI-generated Topic Pages">mean absolute error</a> (MAE), <a href="https://www.sciencedirect.com/topics/engineering/root-mean-square-error" title="Learn more about root mean square error from ScienceDirect's AI-generated Topic Pages">root mean square error</a> (RMSE) and R-squared (R²). The experimental results show that: (1) the proposed ConvLSTM-LSTM deep learning model can effectively forecast wind speed regardless of location; (2) the overall MAE, RMSE, and R² value of the proposed model are improved by 14.84%, 15.04%, and 7.51%, respectively, compared to the ConvLSTM-full connected (ConvLSTM-FC) model, and by 22.12%, 22.80%, and 12.24%, respectively, compared to the convolutional neural network-LSTM (CNN-LSTM) model; and (3) compared with parallel models, the proposed model has better performance in predicting wind speed series with large amplitude variations and rapid frequency changes.</p><ul></ul>