【作者】 张节潭；

【导师】 程浩忠；

【作者基本信息】 上海交通大学 ， 电力系统及其自动化， 2009， 博士

【摘要】 风电具有一些与火电、水电、核电等常规发电机组不同的特性,如间歇性、波动性、反调峰特性等。本文从电源规划的角度分析了风电场给电力系统带来的影响,并以风电场的接入方式为切入点,研究了含大型风电场的电源规划问题和分布式风电源的选址定容规划问题。其主要研究成果有:1)提出了基于等效电量频率法的含风电场的电力系统随机生产模拟方法。该方法利用等效负荷频率曲线计算发电机组的启停次数,在生产模拟中保留了负荷和风电机组的时变特性,不仅可以得到常规算法所能得到的生产模拟结果,而且可以评估风电场接入对常规机组造成的开停机影响,以及与火电机组开机、停机等因素相关的动态费用。2)以等风险度原则为基础,以各检修时段风险度方差最小为目标函数,建立了含风电场的电力系统发电机组检修计划模型,提出了最小累积风险度法和模拟植物生长算法两种求解方法。采用半不变量法结合Gram-Charlier级数展开法求解系统的风险度,考虑了风电场出力的不确定性。最小累积风险度法通过寻找检修区间内累积风险度最小的区间确定待检修机组的检修位置。模拟植物生长算法以发电机组检修的开始时段序号作为决策变量,利用植物的向光性生长机制寻找最优解。3)提出了含风电场的成本最小化电源规划模型,考虑了系统的调峰能力约束、调频能力约束以及污染物排放量约束。在此基础上,计及上网电价差异对电源规划的影响,按照分解协调的思想建立了净收益最大化双层电源规划模型,上层规划为电源投资决策问题,以发电商总收益最大为目标函数,下层规划为生产优化决策问题,并提出了模拟植物生长算法、最小累积风险度法、等效电量频率法相结合的求解方法。4)提出了分布式风电源接入现有配电网的选址定容机会约束规划模型,以独立发电商收益最大为目标函数,允许规划方案以一定的置信水平满足节点电压约束和支路功率传输约束;提出了模拟植物生长算法与随机潮流相结合的求解方法,构建了风电机组出力离散化模型,采用基于半不变量的配电网随机潮流方法评估规划方案是否满足概率性约束条件,考虑了分布式风电源可能导致的双向潮流问题。5)提出了主动管理模式下基于双层规划的分布式风电源选址定容规划模型,打破了被动管理模式下分布式电源接入配电网所遵循的“安装即忘记”原则,以分布式风电源的净收益期望值最大作为上层规划目标,以满足电压和潮流约束下风电源出力切除量期望值最小作为下层规划目标,考虑了主动电压管理在改善支路潮流和节点电压方面的作用,提出了模拟植物生长算法与概率最优潮流算法相结合的求解方法。相关算例验证了所提出的含风电场的电源规划模型、方法的正确性和有效性。更多还原

【Abstract】 Unlike conventional generation sources, such as coal or gas fired power, hydro power, and nuclear power, etc., wind power is with natures of low-controllable, stochastic, intermittent, and anti-peak-shaving. This dissertation analyses the impact of wind power on generation expansion planning, and studies generation expansion planning problems with large-scale wind farms and sizing and siting of distributed wind generation (DWG). The main innovations of the dissertation are as follows:1) According to the natures of stochastic and intermittent of wind turbines, equivalent energy and frequency function (EEFF) method is proposed for power system probabilistic production simulation, which combines the frequency and duration (FD) approach and equiva1ent energy function (EEF) method. EEFF method keeps the time-dependent behaviour of power load and wind turbine generator, and it can evaluate the effect of wind power and load on the commitment of conventional units, as well as the dynamic cost associated with the commitment of thermal units.2) A novel generator maintenance scheduling model based on equivalent risk principle is present, which takes the minimum of risk standard variance of all the maintainance period as planning objective. Heuristic algorithm based on minimum cumulative risk and plant growth simulation algorithm (PGSA) are put forword to solve the model. Both algorithms calculate the system risk with the method of combined semi-invariant and Gram-Charlier expanding, and take into account the uncertainty of wind power. Minimum cumulative risk algorithm takes the minimum cumulative risk period as the maintenance period of the unit. PGSA takes the start period number of the generators to be maintained as decision variables and search the optimal solution with the plant phototropism mechanism.3) An generation expansion planning model of cost minimization with large-scale wind farms is put forward, considering the impacts of wind farms on system peak regulation, frequency regulation, and environmental protection benefit. In order to take into account the impact of different generation price on investment decision, a bi-level generation expansion planning model is put forword, whose top planning problem is the generator investment decision problem with objective of net benefit maximinum, and lower planning problem is production optimizing decision problem. The PGSA combined with minimum cumulative risk algorithm and EEFF method is present to solve the bi-level programming model.4) A planning scheme based on the chance constrained programming of distributed wind generation in the existing distribution networks is proposed, whose planning objective is to maximize the benefit of the independent power producer (IPP). PGSA is put forword to solve the model and the probabilistic power flow method based on semi-invariant is applied to judge whether the planning schemes satisfy the constraints of both the node voltage operation ranges and the branch transmission capacities.5) A novel bi-level programming model for siting and sizing of distributed wind generation under active management (AM) mode is put forward, which breaks the“fit and forget”installation policy of distributed generation in the passive distribution network. The model takes the maximum expectation of net benefit of DWG as the upper level program objective, and takes the minimum expectation of generation curtailment with voltage and thermal constraints as the lower level program objective, and takes into account the impact of active voltage management algorithm on improvement of branch power flow and node voltage. The PGSA combined with probabilistic optimal power flow algorithm is applied to solve the optimal planning of DWG under AM mode.Numerical examples prove the feasibility and effectivity of the proposed model and methodology in this dissertation.更多还原