In order to achieve zero carbon emissions, the decarbonization and cleanliness of the power system becomes more and more important. With the increase of renewable energy integration such as wind and photovoltaic (PV) power, the source-side uncertainty of the power systems has increased significantly, and traditional optimal power flow calculation methods encounter challenges to adapt to the demand of power system operation with large integration of new energy resources. In this paper, effects of wind and PV power output uncertainty are analyzed and represented by probabilistic model first. On this basis, the multi-objective optimization model for optimal power flow is constructed and solved with the modified non-dominated sorting genetic algorithm-II (NSGA-II). In the multi-objective optimization power flow model, the uncertainty of source-side wind and PV power generation, and the constraints of traditional generation and reactive power compensation are considered. Validations on IEEE 14 system show effectiveness of the proposed modified NSGA-II based optimal power flow method in reducing both economic generation cost and electric network power loss.

The significant increase in the proportion of renewable energy sources (RESs) has elevated risks of extreme ramp events and frequency instability in power systems. In recent years, frequency stability events have occurred in several countries/regions worldwide due to flexibility deficiencies. Generation flexibility has emerged as a critical factor influencing the frequency stability of power systems. This paper proposes a domain of attraction (DOA)-based quantitative method to assess the frequency stability region of power systems with a high proportion of RESs, considering generation flexibility constraints. First, ramp rate is adopted as the core indicator to characterize generation flexibility within automatic generation control (AGC) timescale, through which a nonlinear AGC model with rate saturation constraints is established. Second, the concept of DOA is introduced to define the stability region of the nonlinear AGC. Third, a quadratic Lyapunov-based estimation method is employed to quantitatively analyze the DOA of the nonlinear AGC at different generation flexibility levels. Simulation results demonstrate that increased generation flexibility expands the estimated DOA of the nonlinear AGC, whereas generation flexibility deficiency induces AGC instability. Moreover, state trajectory and time-domain simulation verify that the proposed estimation method accurately represents the stability region of the nonlinear AGC.

Data-driven methods are widely recognized and generate conducive results for online transient stability assessment. However, the tedious and time-consuming process of sample collection is often overlooked. The functioning of power systems involves repetitive sample collection due to the constant variations occurring in the operation mode, thereby highlighting the importance of collection efficiency. As a means to achieve high sample collection efficiency following the operation mode change, we propose a novel instance-transfer method based on compression and matching strategy, which facilitates the direct acquisition of useful previous samples, used for creating the new sample base. Additionally, we present a hybrid model to ensure rationality in the process of sample similarity comparison and selection, where features of analytical modeling with special significance are introduced into data-driven methods. At the same time, a data-driven method can also be integrated in the hybrid model to achieve rapid error correction of analytical models, enabling fast and accurate post-disturbance transient stability assessment. As a paradigm, we consider a scheme for online critical clearing time estimation, where integrated extended equal area criterion and extreme learning machine are employed as analytical model part and data-driven error correction model part, respectively. Derived results validate the credible efficacy of the proposed method.

随着交直流混联电网规模的扩大与电力电子化设备的大规模并网,以新能源为主体的新型电力系统的动态特性愈加复杂.物理模型的机理可解释性与数据模型的特性拟合能力具有很强的互补性.如何将融合模型的构建从定性分析向定量分析提升亟待深入研究.文中基于电力系统中数据方法与物理方法的特点,针对4种典型数据-物理融合模型分析了其相对应的应用场景;以并联模式为研究对象,分别对比分析了并联模式与单一物理模型和单一数据模型的泛化误差,并提出了融合模型参数的选取方法;推导了并联模式下融合模型的泛化误差上限,并提出了改进融合模型性能的可行性建议;最后,结合暂态功角稳定问题验证了所提假设与结论的合理性.

The critical clearing time (CCT) is one of the most important indexes for large-disturbance rotor angle stability margin evaluation. In practice, model-driven methods are usually realized based on simplified models to ease the computational burden, but the accuracy is sacrificed. To solve this problem, a data-driven method is adopted in this paper for fast error correction of a model-driven method, creating an integrated method. Both a reliable accuracy and an acceptable computation speed can be achieved with this integrated method. Meanwhile, involvement of model-driven method helps enhance robustness of the integrated method to training sample insufficiency, measurement error and power system scale. In addition, the data-driven method is further transformed on the basis of a cost-sensitive approach where the error tolerance for different actual CCT values should be differentiated during the training process instead of being treated equally in the common data-driven method. To mitigate the negative effect caused by such transformations, an ensemble learning structure is also constructed. In this paper, an integrated extended equal-area criterion (IEEAC) and an extreme learning machine (ELM) are applied as model-driven and data-driven methods, respectively. A genetic algorithm (GA) is used in the ensemble learning structure construction. Validations show that the proposed integrated method with the transformed data-driven method can improve the CCT prediction accuracy and avoid the polarization of the error distribution.

