No abstract

Wind speed forecasting is important for high-efficiency utilisation of wind energy and management of grid-connected power systems. Due to the noise, instability and irregularity of atmosphere system, the current models based on raw historical data have encountered many problems. In this study, a deep novel feature extraction approach is developed based on stacked denoising autoencoders and batch normalisation. Then the deep features extracted from raw historical data are fed to long short-term memory (LSTM) neural networks for prediction. Meanwhile, density-based spatial clustering of applications with noise is employed to process the numerical weather prediction data. By picking out the abnormal samples, the representative training samples are selected to improve the efficiency of the model. For illustration and verification purposes, the proposed model is used to predict the wind speed of Wind Atlas for South Africa (WASA). Empirical results show that deep feature extraction can improve the forecasting accuracy of LSTM 49% than feature selection, indicating that proper feature extraction is crucial to wind speed forecasting. And the proposed model outperforms other benchmark methods at least 17%. Hence, the proposed model is promising for wind speed forecasting.

No abstract

With the acceleration of industrialization and urbanization, most lakes and reservoirs have been in eutrophication state. Eutrophication of water body will produce a series of environmental problems, among which cyanobacteria bloom is one of the most studied and seriously polluted problems. It is of great significance to effectively control the occurrence of cyanobacteria blooms by predicting and simulating the outbreak process of cyanobacteria blooms and accurately forecasting the relevant governance departments. However, there are two problems in the existing analysis of algal blooms: on the one hand, it is difficult to consider the impact of other factors on cyanobacteria blooms by taking chlorophyll concentration as the main influencing factor, and it is also unable to determine the relationship between various factors. On the other hand, only based on the field monitoring data research, lack of comprehensive analysis of the whole water area. The remote sensing image can reflect the change of the whole water area, but the traditional analysis method is difficult to deal with the massive remote sensing data effectively. In this study, eutrophication level was used as characterization index of cyanobacteria bloom, and the remote sensing image and its inversion map were taken as the main research data, and a new method of cyanobacteria bloom prediction based on four-dimensional (4D) fractal CNN was proposed. The prediction model uses 4D fractal CNN to extract the features of multi factor remote sensing images, capture the temporal and spatial characteristics and the interaction between multiple factors, and predict the eutrophication level of water body. In this study, a total of 216 remote sensing images of Taihu Lake Basin were selected from 29 groups with fine weather from 2009 to 2010 obtained by MODIS satellite. The simulation results show that the method proposed in this paper has excellent prediction performance, and the accuracy rate of 85.71% is better than that of common 3D CNN and 4D CNN models.

No abstract