在线经济调度的鲁棒优化方法研究

发布时间：2018-02-25 16:21

本文关键词： 电力系统在线经济调度鲁棒优化不确定性多目标优化　出处：《山东大学》2015年硕士论文　论文类型：学位论文

【摘要】：风电、光伏发电等新型能源发电方式,具有清洁、可再生优点,缓解了我国能源压力,但新能源发电形式多易受气候、环境等因素的影响,具有明显的随机性和间歇性,此类电源大规模接入电网必然会增加电网运行中的不确定性,为电网的安全运行带来隐患。在线经济调度,做为与控制对接的最后一步,在电网扰动消纳上起到重要的作用,因此,对于在线经济调度理论的研究具有十分重要的意义。随机规划、模糊规划以及鲁棒优化等多种不确定性决策方法在电网在线经济调度问题中得到应用。其中,随机规划依据不确定注入量的概率分布信息,通过对各随机场景的统筹考虑,给出具有概率优性的决策结果。然而,概率信息准确获取的困难以及计算的复杂性限制了随机规划决策方法的应用。模糊规划方法采用隶属度函数表示决策者对不确定注入量及其导致后果的态度,通过最大化隶属值,获得满意的决策结果。然而,由于决策结果受隶属度函数影响显著,模糊决策结果的主观性较强。鲁棒优化不同于以上两种规划方法,当不确定量区间确定的情况下,其决策不需要其概率分布特征,也不夹杂决策者的主观意愿,仅根据扰动边界,通过寻找并满足决策中的最“劣”情况,保证决策结果的鲁棒性。目前,鲁棒在线经济调度理论研究存在以下几点问题：①随着新能源的快速发展,可控机组所能接纳的注入波动将无法匹配新能源接入量,百分百吸纳的在线经济调度模型已不符合实际；②在线经济调度模型多偏重于保证安全性,导致模型保守度太高,尽管对保守度改进方面有很多研究,但有进一步提升的空间；③大部分线路约束模型直接运用直流潮流模型,但真正的功率注入转移分布跟直流潮流结果有较大的不确定性,而这部分不确定性往往被忽略。本文首先提出一种电力系统在线经济调度的最大有效静态扰动安全接纳范围法,以机组运行基点与参与因子为决策变量,构建双目标模型,实现有效静态扰动安全接纳范围最大化条件下的经济性最优：其次,提出保守度可控的鲁棒在线经济调度模型,通过在原有的鲁棒在线经济调度模型两层目标处理中,加入控制系数β,改进双目标规划求解方法,实现安全性和经济性的折中,从而达到求解两层目标帕累托最优解的目的；最后,提出计及电网注入转移分布因子不确定性的鲁棒在线经济调度方法。以系统运行成本为目标,结合现在的电网系统量测技术,利用量测数据得到的注入转移分布因子区间表达及其相应的线路传输功率约束,构建综合考虑系统节点注入不确定性和注入转移分布因子不确定性的鲁棒在线经济调度模型,使得优化调度结果得到进一步符合电力系统的运行实际。
[Abstract]:New energy generation methods, such as wind power, photovoltaic power generation and so on, have the advantages of clean and renewable, which alleviate the energy pressure in China. However, the forms of new energy generation are easily affected by climate, environment and other factors, and have obvious randomness and intermittence. The large-scale connection of this kind of power supply to the power network will inevitably increase the uncertainty in the operation of the power network and bring hidden trouble to the safe operation of the power network. As the last step of connecting with the control, the online economic dispatch plays an important role in the elimination of the disturbance of the power network. Therefore, it is of great significance to study the theory of online economic dispatch. Stochastic programming, fuzzy programming and robust optimization are applied to the online economic dispatching problem of power network. According to the probability distribution information of the uncertain injection amount, the stochastic programming gives the decision results with probabilistic optimality through the overall consideration of each random scene. However, The difficulty of accurate acquisition of probabilistic information and the complexity of calculation limit the application of stochastic programming decision method. Fuzzy programming uses membership function to express the decision maker's attitude towards the uncertain injection quantity and its consequences. By maximizing the membership value, a satisfactory decision result is obtained. However, because the decision result is significantly influenced by membership function, the fuzzy decision result is more subjective. The robust optimization is different from the above two programming methods. When the interval of uncertainty is determined, the decision making does not need its probability distribution characteristic, nor does it include the subjective will of the decision maker, only according to the disturbance boundary, it can find and satisfy the worst case in the decision making. At present, the following problems exist in the research of robust online economic scheduling theory: 1 with the rapid development of new energy sources, the injection fluctuations that can be accepted by controllable units will not match the amount of new energy access. The 100% of the online economic scheduling model has been out of line with the actual Y2 online economic scheduling model, which is mainly focused on ensuring security, resulting in too high a conservative degree of the model, although there are many studies on the improvement of conservatism. However, most of the line constraint models with further improvement use DC power flow model directly, but the true power injection transfer distribution is uncertain with DC power flow results. However, this part of uncertainty is often ignored. In this paper, a maximum effective static disturbance safe admission range method for online economic dispatch of power system is proposed, in which the basic point of unit operation and the participation factor are taken as decision variables to construct a two-objective model. Under the condition of maximizing the safe admission range of the effective static disturbance, the economic optimization is achieved. Secondly, a robust online economic scheduling model with controllable conservatism is proposed, which is based on the original robust online economic scheduling model in the two-layer target processing. By adding the control coefficient 尾, the two-objective programming method is improved to achieve the compromise between security and economy, so as to achieve the purpose of solving the two-level objective Pareto optimal solution. This paper presents a robust online economic dispatching method, which takes into account the uncertainty of injection transfer distribution factor, aiming at system operation cost, and combining with the current measurement technology of power system. Based on the interval expression of the injection-transfer distribution factor and the corresponding transmission power constraints, a robust on-line economic scheduling model considering the uncertainties of the system node injection and the injection-transfer distribution factor is constructed. The optimal dispatching results are further in line with the actual operation of the power system.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM73

【共引文献】