基于改进支持向量机算法的光伏发电短期功率滚动预测

Short-term Photovoltaic Power Generation Rolling Forecast Based on Optimized SVM

光伏发电具有明显的波动性与随机性,对其短期功率进行预测可以更准确地实现电网能量管理和运行调度。首先提出了一种基于粒子群优化支持向量机算法(PSO-SVM)的光伏发电短期功率滚动预测模型;通过寻找相似日,以相似日的实际功率和预测日的天气数据作为模型的输入量,对次日一天的发电功率进行预测;再以次日的实际输出功率与预测功率进行滚动对比,当预测点不满足给定预测精度时,以当日实测数据对后期预测点的功率进行修正预测。仿真算例表明所提光伏发电短期功率的滚动预测模型可以更精确地实现功率预测。

Considering the volatility and randomness of photovoltaic （PV） generation systems, short-term forecast of PV power output can accurately achieve the grid scheduling and energy management. This paper proposes a rolling pre- diction model based on support vector machine optimized by particle swarm optimization （PSO-SVM）. Through finding out a similar day to the predicted day, the power output of the similar day and the weather data of the predicted day are taken as the input of the model to forecast the power of the next day. Then, the forecasted power data and actual power of the next day are compared. If the forecasted power cannot satisfy the given forecast accuracy, then the actual power is used to revise the forecasted power. Simulation result shows that the rolling forecast model of the short-term PV power can accurately forecast the power output of PV system.