光伏发电系统的功率预测与接入影响研究

发布时间：2018-07-11 15:04

本文选题：光伏发电 + 神经网络　；参考：《昆明理工大学》2017年硕士论文

【摘要】：随着化石燃料导致的环境问题与能源紧缺问题的出现,太阳能光伏发电系统在世界各国的能源结构的转变中已成为重要部分。但是,分布式光伏发电系统对于外界环境的依赖性较高,造成了其输出功率随着环境因素的变化而变化,其接入主电网后将对电网造成不确定性影响,且分布式光伏发电系统的渗透率越高,电力系统的复杂性与风险也越大。因此,分布式光伏发电系统的功率预测与并网后的配网潮流计算,有助于调度部门提前做好调度计划和风险规避,以提高电力系统的安全性。本文的具体工作如下:(1)将太阳辐射强度、环境温度、天气类别作为光伏系统输出不稳定的主要影响因素,分别基于动量自适应学习速率法的BP神经网络、RBF神经网络以及GRNN神经网络建立了光伏发电系统输出功率的短期预测模型。(2)提出了一种基于决策论极大极小准则的组合预测模型,该组合预测模型由BP神经网络、RBF神经网络、GRNN神经网络组成,对光伏发电系统输出功率进行预测,探索提高输出功率的预测精度。(3)提出了一种改进型的前推回代法,设计不同的接入方案,利用光伏系统输出功率的预测值进行仿真实验,通过前推回代方法计算了含分布式的配网潮流,研究了分布式光伏发电系统接入电网后对配电网络的影响。研究结果表明:将太阳辐射强度、环境温度、天气类别作为影响光伏发电系统输出功率的主要因素能够较好地适用于人工神经网络建立的预测模型,在三种单一预测模型中,GRNN神经网络的输出功率预测精度较高;而基于决策论极大极小准则的组合预测模型的能够进一步提高预测精度;分布式光伏发电系统接入电网后对配电网络的影响为:分布式电源的接入提高了馈线电压的分布,且都能在一定程度上减少系统的网损。
[Abstract]:With the emergence of environmental problems and energy shortages caused by fossil fuels, solar photovoltaic power generation systems have become an important part in the transformation of energy structure in the world. However, the distributed photovoltaic power generation system is highly dependent on the external environment, which causes its output power to change with the change of environmental factors. And the higher the permeability of distributed photovoltaic generation system, the greater the complexity and risk of power system. Therefore, the power prediction of the distributed photovoltaic power generation system and the distribution power flow calculation after the grid connection can help the dispatching department to prepare the dispatch plan and avoid the risk in advance, so as to improve the security of the power system. The main work of this paper is as follows: (1) the solar radiation intensity, ambient temperature and weather type are the main factors that affect the output instability of photovoltaic system. BP neural network based on momentum adaptive learning rate method and GRNN neural network are used to establish short-term prediction models of output power of photovoltaic power generation system. (2) A combined prediction model based on decision theory minimax criterion is proposed. The combined prediction model is composed of BP neural network and RBF neural network and GRNN neural network. The output power of photovoltaic power generation system is predicted, and the prediction accuracy of output power is explored. (3) an improved forward push-back method is proposed. Different access schemes are designed, and the simulation experiments are carried out by using the predicted output power of photovoltaic system. The distribution power flow with distributed distribution network is calculated by the method of forward push back substitution. The influence of distributed photovoltaic system on distribution network is studied. The results show that solar radiation intensity, ambient temperature and weather type are the main factors affecting the output power of photovoltaic power generation system, which can be applied to the prediction model established by artificial neural network. Among the three single prediction models, the output power prediction accuracy of GRNN neural network is higher, while the combined prediction model based on the decision theory minimax criterion can further improve the prediction accuracy. The influence of distributed photovoltaic system on distribution network is as follows: the distribution of feeder voltage is improved by the access of distributed generation, and the network loss can be reduced to a certain extent.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM615

【参考文献】