【作者】 张迪；

【导师】 苗世洪；

【作者基本信息】 华中科技大学 ， 电气工程， 2021， 博士

【摘要】 分布式光伏发电具有清洁环保、布点广泛、成本低廉和可就近利用等特点,已经成为配电网内的主要发电形式之一,也是推动我国能源结构改革和实现“30·60”战略目标的重要支撑技术之一。为了实现以光伏发电为代表的分布式发电健康、稳定和可持续发展,国家发改委和国家能源局于2017年底联合发布了《关于开展分布式发电市场化交易试点的通知》政策文件,以推动我国分布式发电市场的建设。近年来,分布式发电市场化交易政策为分布式光伏发电提供了多样化的市场交易模式,促进了分布式光伏发电的快速发展。然而,分布式发电市场具有的各参与主体利益诉求不一、交互关系复杂、交易模式多样等特点,不仅深刻地改变了分布式光伏发电商的投建行为和收益模式,对分布式光伏发电优化规划技术提出了新的要求,还影响了电网的电源结构及其与电力用户的供电关系,为电网运行的经济性和安全性带来了新的挑战。因此,深入分析分布式发电市场化运行机制,探讨合理的分布式光伏发电规划方法,提出面向分布式发电市场的电网协调优化运行策略,对于分布式光伏发电友好接入电网、促进分布式光伏的高效消纳、推动分布式发电市场高标准建设和保障电力系统经济安全稳定运行均具有重要的理论研究价值和工程实践意义。鉴于此,本文针对分布式发电市场化环境下光伏发电规划与电网协调优化运行技术进行了深入研究,主要研究内容如下:(1)针对分布式光伏发电入网承载力量化评估问题,提出了考虑谐波耦合约束的分布式光伏发电入网承载力量化评估方法,该方法为后续分布式光伏发电规划和电网协调优化运行的研究奠定约束基础。首先,根据分布式光伏发电并网点谐波电压电流耦合关系,推导频率耦合矩阵,建立分布式光伏发电谐波耦合约束模型;其次,在此基础上,以分布式光伏安装容量最大为目标,建立了分布式光伏入网承载力量化评估模型;然后,针对谐波耦合约束的非凸性问题,利用凸松弛转换方法进行松弛处理;最后,将原始非凸模型转换为易于求解的半正定规划模型,实现了模型的高效快速求解。仿真结果表明,将谐波耦合约束引入到分布式光伏发电入网承载力评估模型后,不仅可以得到更加合理的分布式光伏入网承载力评估结果,而且能够有效降低电网运行风险,为电网企业和政府部门提供决策依据。(2)针对多类型交易模式下分布式光伏优化规划问题,提出了考虑扶贫光伏和普通分布式光伏协同规划需求的分布式光伏优化规划技术。首先,结合分布式发电市场化交易政策推出的多类型交易模式,建立分布式发电市场化约束体系;其次,计及扶贫光伏和普通分布式光伏发电协同规划的实际工程需求,构建了多类型交易模式下分布式光伏双层优化规划模型:下层模型以分布式光伏发电商收益最大为目标,决策分布式光伏投建容量,上层模型考虑电网多类型运行需求,决策分布式光伏接入模式;然后,针对模型中出现的决策变量相乘的非凸型约束,利用凸松弛方法进行凸包替代处理;最后,采用卡罗需-库恩-塔克条件转换方法将多下层相互独立的双层规划模型转换为单层规划模型,实现了模型的高效快速求解。仿真结果表明:采用本文所提分布式光伏优化规划方法,在提高分布式发电商收益、减少电网企业经济损失的同时,还可以有效提升电网运行水平,更大程度地接纳分布式光伏投建容量。(3)在分布式光伏发电规划的研究成果基础上,针对分布式发电市场化环境下电网优化运行问题,提出了面向分布式发电市场的主动配电网优化运行策略。首先,根据分布式发电市场多参与主体的特点,分别建立了分布式发电市场化环境下分布式光伏发电商、电力用户、政府机构和电网企业的响应行为模型;其次,考虑分布式发电市场各主体响应行为的动作顺序特点,以分布式发电商响应行为模型为上层模型,建立了多下层相互影响的双层规划模型,并利用最优性条件证明了下层模型转换方法的有效性;然后,采用多目标进化算法,进一步获取了模型的帕累托最优解集;最后,对各主体市场化响应行为进行了分析,研究了电网不同的运行策略对分布式发电市场和其他各主体造成的影响。仿真结果表明:分布式发电市场化环境中各主体间存在着竞争与合作并存的利益交互关系,与以电网经济运行成本最小的优化目标相比,电网企业以分布式发电消纳量最大为优化目标可以有效降低分布式发电商运行成本和负荷用电成本,促进分布式发电市场化交易电量的进一步提升。(4)针对高比例分布式光伏接入背景下输配电网协调优化运行问题,提出了一种输配电网协调优化运行新算法。首先,建立了输配电网协调优化运行模型,在此基础上,利用二阶锥松弛方法实现了原始模型的凸化处理;其次,考虑松弛后的模型具有大量二次不等式约束推导复杂的特点,采用多面体逼近近似表达的方法,实现了二阶锥约束的线性化表达;然后,在最优性条件变换推导的基础上,利用线性化表达方法,分别将配电网运行目标函数和根节点无功功率表达成关于根节点有功功率的分段线性函数;最后,原始输配电网协调优化运行模型被转化为一个易于求解的混合整数规划模型,实现了模型的高效快速求解。仿真结果表明,相较于传统的输配电网协调优化算法,本文所提算法在求解效率和求解准确度方面均具有更加良好的表现。更多还原

【Abstract】 Distributed photovoltaic generation has the characteristics of clean and environmentally friendly,widely distributed,low cost and nearby utilization,which has been one of the main forms of power generation in the distribution network.Moreover,distributed photovoltaic generation would be an important support for my country to promote energy structure reform and one of the technologies to achieve the "30·60" strategic goal.For the healthy,stable and sustainable development of distributed power generation represented by photovoltaic power generation,the National Development and Reform Commission and the National Energy Administration jointly issued the policy document "the Notice on Carrying out Pilot Programs on Distributed Generation Market Trading" at the end of 2017.Since then,the construction of distributed power generation market has been promoted.For more than three years,the marketoriented transaction policy of distributed power generation has provided diversified market transaction models for distributed photovoltaic power generation,and promoted the rapid development of distributed photovoltaic power generation.However,the distributed power generation market has the characteristics of different interest demands,complex interactions,and diverse transaction modes.On the one hand,the market has profoundly changed the investment behavior and profit model for distributed photovoltaic generation companies,and put forward new requirements for planning technology of distributed photovoltaic generation;On the other hand,the market has affected the power supply structure of grid,changed the relationship between