Critical clearing time (CCT) is one of the most important indexes for large-disturbance rotor angle stability margin evaluation. In actual operation, model-driven methods, e.g., integrated extended equal area criterion (IEEAC) methods, are proven to be good methods to calculate CCT. Nevertheless, classical generator models are always used to reduce computation burden, where differences between classical models and detailed models could make CCT predictions useless. To correct error caused by model differences, data-driven method is considered and integrated with classical generator model based IEEAC, making an integrated method. In this paper, extreme learning machine (ELM) is used as the data-driven correction part. Further, cost-sensitive ELM model is constructed by substituting common error summation part with weighted error rate summation part in training objective. Prediction error for cases with small actual value is expected to be lower than that with large actual value by this transformation. Finally, validations on WSCC 9 and New England 39 bus system show effectiveness of proposed integrated method in correcting error caused by model difference and changing error distribution.

Identifying correct model parameters is important for actual power system operation and control. Though existing gradient decent method shows good timeliness, it would converge to wrong results because of inevitable linearization process when applied for strongly nonlinear models. To make up this shortcoming, an estimation and correction combined method is proposed in this paper, by which the gradient method is expected to have better initial values for avoiding the local optimum trap. In the estimation process, pattern matching is utilized based on the constructed post-disturbance trajectory based typical parameters matching database. To construct the typical parameters matching database, correlation coefficient based forward and backward cluster method is applied, with which the typical parameters matching database can be updated conveniently and quickly. In the correction process, a novel comprehensive evaluation index is put forward for gradient decent method to evaluate parameter identification effects reasonably. Finally, the proposed combined parameter identification method is verified with standard high voltage direct current (HVDC) models together with parameter sensitivity analysis, and results show effectiveness.

In traditional risk assessment method, failure of control actions is always neglected due to its low occurrence probability. However, failure of control actions usually leads to power system cascading outages, where severe consequence has an impact on risk assessment results. Hence, it is necessary to consider control actions failure in power system risk assessment. In this paper, relay protection failure is chosen as an example of control actions failure and considered in power system risk assessment. In the risk assessment procedure, the probability evaluation of relay protection failure and system risk computation method are concerned. Firstly, a method for relay protection failure probability evaluation is presented, where intrinsic property and operating state of relay protection devices are combined. Secondly, an improved state space partitioning (SSP) method, suitable for relay protection failure involved power system cascading outages risk assessment, is put forward. The improved SSP method transforms sampling strategy of sampling process in common SSP method, utilizing enumerated state help low-probability event sampling. It enables to avoid the low-efficiency problem caused by extreme low-probability event in common SSP method based risk assessment. Finally, relay protection failure probability evaluation method and the improved SSP method are validated with simulations and results show effectiveness in risk assessment accuracy and efficiency.

知识驱动方法与数据驱动方法是指导工程人员研究电力系统的两大方法论.然而随着电网规模日趋扩大、时变因素日益增多和非线性逐渐增强,基于知识驱动的机理模型方法或基于数据驱动的经验模型方法在电力系统相关应用中将面临更多的挑战.充分利用数据驱动方法与知识驱动方法的互补特性,将二者联合,有望实现应用中综合性能的提升.该文对各研究领域中的数据与知识联合驱动方法进行了整理归纳,进而结合电力系统的特点和需求,梳理了数据与知识联合驱动的典型应用方式,并针对潜在的应用场景进行了详细讨论.最后,在电力系统应用场景中测试验证了数据与知识联合驱动方法的应用效果.

