To address the non-stationarity of wind-turbine bearing vibration signals, the need for collaborative modeling of one-dimensional sequential features and two-dimensional visual features, and the computational cost of sequence modeling, this study proposes an intelligent fault diagnosis model based on multimodal selective-gating Mamba, termed TF-Mamba. A progressive multimodal fusion framework of “feature extraction–CoSSM–explicit selection–global decision-making” is constructed. Specifically, CoSSM enables cross-modal feature interaction through shared-parameter Mamba encoding and reciprocal gating. Bi-SG performs adaptive explicit selection of fused features along the channel and spatial dimensions, and EnSSM models global sequential dependencies for final decision-making. Experiments are conducted on CWRU and the State Grid Wind-turbine Dataset (SGWD), with BMTM-Net and WCamba included as recent baselines. The main comparison yields accuracies of 99.07% and 96.85%, respectively. Across ten repeated runs, TF-Mamba achieves mean accuracies of 99.02±0.05% on CWRU and 96.75±0.12% on SGWD. Paired permutation tests against BMTM-Net give Holm-adjusted p=0.0156 for Accuracy and Macro-F1 on both datasets. In four leave-one-load-out CWRU evaluations, its mean accuracy across the held-out loads is 97.75%, with a mean Macro-F1 of 0.9757. A separate gate-substitution comparison gives Bi-SG an SGWD mean-accuracy advantage of 0.54 percentage points over CBAM. The network contains 2.81 million parameters and has a reported single-sample inference time of 4.68 ms on an RTX 4090. These results support the diagnostic value of progressive time–frequency collaboration and selective fusion under the evaluated dataset and load conditions.
Accurate short-term forecasting of the Market Clearing Price (MCP) is essential for the efficient operation of electricity markets with increasing renewable energy penetration. This challenge is particularly significant in India’s Green Day-Ahead Market (GDAM), where quarter-hourly trading and the variability of renewable generation introduce complex nonlinear price dynamics. This study presents a hybrid Wavelet–XGBoost framework for short-term MCP forecasting by integrating the discrete wavelet transform with the Extreme Gradient Boosting (XGBoost) algorithm, where the wavelet family and decomposition level are selected empirically through a systematic sensitivity analysis rather than assumed a priori. The wavelet decomposition separates the MCP series into multiple frequency components, enabling the model to capture both long-term trends and rapid price fluctuations while reducing the influence of noise. The proposed framework is evaluated using the 2024 GDAM dataset (32,159 valid quarter-hourly observations after data-quality screening) and compared with Naïve, Random Forest, LightGBM, conventional XGBoost, and an LSTM benchmark. On the held-out test period, the proposed model achieves an MAE of 258.6 Rs/MWh, RMSE of 523.7 Rs/MWh, RRMSE of 12.83% and MAPE of 5.65%, and attains the lowest RMSE and RRMSE of all evaluated models. Diebold–Mariano and Wilcoxon signed-rank tests confirm that the proposed model achieves significantly lower squared error than Naïve, Random Forest, XGBoost and LightGBM, and significantly lower typical-case absolute error than the same-architecture XGBoost and LightGBM baselines, while a tail-performance analysis quantifies the systematic underprediction of all models near the regulatory price cap. Furthermore, SHAP-based interpretability analysis identifies lagged MCP and bidding-related variables as the most influential factors affecting price prediction. The proposed framework provides a computationally efficient and interpretable forecasting approach for short-term MCP prediction in India’s GDAM, with particular strength in reducing larger forecasting errors and supporting bidding strategies, market scheduling and renewable energy integration.
Frequent extreme ice storms threaten the operational resilience of modern distribution networks. While coordinating Soft Open Points (SOPs) and Mobile De-icing Devices (MDIDs) offers a flexible spatiotemporal de-icing solution, jointly optimizing MDID traffic routing (discrete) and SOP power control (continuous) constitutes a complex bi-level problem. Traditional heuristics suffer from unacceptable online computational delays, and standard single-agent reinforcement learning struggles with gradient conflicts in mixed-action spaces. To address this, we propose a novel Goal-Conditioned Hierarchical Reinforcement Learning (HRL) framework. A Manager agent optimizes macroscopic MDID routing, while a Worker agent simultaneously regulates high-frequency SOP power flows to prevent cascading failures and electrical over-limits. Simulations on a 114-bus power-traffic coupled network demonstrate that HRL guarantees stable convergence, achieves millisecond-level real-time execution, and exhibits exceptional dynamic robustness against sudden traffic disturbances, significantly enhancing smart grid resilience.
To improve the utilisation rate of new energy sources and establish a new type of power system centred on renewable energy, this paper designs a capacity optimisation scheme for a hybrid energy storage system that integrates advanced adiabatic compressed air energy storage, electrochemical energy storage and hydrogen energy storage. This study employed K-means clustering combined with the cumulative median selection method to identify representative operational scenarios. A battery capacity degradation model accounting for cycle life and an advanced adiabatic compressed air energy storage system model were established, whilst incorporating mechanisms for green hydrogen production and revenue from hydrogen sales. A capacity optimisation model for the hybrid energy storage system was developed, with maximum daily comprehensive net revenue as the objective function, and a coordinated operation strategy for renewable energy and the multi-storage system was formulated. Case studies utilising wind power, photovoltaic and load data from north-western China validated the reliability and economic feasibility of the proposed system. Furthermore, the impact of different hydrogen transmission capacities and various energy storage operation strategies on system capacity allocation was quantitatively assessed.
This study presents a dynamic carbon footprint framework for Battery Energy Storage Systems (BESS) by integrating dispatch-optimised modelling with regional supply-chain data. Three applications — energy arbitrage, demand response, and peak shaving — are evaluated across six European countries to compare two functional units: installed capacity (kWhcapacity) and net delivered energy (kWhdelivered). Results demonstrate that BESS carbon footprints vary by over 300% depending on sizing, operational mode, and location. Operational impacts dominate in high-carbon grids, while production and recycling are primary drivers in low-carbon contexts. Critically, capacity-based metrics can incentivise oversizing, which often leads to underutilisation and poor performance when measured per unit of energy delivered. Sensitivity analysis further reveals that transportation contributes negligibly to the total life-cycle footprint. By providing a context-aware assessment methodology, this research supports informed procurement and regulatory compliance under the emerging EU Battery Regulation.
