基于支持向量分位数回归与智能电网的短期电力负荷概率密度预测方法

发布时间：2018-06-19 07:02

本文选题：智能电网 + 支持向量分位数回归　；参考：《合肥工业大学》2017年硕士论文

【摘要】：电力系统负荷预测是电力系统的计划、规划、调度的重要依据。它能够更好地确保电力系统经济安全和稳定运行。环境的可持续发展是人类社会生存和发展的重要基础。智能电网是未来电网发展的必然趋势,且智能电网的实现离不开精确的电力负荷预测方法的支持。大量分布式电源和可再生能源的并网有效地降低了传统能源的消耗,极大地保护了环境。但由于其不稳定性、间歇性等特点,对电网的稳定运行带来了新的问题。针对智能电网建成后带来的电力系统运行和调度的复杂性和不确定性,需要提出新的方法来提高电力负荷预测的精度。在智能电网快速发展的环境下,影响短期电力负荷预测准确性的除了历史负荷和气象因素外,实时电价对短期电力负荷预测的精度影响也较为显著。而且,实时电价是影响电力负荷预测的不确定因素之一。它对用户用电模式的影响较大。随着智能电网的快速发展,其中一个显著的变化就是人们对根据电能需求结合实时电价调整其用电模式。这样可以达到削峰填谷的目的,而且可以提高电网设备的发电效率和能源的利用效率,同时能够降低用户的电费支出。为了提高考虑实时电价的短期电力负荷预测的精度,更好地反映电力负荷的不确定性,本文提出了支持向量分位数回归(SVQR)方法,通过引入松弛变量构造Lagrange函数,得出了不同分位点下的未来一天任意时刻电力负荷的预测结果。同时采用Epanechnikov核函数,将SVQR方法与核密度估计相结合,进行短期电力负荷概率密度预测,可得出较优的预测结果和未来任意时刻的电力负荷准确的波动范围。本文运用支持向量分位数回归方法进行短期电力负荷预测的过程中,考虑到核函数对于此模型至关重要,由此提出了一个新的方法,基于核的支持向量分位数回归和Copula理论的短期电力负荷概率密度预测方法。本文比较了三种不同的核函数并选择适合目标函数的最优核函数,同时,运用Copula理论来分析实时电价和电力负荷之间的关系。此外,本文选择了可靠性准则(PICP)和预测区间带宽(PINAW)来评价输出区间的准确性。本文采用新加坡的历史负荷和实时电价数据集的四个案例,进行短期电力负荷概率密度预测。结果表明:该方法提供了实时电价和电力负荷之间具有相关关系以及能够较好地解决考虑实时电价的短期电力负荷概率密度预测问题。同时,给出了实时电价和电力负荷的关系图,并提供了概率密度预测曲线图以及预测结果和预测区间图来更好地说明所提出的方法的优越性。
[Abstract]:Power system load forecasting is an important basis for the planning, planning and scheduling of the power system. It can better ensure the economic security and stable operation of the power system. The sustainable development of the environment is an important basis for the survival and development of the human society. The smart grid is the inevitable trend of the development of the power grid in the future, and the realization of the smart grid can not be separated from the essence of the power system. The grid of a large number of distributed power and renewable energy effectively reduces the consumption of traditional energy and greatly protects the environment. But because of its instability and intermittent characteristics, it brings new problems to the stable operation of the power grid. The operation of the power system brought about by the completion of the smart grid has been brought. With the complexity and uncertainty of scheduling, new methods should be put forward to improve the accuracy of power load forecasting. In the rapid development of smart grid, the accuracy of short-term power load forecasting has a significant impact on the accuracy of short-term load forecasting, besides the historical load and meteorological factors. The electricity price is one of the uncertain factors that affect the power load forecasting. It has great influence on the mode of electricity consumption. With the rapid development of the smart grid, one of the significant changes is that people adjust the mode of electricity use according to the demand of electric energy and the real time price of electricity. This can achieve the purpose of cutting peak and filling the valley, and can improve the power grid. In order to improve the accuracy of the short-term power load forecasting and better reflect the uncertainty of the power load, the support vector quantile regression (SVQR) method is proposed to construct the Lagrange function by introducing the relaxation variable. At the same time, the prediction results of power load at any time of the next day at different points are given. At the same time, the Epanechnikov kernel function is used to combine the SVQR method with the nuclear density estimation to predict the short-term power load probability density. The better prediction results and the accurate fluctuation range of the power load at any time in the future can be obtained. In the process of short-term power load forecasting with support vector quantile regression, a new method is proposed, based on kernel support vector quantile regression and Copula theory for short-term power load probability density prediction. Three different kernel functions are compared in this paper. The optimal kernel function suitable for the objective function is selected. At the same time, the relationship between real time electricity price and power load is analyzed by using Copula theory. In addition, the reliability criterion (PICP) and prediction interval bandwidth (PINAW) are selected to evaluate the accuracy of the output interval. In this paper, four cases of the historical load and the real time electricity price data set are adopted in this paper. The result shows that the method provides the relationship between real time electricity price and power load, and can better solve the problem of probability density prediction of short-term electric load considering real time electricity price. At the same time, the relationship diagram of real time electricity price and power load is given, and the probability density is provided. The prediction curve and superiority to better prediction results and prediction interval of the proposed method.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM715

【相似文献】