Quickly identifying faulty sections is tremendously important for power systems, yet challenging due to handling the variations of complex alarm patterns. Existing works have focused on finding fault section clues solely from alarm information (and ignoring power system topology information). So they are only sensitive to alarms from power systems with pre-assumed topology structures, and encounter difficulties when a system's topology changes. To adapt to unknown or varying system topologies, here we present a Topology-Adaptive Deep Model (TADM) for power system multifault diagnosis. TADM mines the underlying mapping from alarm and topology information to each section's fault status. It consists of a deep iterative network (DIN), a one-layer fully connected network (FCN), and section-wise multifault diagnosis (SWMD) subnetwork. TADM first models a fault power system as a graph, from which DIN iteratively integrates the alarm and topology information in the region from each node to its T -hop neighbors, and learns their local correlation. Limited to T 's size, FCN then combines all local correlations to determine the global correlation between alarm and topology information across the entire power system. To implement multifault diagnosis, learned local and global correlations serve as topology-related fault representations for input as an SWMD (to predict all sections' fault states one by one). A comprehensive experimental study demonstrates that TADM outperforms state-of-the-art models in both multifault diagnosis and adapting to system topologies. The source code of the TADM is available onlline 1 .

现代智能电网出现了电力电子化、信息物理融合和大电网复杂互联等新特征,从而对电力系统暂态问题的分析与控制方法产生了极大影响。人工智能(artificialintelligence,AI)在解决数据问题中的优势与暂态问题特点匹配程度高。该文从信息、机理、仿真、分析和控制等角度分析了电力系统暂态问题出现的新特点,归纳总结了现有将AI用于分析电力系统暂态问题时的研究成果,指出了研究中仍需解决的问题,探讨了在数据获取、特征提取和算法应用等方面的若干改进思路,并对AI应用于暂态稳定问题的研究现状进行总结。

Determining the optimal islanding solution is an essential issue of controlled islanding. However, the problem is commonly simplified as a single-objective optimization problem in existing literature, which is not able to ensure the stable operation of sub-systems after separation due to the development of renewable energy and high voltage direct current (HVDC) transmission. Therefore, this paper presents an improved islanding strategy considering renewable energy and HVDC. In the first step, initial splitting surface is obtained based on traditional single-optimization model, which may not be the optimal solution for the current complex system, and then similar feasible solutions are selected. In the second step, multi-infeed short circuit ratio (MSCR) or renewable energy penetration of sub-systems created by feasible solutions in the candidate space are verified to determine the final islanding solution. Simulation results in 39-bus test system indicate that the strategy is effective to improve frequency and voltage stability of sub-systems.

针对新闻推荐系统中用户兴趣模型构建与用户兴趣漂移问题,提出了一种面向新闻推荐的用户兴趣模型构建与更新方法.首先采用向量空间模型与bisecting K-means聚类算法构建了原始用户兴趣模型;然后以艾宾浩斯遗忘曲线为基础构造了遗忘函数,并以此对用户兴趣模型进行时间加权,从而达到对用户兴趣模型更新的目的.实验以基于用户的协同过滤推荐、基于物品的协同过滤推荐为baseline,实验结果表明所构建的原始用户兴趣模型推荐性能更优,在F值上提升了4％,更新后的模型与原始模型相比F值提高了1.3％.

With increase of practical power system complexity, power system online stability assessment and control is more and more important. Application of the traditional model-driven methods is always limited by contradiction between accuracy and efficiency, while data-driven methods demonstrate strong abilities for the online decision-making support with advancement of various data mining techniques. Instead of direct application of data-driven methods in the power system, this paper first discusses feasible integration approaches for the model-driven and data-driven methods based on the existing achievements, and then, proposes to integrate both methods for the power system online frequency stability assessment and control. The integrated method consists of frequency dynamics prediction and load shedding procedure. In frequency dynamics prediction procedure, integration of system frequency response (SFR) model and the extreme learning machine (ELM)-based learning model is applied, where basic physical causality is kept in the SFR model and ELM is used to fit and correct error of the SFR. The ELM also plays a part in load shedding prediction model construction by digging out mapping relationship from samples. Finally, the proposed prediction and control scheme for the frequency stability is verified by simulations on WSCC 9-bus, New England 39-bus, and NPCC 140-bus system. Results show that the reliability, time efficiency, and accuracy are enhanced with the proposed method.

