在高比例新能源的孤岛微电网中,风电的间歇性与波动性会增加微电网运行调度的不确定性,严重情况下可能对系统安全稳定运行造成影响.为此,提出一种计及用户舒适度需求侧管理的海上孤岛微电网运行策略.首先,针对微电网内不同类型负荷用户进行精细化建模以及差异化需求响应策略,提出不同类型用户的评判指标.其次,综合用户舒适度最优以及微电网日运行成本最小为目标建立日前调度框架,并利用yalmip+cplex求解器进行求解得到日前调度计划.最后,以某海上孤岛微电网为例,在不同风电场景下验证所提策略的有效性与普适性.

为了解居民电力用户是否具备调控潜力,分析不同时间段居民电力用户用电情况,提高居民电力用户行为画像准确率为目的,提出基于大数据分析的电力用户行为画像构建方法.通过采集用户用电行为相关的真实数据,建立包含用户基本信息、用电行为、用电特性与缴费行为标签的用户行为标签库;利用改进K均值聚类算法分析用户行为簇别,根据簇别的降序实施目标用户群的筛选,根据居民电力用户不同行为标签与总体调控簇别的划分结果构建用户画像,并以可视化方式呈现.实验结果证明:所研究方法可有效分析出不同时间段的居民电力用户的用电情况,所构建的用户画像准确率较高,可以精准找出具备调控潜力的居民电力用户.

In order to solve the problem of long execution time of complex dispatching instruction ticket, an automatic generation method of parallel strategy for dispatching instructions in distribution network based on clustering and layering is proposed. Firstly, the associated device cluster model is established. The operating devices in different associated device clusters can be executed in parallel, and the devices in the same associated device cluster can be executed in hierarchical order. Secondly, based on the method of cluster analysis, the device parallel operation cluster adjacency table DPCT is generated. Finally, the parallel strategy generation of scheduling instructions based on layering is proposed, and a detailed calculation formula is given. The example analysis shows that the method described in this paper can significantly improve the efficiency of the rapid execution of complex large-scale scheduling instruction tickets. It can better meet the needs of onsite dispatching operation.

通常,配电开关分合闸操作产生的振动信号中蕴含有体现机械状态的重要信息.提出一种基于振动信号分析的新型配电开关故障诊断方法.首先对振动信号求取乔-威廉斯分布获得二维时频矩阵,然后对时频矩阵作分块奇异值分解,用于表征不同机械状态的时频特性,最后结合极限学习机算法对4类实测振动信号的特征向量进行训练和测试.所提方法的优点是有效提取了配电开关振动信号时频域的特征,并且可以在较少样本的情况下训练诊断模型.基于实测数据的实验表明,该方法具有较高的识别精度和较快的收敛速度.

Frequent changes in the distribution model and data quality problems will cause greater interference to the real-time monitoring applications of the distribution management system(DMS),and the new construction mode of DMS puts forward higher performance requirements for multi-user concurrent maintenance.A management scheme of future-state distribution network model based on the polymorphism mechanism is proposed.The work,such as model maintenance based on the model change order,graph model access and verification of IEC 61970/61968 standard,model version management combining the whole model and incremental model,and automatic debug,is carried out under the independent environment.The appropriate decoupling is kept from the online operating environment,and the efficiency problem of multi-user concurrent maintenance is solved.Two actual application cases of DMS and the provincial centralized distribution monitoring system are presented.The practice results show that the proposed management scheme can meet the requirements of multi-source,multi-user and off-line automatic debugging of distribution model management.

In order to solve the problem of calculating speed and accuracy of the large-scale ultra-short term load forecasting in micro-grid engineering practice, the characteristics of various kinds of load types are analyzed. The optimal local shape similarity model is established. The similarity of the two local similarity curves is expressed by similarity coefficient. This paper introduces the human comfort index which is used to characterize the meteorological factors, and improves the human comfort index formula according to the influence of the air quality index on the human behavior. According to the optimal local shape similarity series and real-time data, an optimal local shape similarity ultra-short term load forecasting method based on weighted average is proposed to obtain the initial value of the ultra-short term load forecasting. The factors that affect the accuracy of ultra-short term load forecasting are analyzed, and the calculation methods of various influencing factors are developed. According to the improved human comfort index and various influencing factors, the initial value of the ultra-short term load forecasting is firstly corrected. Based on the deviation between the real-time data and the forecast data, the ultra-short term load forecasting value is secondly corrected using the super-stable adaptive control theory. The second correction value is the final ultra-short term load forecasting value. The feasibility of the proposed method is analyzed by an example. The method has good adaptability to the calculation speed and calculation accuracy in the large - scale ultra - short term load forecasting, and can meet the actual demand of the field engineering.

