Accurate forecasting of distributed photovoltaic power plays a crucial role in ensuring the safety and stability of an active distribution network. However, most existing research on distributed photovoltaic power forecasting exhibits certain limitations, including: 1) insufficient consideration of the dynamic correlations among power sites; and 2) absence of a training loss function capable of simultaneously aligning the amplitude and shape of forecasting values with the true values. Therefore, a dynamic graph network with a shape-amplitude loss function based distributed photovoltaic ultra-short-term power forecasting method is introduced. Firstly, a data-driven method is used to mine the dynamic correlation and the dynamic graph data can be generated to ensure the effective characterization of the correlation among distributed photovoltaics. Secondly, the dynamic graph network is constructed as the power forecasting model to realize the effective utilization of spatial-temporal correlation features. Then, the shape-amplitude loss function which combines the Dynamic Time Warping and Mean Square Error is used as the criterion of model training to ensure the consistency of the forecasting value and the real in situ shape and amplitude. Meanwhile, the dynamic graph network is combined with quantile regression and the quantile loss function is improved inspired from the idea of shape-amplitude. The forecasting performance of the introduced approach is demonstrated via a distributed photovoltaic power dataset in China.

In recent years, in response to national policies, the traditional power system has gradually evolved to the new power system in order to achieve low-carbon power generation. System operators, loaded virtual power plant and end users are the three main players in the current power market. However, most current studies take demand response as a constraint, and rarely take loaded virtual power plant as the main body. This will lead to the failure to integrate the user's demand response resources to participate in the electricity market bidding, so as to provide a feasible demand response scheme for each individual user. This paper takes the loaded virtual power plant as the main body, adopts the layered bidding strategy, establishes the demand response model of the loaded virtual power plant and the response contract to participate in the day-ahead market bidding, and uses the simulation to solve the mixed integer linear programming model, so as to obtain the optimal allocation scheme.

Wind farms located in remote areas have the problems of weak stability and low grid-connection strength due to the long electrical distance from large power grids and small Short-Circuit Ratio (SCR). Grid-Forming Control technology (GFM) has the characteristics of voltage support and active inertia, which can replace synchro to realize grid support and maintain the stability of power system. Therefore, GFM has broad development and application prospects in new energy power system. Based on this, this paper proposes a grid-forming energy storage system with hybrid ultracapacitors as the energy storage medium and virtual synchronous machine (VSG) control as the control strategy. The energy storage system is connected to the AC side of the wind farm, which can coordinate the energy storage to ensure that the system can smooth the frequency fluctuations and Reduce the Rate of Change of Frequency (ROCOF) when the wind resources fluctuate and the system switches the load. Raise the lowest frequency point; Provide frequency and voltage support when short circuit fault occurs in the system, prevent frequency collapse, and improve the success rate of Low Voltage Ride-Through (LVRT) of wind turbines. Improve grid-connected strength. Finally, based on the MATLAB/Simulink simulation platform, the good dynamic response ability of the energy storage system under the above-mentioned working conditions in each short-circuit ratio scenario is verified.

The widespread adoption of distributed solar photovoltaic technology has enabled distribution networks to manage the voltage levels at critical nodes through precise control of the reactive power supplied to the grid. This study has developed an adaptive voltage control scheme specifically designed for distribution networks integrated with large-scale photovoltaic resources. The scheme employs an advanced optimization search algorithm to automatically adjust the PI parameters of the voltage controller, ensuring flexible adaptation to real-time system fluctuations. The PI gain of the voltage controller minimizes the cost function that reflects the performance of the voltage controller through online updating. A distribution network model containing a 5MW PV system is constructed in MATLAB/Simulink to test the proposed self-adaptive voltage controller. The simulation results show that the proposed voltage controller can adapt well to load variations and grid faults, and improves the reliability and voltage stability of a power system, compared with the traditional PI controller.

In island microgrids, to tap the voltage control potential of different DGs, this paper presents a unified distributed cooperative voltage control approach for grid-following (GFL) and grid-forming (GFM) DGs. The proposed approach consists of a unified modeling method and a novel range-consensus-based distributed control algorithm. The modeling method employs power coupling and extended state observer (ESO) to unify different inner dynamics and control structures of DGs, making the voltage regulation of different DGs can be addressed by designing the same distributed controller. The proposed distributed algorithm introduces deviation-based states to converge the voltages to an expected range instead of a consensus state, making the algorithm is more in line with the voltage control standard of the grid code than the common state consensus ones. Besides, the local feedback designed in the algorithm improves the anti-interference ability of the system. The Lyapunov technique verifies the stability of the proposed method. Finally, the effectiveness of the proposed scheme is validated by several simulation results and a hardware-in-the-loop (HIL) test.

