基于灰色PSO算法的分布式电网多目标调峰优化调度研究

发布时间：2018-09-14 08:53

【摘要】：目前我国新能源发展已经走在了世界前列,成为全球风电规模最大、光伏发电增长最快的国家。然而风电、光伏这类分布式电源(Distributed Generation,DG)出力具有随机性和间歇性,这种不确定性使系统的调峰调度任务更加艰巨。而传统的依赖发电机组进行调峰的方式已经无法完成目前新能源大规模并网所带来的调峰任务。因此本文针对目前新能源大规模并网,传统的调峰调度运行方式不能满足新能源出力消纳要求的问题,探索一种用户侧主动参与调峰调度的可行方案。论文首先简要介绍了我国现行的调峰调度模式,并且对其调峰调度模式的优缺点进行分析。根据新能源发电的出力特性说明新能源大规模并网对电力系统所带来的影响。论文针对目前调峰调度方式在新能源消纳难以奏效的情况,挖掘需求侧调峰资源、研究需求侧管理方法,将新能源发电的特性与不同种类负荷的特性进行联动,寻求一种可以实现功率互补,减小系统调峰压力的协调调度模式。其次,论文针对DG、储能单元、可控负荷、可中断负荷等这些不同类型的分布式能源(Distributed Energy Resource,DER)数量多、分布分散、类型不一、难以管理的问题,提出利用虚拟电厂(Virtual Power Plant,VPP)将多个DER进行协调优化,实现调峰资源合理优化配置,构建以负荷方差最小以及运行成本最小为目标的多目标调峰调度模型。论文针对传统多目标粒子群(Multi-objective Particle Swarm Optimization,MOPSO)算法精度低、容易陷入局极小、更新策略具有随机性和对种群的全局最优解选取缺乏指导所造成优化结果的客观性和可信度不足的缺点,提出了基于灰色关联度的多目标粒子群改进算法。最后,论文以IEEE33节点配电系统为算例进行仿真,通过验证表明论文提出的方法能够有效增强系统调峰能力、提高新能源消纳水平以及降低电网运行成本。
[Abstract]:At present, China's new energy development has been in the forefront of the world, become the world's largest wind power, photovoltaic power generation growth of the fastest country. However, the distributed power generation (Distributed Generation,DG), such as wind power and photovoltaic, has randomness and intermittency, which makes the task of peak-shaving scheduling more difficult. However, the traditional peak-shaving method based on generator sets has been unable to complete the peak-shaving task brought by large-scale grid connection of new energy sources. Therefore, in view of the problem that the traditional peak-shaving operation mode can not meet the demand of new energy, this paper explores a feasible scheme of active participation of user side in peak-shaving scheduling. Firstly, the paper briefly introduces the current peak-shaving scheduling mode in China, and analyzes the advantages and disadvantages of the peak-shaving scheduling mode. According to the power generation characteristics of new energy generation, the influence of large-scale grid connection of new energy on power system is explained. Aiming at the situation that the current peak-shaving dispatching mode is difficult to work in the new energy consumption, the paper excavates the demand-side peak-shaving resources, studies the demand-side management method, and links the characteristics of new energy generation with the characteristics of different kinds of loads. To seek a coordinated scheduling mode which can realize the complementary power and reduce the peak-shaving pressure of the system. Secondly, this paper aims at the problems of DG, energy storage unit, controllable load, interruptible load and so on, which are many in quantity, scattered in distribution, different in type and difficult to manage. A multi-objective peak-shaving scheduling model with minimum load variance and minimum operating cost is proposed by using virtual power plant (Virtual Power Plant,VPP) to coordinate and optimize multiple DER to realize rational allocation of peak-shaving resources. In this paper, the traditional multi-objective particle swarm optimization (Multi-objective Particle Swarm Optimization,MOPSO) algorithm has the disadvantages of low precision, easy to fall into local minima, randomness of update strategy and lack of guidance to select the global optimal solution of the population, which leads to the lack of objectivity and credibility of the optimization results. An improved multi-objective particle swarm optimization algorithm based on grey correlation degree is proposed. Finally, the simulation of IEEE33 node distribution system shows that the proposed method can effectively enhance the peak-shaving ability of the system, improve the level of new energy consumption and reduce the power grid operation costs.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【参考文献】