在电力现货市场环境下,售电公司需要面向市场电价及用户负荷的双重不确定性,在日前申报的环节易造成额外购电成本。然而现有基于条件风险价值(conditional value at risk, CVaR)等随机优化方法的购电方案与风险管理策略中存在等概率缩减关键场景与主观进行置信度选值的问题,为此基于传统的中性风险模型及CVaR优化模型,引入基于K-means的场景聚类缩减方法,提出基于外推内插法的置信度选值优化方法,综合形成改进CVaR的售电公司日前申报优化模型及其求解策略。仿真算例结果验证了改进CVaR优化模型能有效降低售电公司的综合购电成本及潜在风险损失,并探究了在不同的风险厌恶程度与市场波动程度的情况下对日前申报优化策略的影响,体现了改进优化申报策略的适用性与鲁棒性。

In the electricity spot market, electricity retailers face dual uncertainties that arise from market electricity prices and user loads. The day-ahead bidding process can incur additional purchasing costs owing to these uncertainties. However, existing stochastic optimization methods for electricity purchasing strategies and risk management, such as the conditional value at risk (CVaR), suffer from problems related to equiprobable reduction in key scenarios and subjective confidence level selection. To address these challenges, this study introduces a scenario reduction method based on k-means and a confidence level optimization method based on extrapolation-interpolation, proposing an improved CVaR day-ahead bidding optimization model and its solution strategy based on the traditional neutral risk model and CVaR optimization model. The simulation results validate that the improved CVaR optimization model effectively reduces the overall purchasing costs and potential risk losses for the electricity retailer. This study explores the impact of the day-ahead bidding optimization strategy under different levels of risk aversion and market volatility, demonstrating the applicability and robustness of the improved optimization strategy.

相比传统电网,韧性城市电网展现出了出色的适应多种扰动和灾害的能力,但其复杂性也使得韧性城市电网风险预警面临更大的挑战,亟需大数据和机器学习等先进技术的引入。首先,构建韧性城市电网风险评估指标体系,采用主客观结合的综合赋权法对指标赋权,通过大数据技术获取的实时数据流得到韧性城市电网风险评估指标的动态权重;然后,构建韧性城市电网风险评估标准云,计算韧性城市电网风险等级隶属度,确定风险等级;最后,基于随机森林构建韧性城市电网风险预警模型,并进行算例分析,通过与其他模型对比,发现所构建的模型表现出高精度的特征。所建模型具有较好的风险预警效果,从而能够及时采取有效风险管控措施,保障韧性城市电网稳定运行。

Resilient urban power grids, while showcasing remarkable adaptability to diverse disturbances and disasters, pose significant challenges in risk warning due to their complexity. This underscores the necessity of integrating advanced technologies such as big data and machine learning. This study proposes a novel approach to resolve these issues. First, a resilient urban power grid risk assessment index system was established, employing a comprehensive weighted approach that combined subjective and objective factors to weigh the indicators. Leveraging real-time data flow obtained through big data technology, dynamic weights for risk assessment indicators were determined. Subsequently, a resilient urban power grid risk assessment standard cloud was developed, and the membership degree of the resilient urban power grid risk level was computed to ascertain the risk level. Finally, a resilient urban power grid risk warning model was formulated using random forest, and a thorough numerical analysis was conducted. Compared with other models, the constructed model exhibited high precision characteristics. The findings demonstrated that the developed model exerted a substantial risk-warning effect, enabling timely implementation of effective risk control measures to ensure the stable operation of resilient urban power grids.

In the smart grid paradigm, growing integration of large-scale intermittent renewable energies has introduced significant uncertainties to the operations of an electric power system. This makes real-time dynamic security assessment (DSA) a necessity to enable enhanced situational-awareness against the risk of blackouts. Conventional DSA methods are mainly based on the time-domain simulation, which are insufficiently fast and knowledge-poor. In recent years, the intelligent system (IS) strategy has been identified as a promising approach to facilitate real-time DSA. While previous works mainly concentrate on the rotor angle stability, this paper focuses on another yet increasingly important dynamic insecurity phenomenon-the short-term voltage instability, which involves fast and complex load dynamics. The problem is modeled as a classification subproblem for transient voltage collapse and a prediction subproblem for unacceptable dynamic voltage deviation. A hierarchical IS is developed to address the two subproblems sequentially. The IS is based on ensemble learning of random-weights neural networks and is implemented in an offline training, a real-time application, and an online updating pattern. The simulation results on the New England 39-bus system verify its superiority in both learning speed and accuracy over some state-of-the-art learning algorithms.

在电力现货市场运营中普遍存在发电侧市场力形式多、影响大、防范难、监管难的问题,给电力市场运营带来巨大风险和挑战,因此对电力市场力风险有效防范研究具有重要现实意义。基于激励相容、个体理性及社会福利最大化的设计目标,提出了一种将市场风险防范措施嵌入到现货市场出清算法的市场力风险防范方法,综合考虑了市场主体的当日市场交易行为和历史行为表现,通过在市场出清环节建立违规惩罚成本和系统价格权拍卖结算机制,降低违规行为市场主体出清的优先级,有效抑制市场主体行使市场力的意图,从而将风险防范前置到出清环节,实现防范风险于未然,有效解决市场力风险发生后的处置难、影响恢复难问题。最后,利用IEEE 30节点系统算例,验证了所提方法的可行性和有效性。

In the operation of the electricity spot market, there are generally many forms of market power on the power generation side, great influence, difficult prevention, and difficulty in supervision, which bring huge risks and challenges to the operation of the electricity market. Aiming at the design goals of incentive compatibility, individual rationality and social welfare maximization, this paper proposes a market power risk prevention method that embeds market risk prevention measures into the spot market clearing algorithm. The market main body of the same-day market trading behavior and historical behavior performance is comprehensively considered. Through the establishment of a violation penalty cost and a systematic price right auction settlement mechanism in the market clearing link, the priority of clearing market players for violations is reduced, and the intention of market players to exercise market power is effectively restrained, so that risk prevention will be put before risk prevention. It is placed in the clearing link to prevent risks before they occur, and effectively solve the problems of difficult disposal and impact recovery after market power risks occur. Finally, the IEEE 30-node system example is used to verify the feasibility and effectiveness of the method.

Despite significant successes achieved in knowledge discovery, traditional machine learning methods may fail to obtain satisfactory performances when dealing with complex data, such as imbalanced, high-dimensional, noisy data, etc. The reason behind is that it is difficult for these methods to capture multiple characteristics and underlying structure of data. In this context, it becomes an important topic in the data mining field that how to effectively construct an efficient knowledge discovery and mining model. Ensemble learning, as one research hot spot, aims to integrate data fusion, data modeling, and data mining into a unified framework. Specifically, ensemble learning firstly extracts a set of features with a variety of transformations. Based on these learned features, multiple learning algorithms are utilized to produce weak predictive results. Finally, ensemble learning fuses the informative knowledge from the above results obtained to achieve knowledge discovery and better predictive performance via voting schemes in an adaptive way. In this paper, we review the research progress of the mainstream approaches of ensemble learning and classify them based on different characteristics. In addition, we present challenges and possible research directions for each mainstream approach of ensemble learning, and we also give an extra introduction for the combination of ensemble learning with other machine learning hot spots such as deep learning, reinforcement learning, etc.