In the context of global energy transition, photovoltaic (PV) power plays a key role in mitigating climate change. Ultra-short-term PV power prediction is vital for grid dispatch and operational security. However, PV output is highly sensitive to weather, particularly during severe conditions. The performance of existing models in these non-stationary environments is hindered by scarce samples, noise in observations, and nonlinear coupling among meteorological variables. To address these issues, this study proposes an end-to-end multimodal prediction framework: SeqVAEWD-TimeXer-ACNetLG (STALG). SeqVAEWD performs generative enhancement of severe weather samples in latent space, paired with Wavelet Denoising for noise suppression. The ACNet adapter captures multi-scale, high-order meteorological correlations, while TimeXer-ACNetLG combined with LightGBM ensemble enables robust prediction of complex temporal features. Experiments on mountain and plain PV stations show STALG reduces MSE by 38.3\% and MAE by 26.2\%, with RMSE optimized by up to 60\% under severe weather. Prediction accuracy stably exceeds 90\%, meeting State Grid specifications to support stable PV grid integration under harsh conditions.

Stability and sustainability are the operational bottom lines of modern power grids. However, the inherent volatility and non-stationarity of wind energy, particularly in complex terrains, severely threaten power grid stability. To address this challenge, we propose an end-to-end architecture named the Adaptive Multi-scale Routing Wind Power forecasting (AMR-Wind) framework. The framework is principally composed of three sequential modules: an Adaptive Frequency Disentanglement Module (AFDM), an inverted Transformer (iTransformer), and a Scale-Routing Gated Recurrent Unit (SRGRU). The AFDM utilizes a differentiable filter bank to dynamically disentangle complex spectral signatures and mitigate mode mixing. The iTransformer is employed to effectively capture the complex multivariate dependencies between these disentangled modes and exogenous meteorological features. The SRGRU utilizes hierarchical temporal routing to synchronize localized high-frequency ramp events with macroscopic evolutionary trends. Comprehensive evaluations across four diverse wind farms demonstrate that AMR-Wind reduces the RMSE by an average of 8.4% and improves the $R^2$ by at least 1.0% compared to state-of-the-art baselines. Ablation studies further confirm the modules’ strong synergistic effects, yielding a 7.6% reduction in forecasting errors. This framework reduces the error in wind energy prediction, providing a reliable tool for the stability and sustainability of the power grid.