Mobile energy storage (MES) devices have excellent characteristics such as strong flexibility and wide application scenarios, and can play a role in shaving peaks and filling valleys in microgrids with high new energy penetration rates and improving power quality. Considering the plug-and-play application scenarios of MES devices in microgrid, this paper uses simulation to study the transient response of plug-and-play MES devices in the multi-inverter parallel connection system under the control of different control strategies, and the working situation after entering steady state. This paper analyzes and compares the situation of voltage source converter droop control, voltage source and current source hybrid droop control and plug-and-play under master-slave control, and gives an analysis of the advantages and disadvantages of various strategies and suggestions for whether they are suitable for mobile energy storage plug-and-play applications.

This paper introduces an improved deep Q-network (MDQN) algorithm, utilizing deep reinforcement learning, to tackle the real-time dispatch problem in virtual power plants. Extensive comparative experiments have consistently shown that the MDQN algorithm surpasses existing algorithms, such as DDPG, SAC, and TD3, in terms of both training efficiency and performance. Notably, the MDQN algorithm strictly adheres to operational constraints within the action space, ensuring the feasibility of scheduling plans during real-time operations. The experimental findings underline the promising potential of the MDQN algorithm in effectively addressing real-time optimization and dispatch challenges encountered in virtual power plants. These results contribute to the existing research in the field and hold significant implications for practical applications in the virtual power plant domain.

柔性直流配电网的低阻尼特性与开关器件的弱耐流能力之间的矛盾日益尖锐,给继电保护带来了严峻挑战.为此,提出一种基于Daniel趋势检验的柔性直流配电系统纵联保护方案,提取信号低频带数据信息,通过正、负极电压的Spearman秩相关系数状态值之和判别故障极;利用线路两侧电流的Spearman秩相关系数状态值乘积识别故障区域.仿真结果表明,所提保护方案能够可靠识别故障,且具备较强的耐受过渡电阻能力和抗噪声能力.

面向气-电耦合配网,提出一种基于信息间隙决策理论(information gap decision theory,IGDT)的多能协同优化调度方法.首先,采用二阶锥优化(second-order conic program,SOCP)松弛描述配电网与天然气管网的能量流特征,利用电/气/冷/热多能互补及协同转化作为运行灵活性提升的重要手段,以经济性最优为原则,建立气-电耦合配网调度的确定性优化模型.在此基础上,提出基于IGDT的多能协同滚动调度方法,生成时变的动态风险边界.最后,在基于IEEE 33节点配电网和比利时20节点天然气网络的多能耦合系统中对所提方法进行了测试验证.算例结果表明,所提调度方法可以充分发挥气-电耦合配网的多能互补优势,并克服了传统滚动调度方法对于可再生能源预测精度的依赖,有效提升了系统对不确定性风险的承受能力,进而在鲁棒性与运行经济性之间取得合理权衡.

Energy router is an important component in the energy internet, which can realize the interconnection, exchange and routing of energy and information among all links. With the deep integration of the cutting-edge digital information technologies such as power generation technology, power grid technology, big data and Internet of Thing, the requirements for the friendliness, flexibility, intelligence, reliability, and stability of applications to energy router are increasing. Taking scientific research institutes and universities as potential customers, this paper researches and develops a power electronics energy router with clear positioning, strong support and high efficiency that adapts to multi-scenarios and multi-modes. The energy router is able to achieve functions of displaying operation modes and verifying control principles to meet the needs of scientific research experiments and talents training.

在"双碳"目标下,智能发展、低碳转型、多元融合,商业模式日新月异,能源互联网正稳健发展.以储能+高效变流为基础,以数字化、信息化、智能服务为核心的移动储能,在构建未来城-镇-乡-村"移动能源互联网"中将承担重要角色.提出的新型智能化储能车能够实现能量双向流动,接入配电网后能够根据所采集到的支路电压、电流计算输出功率,自动调整自身输出功率,从而实现对台区负荷曲线的调整,减小高峰时段的供电压力.该智能化储能车结构紧凑、机动灵活,可在负荷高峰时段接入,其余时段退出配电网,不涉及基础建设.该新型智能化储能车作为可灵活调度资源,可应用于削峰填谷、不停电作业、台区动态增容、分布式发电及微电网、巨量电动汽车充电、用户侧用电备用等多种场景,助力高弹性配电网建设,具有广阔的发展前景.

不同资源聚合下的虚拟电厂调节能力各异,给虚拟电厂参与辅助服务决策带来困难,甚至影响电网安全可靠运行.针对此类情况,提出基于虚拟电厂可调空间统一模型的调峰优化运行策略.首先,针对虚拟电厂内部资源特性的不同,分别建立了半控型资源和可控型资源的出力模型.然后,提出一种虚拟电厂可调空间模型,可以针对不同资源聚合下的虚拟电厂调峰能力进行评估.最后,为验证可调空间模型的有效性,面向调峰辅助服务市场建立虚拟电厂日前优化调度模型,并采用动态鲁棒优化下的列约束生成算法对其进行求解.算例结果表明,所提虚拟电厂可调空间可以适应虚拟电厂不同聚合方式带来的变化,运行策略能够合理地平衡虚拟电厂参与调峰辅助服务时运行的安全性与经济性.