In the traditional anti-misoperation locking system,the electric locks are all unlocked through the key hole,thus causing the disadvantage of being technically unlocked.An intelligent non-contact unlocking system is proposed,in which the computer key supplies power to the intelligent lock through magnetic coupling power transmission technology,and interchanges instructions with intelligent lock through short distance wireless communication technology for controlling the internal mechanical structure of intelligent lock to realize non-contact unlocking.Meanwhile,the design method and testing process of the key technology of the wireless power supply are described.Testing results are basically identical with the results of theoretical analysis.The new design can meet the requirement on technical specifications of interlocking,thus eliminating the possibility of technical unlocking and improving the safety of electric power production.

The large-scale distributed generation such as wind and photovoltaic integrating into distribution network changes the network structure and flow direction and increases the uncertainty of power flow in distribution network.Meanwhile,the charging load of Electric Vehicles (EV) has randomness in time and space,and the superposition of daily load in distribution network with charging load of EV changes the original supply and demand,so traditional reliability assessment methods are no longer applicable.Based on statistical data of gasoline vehicle,this paper analyzes charging start instant and daily mileage and models the charging load of multi-EV,the sequential Monte Carlo method is applied to evaluate the reliability of distribution system containing distributed generation and EV.The main feeder F4 in IEEE-RBTS Bus6 test system is used to verify the validity of the reliability assessment method and make quantitative analysis of influence on system reliability from battery type of EV,location of charging point and numbers of EV respectively.

In order to solve the problem of single-phase-to-earth fault location in non-solid-earthed distribution, the abnormal characteristics of bus voltage under non-solid-earthed distribution fault are analyzed. A method of single-phase-to-earth fault identification is proposed, and concrete steps and formulas are given. The power flow trend of the distribution network under unbalanced capacitance injection is analyzed in the case of non-solid-earthed fault. When the unbalanced capacitance is injected, the steady-state current change upstream of the fault point in the fault phase is different from the steady-state current change downstream of the fault point, and the steady-state current change with the non-fault phase is also different. Based on the above analysis, the non-solid-earthed fault location problem is transformed into the steady-state current comparison problem. The equivalent circuit diagram of unbalanced capacitance current and the detailed calculation formula are given, and the approximate calculation formula and phasor diagram of the fault phase current are given. Based on the mutual cooperation of DMS (Distribution Management System) and unbalanced capacitance, the method of non-solid-earthed distribution fault location based on unbalanced capacitance injection is proposed, and detailed steps and specific rules are given. The simulation results show that the proposed method has good judgment adaptability for permanent non-solid-earthed fault and intermittent non-solid-earthed fault with longer duration.

基于对调度控制系统、配电自动化主站系统技术特点的分析,提出调配一体系统设计的基本原则,即一体化平台、一体化维护、一体化展示、核心应用独立运行;从调配一体建模、调配应用运行框架、调配应用协同三方面,探讨了基于智能电网调度控制系统基础平台(简称D5000平台)调配一体系统的关键技术;结合实际案例,介绍了D5000平台的调配一体系统的特点和研究成果.考虑到配网自动化建设因地制宜、分步实施的特点,作为现阶段提高配电网调度技术的支撑手段,推行调配一体系统有着积极长远的意义.

为了减少大型电力系统的潮流计算时间,研究简化电网的潮流计算方法具有重要意义.现有方法将传统Ward等值技术用于电网简化后,简化电网却产生与原始电网不同的潮流结果.针对这种情况,提出了一种改进的简化电网的直流潮流计算方法.在已有研究的基础上,给出了基于功率转移分布因子的直流潮流算法以及使用Ward技术对电网的简化方法.结合上述两部分,给出了基于功率转移分布因子的简化电网潮流计算方法.通过仿真算例表明,所得的结果具有较高的精度,可用于简化大型电网的潮流计算、潮流预报等应用中.

作为"中国一流调度机构"的福建电力调度通信中心如何能在管理上更上一个台阶,从根本上解决以往管理上的问题,做到闭环控制、偏差管理、可控在控,一直是中心探索的问题。如何建立一套适用于电力企业的实际应用需要,并能为企业的可持续发展提供良好支持的管理信息系统成为规划设计的重点。分析了管理信息系统的建设背景,从研究、规划设计、软硬件设计和平台功能开发等方面阐述了福建DMIS系统中平台化动态建模技术在调度业务中的应用。

介绍了电网故障分析定位系统GIS平台建设的要求;探讨了电网故障分析定位系统与雷电信息定位系统、森林防火系统和保护故障信息系统的集成;研究了电网故障的分析和故障定位算法;讨论了系统的应用和存在的问题。