grid and power users,brought new challenges to the economy and safety of grid.Therefore,it is necessary to deeply analyze operation mechanism of distributed generation market,explore reasonable planning methods of distributed photovoltaic generation,and propose coordinated and optimized operation strategies to grid for distributed generation market.These are important theoretical research value and engineering practical significanc to improve distributed photovoltaic generation to access the grid friendly,promote the efficient consumption of distributed photovoltaic generation,realize the high-standard construction of the distributed generation market,and ensure the safe and stable operation of power system.Therefore,this article conducts indepth research on photovoltaic generation planning and power grid coordinated and optimized operation technology in the marketization environment of distributed generation.The main contents are as follows:(1)To solve the problem of evaluating the power grid hosting capacity of distributed photovoltaic generation accessed to the grid,a method for evaluating the power grid hosting capacity of distributed photovoltaic power generation considering harmonic coupling constraints was proposed.This method would lay a constraining foundation for subsequent distributed photovoltaic generation planning and coordinated and optimized operation of grid.Firstly,according to the harmonic voltage and current coupling relationship of distributed photovoltaic generation grid connection point,the frequency coupling matrix was derived,and the harmonic coupling constraint model of distributed photovoltaic generation was established.Secondly,on this basis,a model for evaluating the power grid hosting capacity of distributed photovoltaic generation accessed to the grid was established,which aimed to maximize the installed capacity of distributed photovoltaics.Thirdly,considering the problem of harmonic coupling constraint non-convexity,the convex relaxation conversion method was used for relaxation processing.Finally,the original non-convex model was converted into a positive semi-definite programming model that can be solved by mature solving software,which realized the efficient and fast solution.The results are as follows: in the evaluation model of power grid hosting capacity of distributed PV accessed to network,it is not only necessary to consider harmonic coupling constraints,but also can effectively reduce grid operation risks.Besides,it can provide reasonable evaluation results for grid companies and government departments.(2)For the problem of distributed photovoltaic optimization planning with multi-type transaction mode,a distributed photovoltaic optimization planning method considering the needs of distributed photovoltaic generation collaborative planning was proposed.Firstly,based on the multi-type transaction mechanism introduced by the distributed generation marketoriented transaction policy,the distributed generation market-oriented constraint was established.Secondly,taking into account the actual engineering needs of the coordinated planning of poverty alleviation photovoltaic and ordinary distributed photovoltaic power generation,a distributed photovoltaic bi-level optimization planning model with multi-type transaction mechanism was constructed.The lower level model aimed at maximizing the revenue of distributed photovoltaic generation companies,and distribution PV construction capacity was decided.The upper level model considers the multiple types of operation requirements of power grid,and the distributed PV access mode was decided.Thirdly,for the non-convex constraints with the multiplication terms of the decision variables,the convex relaxation method was taken to perform convex hull substitution.Finally,Karush-Kuhn-Tucker Conditions was used to convert bi-level model with the multi-low-level independent into a single-level model,which could realize the efficient and rapid solution.The results are as follows: The proposed method improves the income of distributed generators and reduces the economic losses of grid.At the same time it also effectively improves the grid operation level and more bearing capacity of PVs.