• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

电力建设 ›› 2016, Vol. 37 ›› Issue (7): 27-32.doi: 10.3969/j.issn.1000-7229.2016.07.004

• 理论研究 • 上一篇    下一篇

光伏发电系统的时序概率模型研究

金黎明1,周宁1,冯丽1,范飞2,赵渊2   

  1. 1.国网重庆市电力公司,重庆市 400015;2.输配电装备及系统安全与新技术国家重点实验室(重庆大学),重庆市 400044
  • 出版日期:2016-07-01
  • 作者简介:金黎明(1982),男,硕士,高级工程师,长期从事电力系统运行分析和控制的研究工作; 周宁(1972),男,硕士,高级工程师,长期从事电力系统自动化研究工作; 冯丽(1975),女,博士,高级工程师,长期从事电力系统运行分析和控制研究工作; 范飞(1989),男,硕士研究生,从事电力系统规划与可靠性的研究工作; 赵渊(1974),男,教授,博士生导师,长期从事电力系统规划与可靠性的研究工作。
  • 基金资助:

    国家自然科学基金项目(50977094)

Chronological Probability Model of Photovoltaic Generation System

JIN Liming1, ZHOU Ning1, FENG Li1, FAN Fei2,ZHAO Yuan2   

  1. 1. State Grid Chongqing Electric Power Company, Chongqing 400015, China;2. State Key Laboratory of Power Transmission Equipment & System Security and New Technology(Chongqing University), Chongqing 400044, China
  • Online:2016-07-01
  • Supported by:

    Project supported by National Natural Science Foundation of China(50977094)

摘要:

随着光伏发电在电网中的应用越来越广泛,如何建立有效的光伏发电概率模型成为亟待研究的问题。传统的光伏发电模型一般基于参数估计,需要对辐照度的概率分布形式做出预先假设,且无法计及日总辐射与小时辐射间的内在“加和”约束。为了克服传统光伏发电模型存在的不足,利用解集及条件核密度估计技术提出一种新的光伏发电时序概率模型。该模型属于非参数模型,无需对辐照度的概率分布形式做出任何限制,不仅能够计及日/小时辐照度之间的时序相关性,而且能够计及日总辐射与小时辐射间的内在“加和”约束,从而可以更加精确地反映光伏发电的随机变动规律。算例分析表明该模型能够以更高精度反映辐照度的变动规律,因而具有明显的优越性和实用性。

关键词: 光伏发电, 概率模型, 加和约束, 时序相关性, 条件核密度估计

Abstract:

With the growing use of photovoltaic (PV) generation in power system, establishing effective probabilistic model for PV generation becomes an urgent problem to be settled. Conventional chronological probability models of PV generation are based on parametric estimation, which require to assume the probability distribution type of irradiance, and cannot consider the additive constraint between days and hours irradiance sequence. In order to overcome the drawbacks of conventional models, this paper proposes a new photovoltaic sequence probabilistic model based on disaggregation theory and conditional kernel density estimation. Without limiting the probability distribution type of irradiance, the proposed nonparametric model is the non-parametric model, and can capture not only the chronological correlation, but also the additive constraint between days and hours irradiance sequence, which can more accurately reflect the random fluctuation law of photovoltaic generation. The example analysis shows that the model can reflect the change rule of irradiance with higher precision, which has obvious superiority and practicality.

Key words: photovoltaic generation, probabilistic model, additive constraint, chronological correlation, conditional kernel density estimation

中图分类号: