基于功率预测的风电并网优化调度研究

发布时间：2018-01-14 04:28

本文关键词：基于功率预测的风电并网优化调度研究　出处：《北京交通大学》2014年硕士论文　论文类型：学位论文

【摘要】：风能是一种可再生清洁能源。大力发展风力发电对改善能源结构、应对气候变化和能源安全问题具有十分重要的意义。然而风能具有波动性、间歇性和不可控性等特点,大规模风电接入电网对电力系统可靠运行和经济调度带来了巨大挑战。因此,研究风电功率预测和含风电场的电力系统优化调度为提高风电利用水平和智能电网建设具有重要的经济意义和实际价值。本文基于含风电的电力系统经济调度问题,展开了如下研究： 1.将具有较大聚类系数和较小平均路径的NW型小世界网络作为粒子群优化算法的拓扑结构,提出了小世界邻域粒子群优化(SW-PSO)算法。仿真分析表明,该算法在求解高维寻优问题上具有搜索速度快、寻优精度高的优点,适宜于优化神经网络预测模型和求解大规模非线性数学规划问题。 2.研究了基于SW-PSO算法的BP神经网络预测模型。采用风电场实际数据建立风机风速-功率曲线,利用NWP信息实现了日前风电功率预测。由于NWP数据自身存在误差影响最终风电功率的预测精度,将风电功率历史数据与NWP数据相结合实现了风电功率超短期滚动预测。 3.在风电功率预测的基础上,将风电预测误差和穿透功率计入系统旋转备用当中,建立了含风电场电力系统动态经济调度数学模型。该模型是一个含有多约束条件的大规模非线性数学规划问题,利用SW-PSO算法对其进行求解,并采用调整策略修正粒子保证了其在可行域中飞行寻优。通过算例验证了SW-PSO算法在求解风电并网优化调度问题方面的有效性和可行性。 4.将机会约束规划引入到含风电场的电力系统动态经济调度模型当中,以概率形式描述相关约束条件,建立了风电并网电力系统随机优化调度模型。算例表明应用蒙特卡洛随机模拟技术和SW-PSO算法对求解电力系统随机优化调度问题是行之有效的,能够为决策者提供既满足电网安全稳定运行又符合新能源消纳需求的合理规划方案。
[Abstract]:Wind energy is a renewable and clean energy. The development of wind power to improve the energy structure, has very important significance to deal with the problem of climate change and energy security. However, the wind is volatile, intermittent and uncontrollable, large scale wind power integration has brought great challenges for reliable operation and economic dispatch of power system. Therefore, study on wind power forecasting and wind power system scheduling to improve wind power has important economic significance and practical value of utilization and development of smart power grids. The power system with wind power based on the economic dispatch problem, are done as follows:
1. will have a relatively large clustering coefficient and small average path of the NW type small world network as the topology of particle swarm optimization algorithm, proposed the small world neighborhood particle swarm optimization (SW-PSO) algorithm. Simulation results show that the algorithm has fast search speed in solving high dimension optimization problems, find the advantages of high precision, suitable for the model and solving large-scale nonlinear programming problem to predict the optimization of neural network.
2. of the BP neural network forecasting model based on SW-PSO algorithm. The actual wind data set fan speed power curve, the day before the wind power prediction using NWP NWP data information. Due to existence of wind power prediction accuracy of the final error, the wind power history data combined with NWP data the ultra short term wind power forecast.
3. based on wind power prediction, the prediction error and wind power penetration into the spinning reserve, established in wind power system dynamic economic dispatch mathematical model. The model is a multi constraint conditions of large-scale nonlinear programming problems, using SW-PSO algorithm to solve it, and the modified particle adjustment strategy to ensure its flight search in the feasible region. The examples demonstrate that the SW-PSO algorithm in solving the optimization problem of wind power integration in terms of effectiveness and feasibility.
4. the chance constrained programming is introduced into the power system dynamic economic dispatch model with wind farms, with probability to describe the relevant constraints, a stochastic optimal scheduling of wind power grid power system model. Examples show that the application of Monte Carlo simulation technique and SW-PSO algorithm is effective for solving stochastic optimal scheduling problem of power system, can provide both to meet the safe and stable operation of power grid and reasonable planning of the new energy consumptive demand for decision makers.

【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM614

【参考文献】