低压配电台区通过柔性直流互联,近期可实现台区动态增容、故障快速转供,提升供电可靠性及分布式电源接纳能力,远期通过低压交直流灵活组网,适应规模化多模式源、荷接入,实现源网荷储柔性高效互动目标.文中提出应对大规模分布式电源接入、终端电气化率提升、新基建建设需求以及季节性负荷波动等适用于配电台区柔性互联的典型场景,从规划建设、互联拓扑及网架结构设计、关键设备、运行控制与快速保护等五方面对台区柔性互联系统关键技术进行了综述,提出台区柔性互联系统的一次建设方案和二次物联架构,对台区柔性互联系统的高级应用和未来发展模式进行了展望.

能源互联网的发展推动了分布式资源配置规模与利用效率的逐步提高，规模化灵活资源虚拟电厂的构建将成为新型电力系统灵活性提升的关键研究领域。文中聚焦于规模化灵活资源可信聚合以及灵活资源辅助电力系统所需的信息-能量-价值耦合互动机理问题，提出了规模化灵活资源虚拟电厂的科学问题、技术路线与理论框架。在此基础上，分别从动态聚合、安全通信、协同调控、可信交易等4个领域，论述了具体的核心关键技术及其关联支撑关系。最后，围绕规模化灵活资源虚拟电厂的技术挑战与创新应用进行了总结和展望。

The integration of new energy generation, energy storage and flexible load parallel operation of the power terminal is an important part of DC distribution network optimal operation. In view of the limitations of the fixed energy optimization execution timescale, an adaptive timescale energy optimal predictive control in the DC power consumption zone with flexible dispatching units is proposed. The method is based on the adaptive adjustment of the energy output time scale of the power output ratio of the intermittent power supply in the DC power partition. At the same time, the lower-level comprehensive upper-level DC distribution network optimization instructions which are based on model free adaptive predictive control to achieve adaptive timescale energy optimal predictive control in the DC power consumption zone. The simulation results show that the proposed energy optimal predictive control method can realize the adaptive adjustment of the execution timescale, and under the condition that the flexible dispatching units are abundant, else and it can realize the fast and accurate tracking and execution of the energy optimal dispatching instruction of the superior DC distribution network.

Traditional prediction methods can’t solve the long time dependence problem in time series. In order to mine the effective Information contained in massive amounts of data and improve the prediction accuracy, a wind power prediction model based on the depth learning method Long Short-Term Memory model is proposed. Principal component analysis(PCA) is used to extract the features of input samples and predict the power of reference wind farm. Based on the Gaussian process regression method, an upscale model is established to predict the total regional wind power .Simulation results show that LSTM-GPR combinatorial method is adopted to improve the power prediction model of wind farm. Compared with BP neural network and support vector mac hine (SVM) model, the normalized root mean square errors were reduced by about 4% and 7% ,the prediction accuracy of regional wind power prediction model is improved. It is feasible and advanced to apply LSTM-GPR combinatorial method in the field of wind power prediction.

针对海岛微电网分布式电源和负荷的独有特性,以海岛系统运行经济性为目标,考虑新能源消纳因素,提出一种海岛微电网能量优化调度方法.首先在分析典型海洋能发电出力特性的基础上,建立了含风、光、柴、储、波浪能、潮汐能以及可控负荷的海岛微网能量优化调度模型;而后面向日前和日内2个时间尺度,采用CPLEX完成混合整数规划,利用日前调度计划与日内实时滚动推演结果比对,依据日内滚动计算结果修正日前调度计划,实现当日微网能量调度全局最优.算例分析表明,由于利用了日前和日内相结合的滚动计算方法,使对岛上源、荷的预测精度逐级提高,该优化调度方法在兼顾海岛微电网功率平衡和可再生能源利用率的同时,最大限度降低了海岛微电网运维成本.

以含风光储的交直流混合微电网为研究对象,通过聚类分析方法对气象条件进行分类,搭建风光储出力概率模型并进行序列化.提出一种序列运算方法计算交直流混合微电网用电不足期望值,从而迭代求得可信容量,针对风光装机容量配比以及直流负荷敏感度对可信容量的影响进行了分析,并将计算结果与蒙特卡洛法进行了对比,验证该文所提方法的正确性与有效性.

Aiming at the problem of voltage rise caused by the high-penetration distributed photovoltaic cluster connected to the distribution network, this paper proposes a bottom-up, trigger-type hierarchical voltage control strategy of distributed photovoltaic cluster. When the voltage is exceeded the limit, the local voltage control layer is based on the photovoltaic inverter Q(U) segmented droop control strategy to achieve rapid response to the voltage change of the point of connection. If the voltage is still exceeded the limit after local voltage control, in the cluster voltage control layer, in accordance with the priority to use the reactive power adjustment capability of distributed photovoltaic, and the principle of minimizing the reduction of photovoltaic active output, adjust the active and reactive power output of distributed photovoltaic based on the voltage sensitivity analysis method to maximize the consumption of distributed photovoltaic while ensuring the safe operation of the power grid. Finally, simulations verify the effectiveness of the proposed control strategy.