光伏微电网发电预测与经济运行研究

发布时间：2018-01-29 13:48

本文关键词： 光伏微电网光伏预测经济调度小样本多时段优化　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着社会经济的发展,能源短缺和环境污染问题日益突出,开发与利用可再生能源已成为解决能源和环境问题的重要途径。微电网作为分布式可再生能源友好接入大电网的一种方式,引起了广泛的关注。本文以光伏微电网为研究对象,围绕其发电预测与经济运行问题展开研究,研究工作主要包括:1)分析光伏微电网特性,包括微电网结构特性与各元件特性。重点阐述光伏发电系统的发电原理与输出功率特性、光伏微电网主要元件的数学模型,为光伏微电网发电预测方法研究与经济运行问题研究提供研究基础。2)研究光伏发电预测问题。针对投运初期的光伏电站历史数据不足的问题,提出一种对样本量需求较少的神经网络光伏预测方法。根据光伏发电各环节影响因素的解耦特性,利用改进的双层神经网络模型进行未来一天内的光伏输出功率预测,实现在较少样本量下,对神经网络预测模型的有效训练。基于实际数据,验证了所提出的光伏预测方法在小样本条件下的有效性。3)研究光伏微电网并网经济运行问题。综合考虑光伏发电超短期、短期预测技术的特点与微电网经济运行问题的需求,提出一种混合时长多时段滚动优化方法。在日前调度层的优化模型中,通过前密后疏的时段划分,有效利用光伏发电预测数据的时间粒度与预测精度特点,为实时调度层提供可靠的运行计划,从而保证了系统运行的经济性和安全性。与此同时,所提出的混合时长多时段滚动优化方法减少了日前调度层优化模型的变量数量,减少了问题求解的计算耗时。
[Abstract]:With the development of social economy, energy shortage and environmental pollution problems have become increasingly prominent, the development and utilization of renewable energy has become an important way to solve the problem of energy and environment. The micro grid as a distributed renewable energy friendly grid, has attracted wide attention. In this paper, photovoltaic micro grid as research object, focuses on the research of the power of prediction and economic operation problem, the main work includes: 1) analysis of photovoltaic micro grid characteristics, including the structure characteristics of micro grid and characteristics of each component. Focuses on the principle of power generation and power output characteristics of photovoltaic power generation system, the mathematical model of the main components of photovoltaic micro grid, photovoltaic micro grid power generation prediction method study and research the economic operation provides research foundation for the study of photovoltaic power generation.2) prediction. According to the historical data of the initial operation stage of the photovoltaic power station provided problems. A prediction method for the small sample of demand of the neural network. According to the photovoltaic PV power decoupling characteristics of each link factors, for the PV output power in the next day's forecast by the double improved neural network model, achieved with less sample size, effective prediction model of neural network training. Based on the actual data. Verify the PV prediction effectiveness of the proposed method in.3 under the condition of small sample) of photovoltaic micro grid economic operation problem. Considering the characteristics of photovoltaic power generation of ultra short term, short-term forecasting technology and micro grid economic operation problem, this paper proposes a hybrid optimization method of multi period length of rolling optimization model in Japan. Before scheduling layer, the density after thinning of the time division, the effective use of photovoltaic power generation forecast data and prediction accuracy of time granularity characteristics, to provide reliable real-time scheduling layer The operation plan, so as to ensure the safety and economy of system operation. At the same time, the proposed hybrid long time rolling optimization method to reduce the number of variables of the optimization model for day ahead scheduling layer, reduce the computation time of problem solving.

【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM615

【相似文献】