【作者】 郭慧；

【导师】 罗建； 汪飞；

【作者基本信息】 上海大学 ， 电力电子与电力传动， 2019， 博士

【摘要】 区域能源互联网是一个能源与信息深度融合的分布式复杂系统。若不同能源系统单独规划、独立运行,彼此间缺乏协调性,会导致能源利用率低、自愈能力弱、系统安全可靠性低等问题。因此,科学合理的规划方法和运行策略,对于分布式多能源系统的全局效率和经济性至关重要。由于分布式多能源系统的系统规划与运行优化是一个大规模的混合整数非线性优化问题,采用集中方式难以求解。可再生能源发电功率和负荷需求功率的不确定性以及随机性,增加了问题的复杂度。因此,本文根据能源传输损耗和源荷互补特性,基于分解协调思想实现不同形式能源与负荷的合理分区以及网络重构的协同规划。在此基础上,根据优化目标的不同时空解耦特性,开展分布式分层优化控制策略的研究,降低多能源系统运行优化的复杂度和耦合性,并针对可再生能源发电和负荷需求功率预测误差及其随机因素产生的调度偏差,引入滚动时域优化思想进行修正。本文主要研究内容及创新工作如下:1)针对分布式能源和负荷具有数量多、容量小、分布广、时变性等特点,为实现不同能源基础设施的合理配置,减少能源的远距离传输损耗,创新性提出基于能源传输损耗和源荷互补特性的聚类分区优化模型。同时,优化改进传统k-means求解算法,通过聚类中心数的预先确定以及初始聚类中心的优化选取,实现大规模分布式能源和负荷的合理分区,以及能量枢纽的优化选址。2)在大规模分布式能源和负荷实现合理分区的基础上,针对区域内分布式能源和负荷的互联以及区域间的网络互联问题,提出基于改进最小生成树的分布式能源网络重构方法,实现区域能源网内的星-枝状拓扑规划以及区域能源网间的网状拓扑规划,以建设一个经济可靠的能源互联网络。3)进一步考虑区域能源互联网的运行优化问题,为适应分布式能源的即插即用性,解决多源、多荷、多态运行模式下,区域能源互联网的多能互补和需求响应问题,提出区域能源互联网对等自主的分布式分层运行优化策略。根据优化目标的时间解耦特性,在较长时间尺度的日前调度引入代表能量分配的可变调度因子对能量枢纽模型进行优化,实现多能源经济优化调度。考虑可再生能源发电和负荷功率预测误差,提出基于模型预测控制的日内优化调度,通过日前/日内调度一致性权重,保证储能长时间尺度参与系统能量平衡的能力。针对较短时间尺度,反应电能质量的静态电压偏差和功率实时优化分配问题,创新性提出计及储能SOC的电压/电流双补偿型分布式自适应下垂控制,实现电压偏差调节和电流分配精度调节。根据优化目标的空间解耦特性和不同区域能源网的信息隐私问题,以及能源传输的实时性和异步性,创新性提出一种基于撮合交易竞价机制的分布式能量路由控制方法。通过采用最短路径算法的智能型电能路由器为特定的能源交易方选取最小网损路径,实现端到端的电能传输以及阻塞管理。4)为验证所提出的运行优化控制策略的效果,采用半实物仿真技术对分布式多能源系统的运行优化进行研究,提出并应用缩放方法来解决实物装置的容量或功率等级与仿真对象不匹配的问题。通过搭建半实物仿真平台,验证并展示了部分运行优化控制策略的效果。更多还原

【Abstract】 Regional energy internet is a distributed and complex system with deep integration of energy and information.If different energy systems are planned and operated independently,it will result in the low energy utilization,weak self-healing ability,and low system reliability.Therefore,scientific reasonable planning methods and operation strategies are essential for the global efficiency and economy of distributed multi-energy systems.Because the system planning and operation optimization of a distributed multi-energy system is a large-scale mixed integer nonlinear optimization problem,it is difficult to be solved by a centralized manner.Besides,the uncertainty and randomness of renewable energy generation and load demands increase the complexity of the problem.According to the transmission loss and energy balance,the coordinated planning of multi-energy,including clustering partition and network reconfiguration,is realized based on the decomposition-coordination theory.Following that,according to the different spatiotemporal decoupling characteristics of optimization objectives,a distributed hierarchical optimization strategy is implemented to reduce the complexity and coupling of multi-energy system’s operation optimization.In order to correct the optimal dispatch deviation caused by the forecasting error of renewable energy generation and load demands,the rolling horizon optimization method is introduced.The main research and innovation work of this paper are as follows:1)Distributed energy and loads have the characteristics of large quantity,small capacity,wide distribution,and time variability.In order to realize the reasonable configuration of different energy infrastructures and reduce the long-distance transmission loss of energy,a clustering partition optimization model based on transmission loss and energy balance is proposed innovatively.Simultaneously,the traditional k-means algorithm is optimized and improved to realize the reasonable partition of large-scale distributed energy and loads,as well as the location optimization of energy hub,by the pre-determination of cluster center’s number and the optimal selection of initial cluster centers.2)On the basis of achieving reasonable partition of large-scale distributed energy and loads,as for the interconnection of distributed energy and loads within and between regional energy networks,a distributed energy network reconfiguration method based on improved minimum spanning tree algorithm is proposed to build an economical and reliable energy internet,by realizing the star-branch topology planning within regional energy networks and the mesh topology planning between regional energy networks.3)Based on the above,the operation optimization of regional energy internet is need further consideration.In order to adapt to the plug-and-play of distributed energy,and solve the multi-energy complementarity and demand response of regional energy internet under multi-source,multi-load and multi-state operation modes.A peer-to-peer autonomous distributed hierarchical control strategy is proposed,based on different spatial and temporal scales of the optimization objectives.In the day-ahead scheduling of a long time scale,a variable dispatch factor representing energy allocation is introduced to optimize the energy hub model,and to realize the economic dispatch of multi-energy.Considering the prediction error of renewable energy generation and load demands,a daily scheduling based on model predictive control is proposed,and the ability of storage to participate in the system energy balance over a long period is ensured,by introduce the weight consensus of day-ahead scheduling and daily scheduling.In the short time scale,as for the voltage deviation reflecting power quality and the power real-time optimal allocation,an SOC-based voltage/current double-compensated distributed adaptive droop control is proposed innovatively,to achieve the adjustment of voltage deviation and current allocation.Considering the spatial decoupling characteristics of optimization objectives and the information privacy of different regional energy networks,as well as the real time and asynchrony of energy transmission,a distributed energy routing control method based on market transaction is proposed innovatively.In order to achieve the end-to-end power transmission and congestion managements,a minimum loss path with no congestion is selected for the specific energy trader by the intelligent electricity router based on the shortest path algorithm.4)To verify the effectiveness of the proposed operation optimization strategy,the hardware in the loop simulation technology is used to study the operation optimization of distributed multi-energy systems,and a scaling method is proposed and applied to solve the problem that the capacity or power level of physical device does not match the simulation object.Through the established hardware-in-the-loop simulation platform,the effectiveness of the proposed optimal control strategy is verified and displayed partly.更多还原