Controlled islanding is an effective remedy to prevent large-area blackouts in a power system under a critically unstable condition. When and where to separate the power system are the essential issues facing controlled islanding. In this paper, both tasks are studied to ensure higher time efficiency and a better post-splitting restoration effect. A transient stability assessment model based on extreme learning machine (ELM) and trajectory fitting (TF) is constructed to determine the start-up criterion for controlled islanding. This model works through prompt stability status judgment with ELM and selective result amendment with TF to ensure that the assessment is both efficient and accurate. Moreover, a splitting surface searching algorithm, subject to minimal power disruption, is proposed for determination of the controlled islanding implementing locations. A highlight of this algorithm is a proposed modified electrical distance concept defined by active power magnitude and reactance on transmission lines that realize a computational burden reduction without feasible solution loss. Finally, the simulation results and comparison analysis based on the New England 39-bus test system validates the implementation effects of the proposed controlled islanding strategy.

Due to an increase in intermittent renewable energy penetration, the mechanical inertia of power systems has gradually decreased, threatening system frequency stability. As one effective solution to this problem, demand response (DR) technologies, which enable large-scale residential loads to regulate system frequencies via load aggregators, have been widely used. Aggregated loads, while treated as a whole from the perspective of the system operator, present some aggregate characteristics related to the specificities of individual loads within each aggregator. To construct such a relationship, DR aggregate characteristics based on load heterogeneity in response latency and user comfort requirements are analysed in this paper. As a result, an aggregate model for available response capacity evaluation is constructed. To realize the effective utilization of DR resources, feasible control strategies based on aggregate characteristics are discussed and compared with an emphasis on frequency regulating effects. Furthermore, a two stage-based DR section related to control strategies is introduced into a conventional system frequency response model for post-disturbance frequency nadir prediction. Finally, simulations are performed to verify the validity and accuracy of the proposed method.

Accurate and prompt transient stability prediction is one of the effective ways to reduce the risk of blackout or cascading failures. In an effort to achieve improvements in time efficiency and prediction accuracy, a new transient stability prediction method combining trajectory fitting (TF) and extreme learning machine (ELM) based on two-stage process, named hybrid method, is proposed here. ELM-based method is implemented in central station to ensure the time efficiency, while TF-based method is adopted in local station to guarantee the accuracy. Furthermore, data corruption is taken into consideration to assure the robustness of the proposed algorithm. The hybrid method is validated with the New England 39-bus test system and the simulation results indicate its effectiveness and reliability.

电力系统频率态势在线预测有助于快速准确地制定扰动后的控制措施,降低事故影响。单一依靠物理或数据模型的频率态势在线预测方法在实际应用中存在计算速度与精度之间的矛盾。采用基于物理—数据融合建模思路,提出一种频率态势在线预测方法:将暂态频率影响因素划分为关键因素和非关键因素,对关键因素采用系统频率响应模型以保留电气信息间因果联系,对非关键因素采用基于极限学习机的误差校正模型以表征关联关系。该方法具有样本依赖性小、通信容错率高、计算效率受系统规模影响小的特点。通过标准测试系统仿真验证,表明所述方法能够快速、准确地预测系统受扰后的频率态势特征。

Because of the differences of grid network architecture, generator parameters, and load characteristics, frequency and voltage at grid nodes show certain spatial distribution characteristics when faults happen. Therefore, the same stability control measures may result in different control effects if applied at different nodes. Meanwhile, in order to improve the control effect and reduce stable control cost, the impacts of the interactions between strongly coupled factors, such as frequency, voltage, active power, reactive power and other quantities, on the effects of stability control measures need to be considered. A new optimization control strategy of under‐frequency load shedding/under‐voltage load shedding is proposed in this paper based on the load control sensitivity index. The method takes into consideration the spatial distribution characteristics of frequency and voltage, as well as the real‐time interaction characteristics of the coupled electrical quantities. Compared with the traditional load‐shedding scheme, the proposed scheme can achieve better control results. The effectiveness of the proposed method is verified through a time‐domain simulation of two actual provincial power grids. © 2018 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

With the development of ultra-high-voltage direct current (UHVDC) transmission systems, the receiving AC systems face the problems of power flow transferring, integrating capability and voltage stability. A novel UHVDC hierarchical connection (HC) mode has been proposed and adopted in several real projects in China. The more complex electric coupling relationship of the entire systems made the power interaction characteristics different from those of previous multi-infeed HVDC systems. Thus, it is important to analyze the power stability of such systems in a systematic manner to avoid adverse interactions. In this study, the equivalent model of HC-UHVDC systems is established and several evaluation indices concerning power stability are discussed and improved. Furthermore, the power integration capability and reactive power interaction are evaluated based on real data. The results of this research can be used to provide technical support to real HC-UHVDC projects. (C) 2017 Elsevier B.V. All rights reserved.