基于SVR残差修正的光伏发电功率预测模型

作为一种重要的分布式电源,光伏发电发展迅速且当前部分地区的渗透率不断升高,对区域电网的安全稳定运行造成了严重的影响。光伏功率超短期预测可以为区域电力调度提供必要的数据支撑,促进新能源消纳目标的实现。但是光伏电源自身的波动性特性使光伏功率预测的精度难以提高。鉴于此,本文提出了一种考虑功率修正基于差分自回归移动平均模型(ARIMA)的光伏发电功率预测模型。首先,以光伏电站现场采集的功率时间序列数据建立ARIMA模型进行预测日发电功率的初步预测；其次,利用前一个气象相似日的预测残差数据建立支持向量回归模型对预测日的ARIMA预测残差进行预测；最后,对初步预测结果进行修正。现场实际数据建模证明了本文方法的有效性。#$NL关键词：光伏发电; 功率预测; SVR; ARIMA#$NL中图分类号：TM615

Prediction model of photovoltaic power generation based on SVR residual correction

At present, as an important distributed power source, photovoltaic (PV) power develops rapidly and the penetration rate is increasing continually, which has serious impact on the safe and stable operation of power grid. Ultra-short-term prediction of PV power can provide necessary data support for regional power dispatch which contribute to realization of new energy consumption. However, the volatility characteristics of the PV power make it difficult to improve the accuracy of power prediction. Therefore, a PV power prediction model based on autoregressive integrated moving average model (ARIMA) considering power correction is proposed. Firstly, the ARIMA model is established using time series power data collected by the PV power monitoring system, and preliminary prediction results can be obtained. Secondly, the prediction residuals of the previous meteorological similar day are used to establish a support vector regression (SVR) model to gain the residuals of the prediction day, Finally, the preliminary prediction results are revised by prediction residuals. The effectiveness of proposed ARIMA+SVR model is verified by actual data.