(3)On the basis of the research of distributed photovoltaic generation planning,in view of the optimal operation of grid in the marketization environment of distributed generation,an optimization operation technology of active distribution network for the distributed generation market was proposed.Firstly,according to the characteristics of multiple participants in the distributed generation market,the response behavior models of distributed generation companies,power customers,government agencies and grid companies in the marketization environment of distributed generation were established respectively.Secondly,considering the characteristics of the action sequence of the response behavior of the distributed generation market,a bi-level model with multi-low-level interacting was established,whose upper model was the response behavior model of the distributed power generation company.Moreover,the optimality condition was taken to prove the effectiveness of the model conversion method.Thirdly,the multi-objective evolutionary algorithm was used to further obtain the Pareto optimal solution set of the model.Finally,the response behavior of each entity was analyzed,and the impact of different grid operation strategies on the distributed generation market and other entities was studied.The results are as follows: there is an interactive relationship of competition and cooperation interests between the various entities in the marketization environment of distributed generation.Compared with the optimization goal of the minimum economic operation cost,the grid enterprise aims to maximize the consumption of distributed generation,which can be effective reduce the operating costs of distributed power generation companies and power customers,and promote further increase of distributed generation market trading electricity.(4)With the continuous increase of the scale of distributed photovoltaic generation in the distribution network,the problem of coordinated and optimized operation of the transmission and distribution network for high proportion of distributed photovoltaic access has become prominenter.Therefore,based on the AC optimal power flow,a non-iterative coordination optimization algorithm for transmission and distribution network was proposed.Firstly,a twolevel grid coordination model for transmission and distribution network was established.On this basis,the convexity of the original model was solved by using the second-order cone relaxation method.Secondly,considering the complexity of quadratic inequality constraint derivation of the relaxed model,a polyhedron approximation method was adopted to realize the linear expression of the second-order cone constraints.Thirdly,based on the derivation of optimality condition transformation and linear expression method,the distribution network operation objective function and root node reactive power were respectively expressed as piecewise linear functions about the root node active power.Finally,the original model was transformed into a mixed integer programming model that can be solved by a mature algorithm,which can realize the efficient and rapid solution.The results are as follows: compared with the traditional optimization algorithm,the algorithm proposed in this paper has a better performance in terms of solution efficiency and accuracy.更多还原