With the development of the Smart Grid, accurate prediction of power generation is becoming an increasingly crucial task. The primary goal of this research is to create an efficient and reliable forecasting model to estimate the full-load power generation of a combined-cycle power plant (CCPP). The dataset used in this research is a subset of the publicly available UCI Machine Learning Repository. It contains 9568 items of data collected from a CCPP during its full load operation over a span of six years. To enhance the accuracy of power generation forecasting, a novel forecasting method based on Transformer encoders with deep neural networks (DNN) was proposed. The proposed model exploits the ability of the Transformer encoder to extract valuable information. Furthermore, bottleneck DNN blocks and residual connections are used in the DNN component. In this study, a series of experiments were conducted, and the performance of the proposed model was evaluated against other state-of-the-art machine learning models based on the CCPP dataset. The experimental results illustrated that using Transformer encoders along with DNN can considerably improve the accuracy of predicting CCPPs power generation (RMSE = 3.5370, MAE = 2.4033, MAPE = 0.5307%, and R2 = 0.9555).
20 October 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Pansharpening is an image fusion technique that integrates low-resolution multispectral (MS) images with high-resolution panchromatic (PAN) images to generate high-resolution multispectral images with enhanced spatial and spectral fidelity. Existing convolutional neural networks (CNNs) for pansharpening are limited in their ability to model long-range dependencies, whereas Transformer-based models often incur significant computational overhead. To efficiently capture global correlations among remote sensing image patches, we propose the global context injection Transformer (GCIFormer) as a core component in both the spatial and spectral branches. The proposed GCIFormer captures global dependencies at the window level and subsequently infuses them into the image representation at the patch level. Moreover, we design an axial fusion module (AFM) to capture long-range dependencies between the two modalities while mitigating information loss. AFM exploits axial features from different modalities by fusing their feature maps along respective axial directions. In addition, we introduce a PAN reconstruction loss to enhance the retention of high-frequency spatial details. Extensive experiments conducted on multiple benchmark datasets demonstrate that the proposed model outperforms state-of-the-art methods, showcasing its superior performance in pansharpening applications.
1 January 2026 in IEEE Transactions on Geoscience and Remote Sensing