To address the issue of significant unpredictability and intermittent nature of renewable energy sources, particularly wind and solar power, this paper introduces a novel optimization model based on online reinforcement learning. Initially, an energy management optimization model is designed to achieve plan adherence and minimize energy storage (ES) operation costs, taking into account the inherent challenges of wind power-photovoltaic energy storage systems (WPESS). An online reinforcement learning framework is employed, which defines various state variables, action variables, and reward functions for the energy management optimization model. The state-action-reward-state-action (SARSA) algorithm is applied to learn the joint scheduling strategy for the microgrid system, utilizing its iterative exploration mechanisms and interaction with the environment. This strategy aims to accomplish the goals of effective power tracking and reduction of storage charging and discharging. The proposed method's effectiveness is validated using a residential community with electric vehicle (EV) charging loads as a test case. Numerical analyses demonstrate that the approach is not reliant on traditional mathematical models and adapts well to the uncertainties and complex constraints of the WPESS, maintaining a low-cost requirement while achieving computational efficiency significantly higher than that of the model predictive control (MPC) and deep Q-network (DQN) algorithm.

A distributionally robust scheduling strategy is proposed to address the complex benefit allocation problem in regional integrated energy systems (RIESs) with multiple stakeholders. A two-level Stackelberg game model is established, with the RIES operator as the leader and the users as the followers. It considers the interests of the RIES operator and demand response users in energy trading. The leader optimizes time-of-use (TOU) energy prices to minimize costs while users formulate response plans based on prices. A two-stage distributionally robust game model with comprehensive norm constraints, which encompasses the two-level Stackelberg game model in the day-ahead scheduling stage, is constructed to manage wind power uncertainty. Karush-Kuhn-Tucker (KKT) conditions transform the two-level Stackelberg game model into a single-level robust optimization model, which is then solved using column and constraint generation (C&CG). Numerical results demonstrate the effectiveness of the proposed strategy in balancing stakeholders' interests and mitigating wind power risks.

This paper addresses the challenge of maximizing power capture from new energy sources, including coal, wind, solar, and hydroelectric power, which often lack sufficient inertia support. This deficiency can lead to frequency instability and cascading failures within the power system. A cooperative optimization model for the start‐up of multiple power supplies, designed to enhance the integration of new energy sources while maintaining system stability is proposed. The model incorporates primary frequency modulation and the intrinsic inertia support capabilities of self‐synchronous voltage source field stations, considering dynamic frequency constraints. Additionally, it employs new energy units with primary frequency modulation to provide inertia support during curtailment, particularly when conventional units cannot meet frequency standards due to existing constraints. Extensive simulations and comparative analyses demonstrate that the proposed model improves new energy utilization by up to 37.5% and reduces operational costs by approximately 16%, enhancing overall operational efficiency in high energy consumption scenarios.

Electric power audit text classification is one of the important research problem in electric power systems. Recently, kinds of automatic classification methods for these texts based on machine learning or deep learning models have been applied. At present, the development of computing technology makes “pre-training and fine-tuning” the newest paradigm of text classification, which achieves better results than previous fully-supervised models. Based on pre-training theory, domain-related pre-training tasks can enhance the performance of downstream tasks in the specific domain. However, existing pre-training models usually use general corpus for pre-training, and do not use texts related to the field of electric power, especially electric power audit texts. This results in that the model does not learn too much electric-power-related morphology or semantics in the pre-training stage, so that less information can be used in the fine-tuning stage. Based on the research status, in this paper, we propose EPAT-BERT, a BERT-based model pre-trained by two-granularity pre-training tasks: word-level masked language model and entity-level masked language model. These two tasks predict word and entity in electric-power-related texts to learn abundant morphology and semantics about electric power. We then fine-tune EPAT-BERT for electric power audit text classification task. The experimental results show that, compared with fully supervised machine learning models, neural network models, and general pre-trained language models, EPAT-BERT can significantly outperform existing models in a variety of evaluation metrics. Therefore, EPAT-BERT can be further applied to electric power audit text classification. We also conduct ablation studies to prove the effectiveness of each component in EPAT-BERT to further illustrate our motivations.

In the face of the dual crisis of energy shortages and global warming, the vigorous development of renewable energy represented by wind-solar energy is a significant approach towards achieving energy transition, carbon peaking, and carbon neutrality goals. Targeting the park-level integrated energy system (PIES) with high penetration of wind-solar energy, we propose a day-ahead dispatching strategy that takes into account the flexible supply and the reward-punishment ladder-type carbon trading mechanism (RPLTCTM). Firstly, RPLTCTM and carbon capture equipment (CCE) are considered in the dispatching model, and the mechanism of coordinated operation of CCE and RPLTCTM is explored to further improve the system’s ability to restrain carbon emissions. Secondly, power-based flexibility indicators (PFIs) are adopted to quantitatively evaluate the flexibility supply, and based on the load demand response characteristics, the dispatchable resources on the load side are guided to improve the system’s operation flexibility. On this basis, a multi-objective optimal dispatching model that takes into account the carbon emission cost, energy cost, and flexibility supply are constructed, and the original problem is transformed into a mixed-integer single-objective linear problem through mathematical equivalence and flexibility cost. Finally, simulation examples validate that the economy, flexibility, and low-carbon level of the dispatching plan can be synergistically improved by the proposed strategy.

Electric vehicles (EVs) and small photovoltaic (PV) installations advance residential power grids by lowering charging costs and fostering eco‐friendly operations. Yet, the variable nature of EV charging presents challenges to grid reliability. This research introduces a Monte Carlo‐based simulation for predicting EV charging loads and a systematic charging method that integrates a ‘green electricity’ pricing scheme with a joint optimization model for PV and EV management. By applying an improved ant lion optimizer (IALO) algorithm enriched with differential evolution features, an optimization strategy that markedly enhances grid performance is devised. In a park scenario, this ‘green electricity’ model reduced the mean square error of EV charging load by 11.82%, smoothed the power load curve, and improved grid stability. When compared with particle swarm optimization (PSO) and grey wolf optimizer (GWO) algorithms, the IALO algorithm boosted overall revenue by 16.8% and 12.8%, increased PV utilization by 162.3% and 37.1%, and significantly cut carbon emissions by 159.6% and 31.6%, respectively. These outcomes affirm the financial, environmental, and functional benefits of our proposed approach.