【摘要】 针对光伏发电系统输出功率的随机性和不稳定性导致预测难的问题，采用具有相同季节和天气类型的相似日历史数据作为训练样本，利用最小二乘支持向量机(least squares support veotor machine, LSSVM)建立光伏发电功率预测模型。针对LSSVM模型中核函数宽度函数和惩罚系数选择难的问题，采用自适应柯西变异粒子群算法对这两个参数进行优化，以提高LSSVM模型的预测精度。根据国外光伏电站实测数据对建立的预测模型进行训练。仿真结果表明，所提出的预测模型具有较高的预测精度以及很好的适应性。更多还原

【Abstract】 In view of the difficulty of forecasting caused by the randomness and instability of the output power of photovoltaic(PV)power generation systems, the similar daily historical data with the same season and weather type were used as training samples, and the least squares support vector machine(LSSVM) was used to establish the PV prediction model. Aiming at the difficulty of choosing the width function of the kernel function and the penalty coefficient in the LSSVM model, the adaptive Cauchy mutation particle swarm optimization algorithm was used to optimize these two parameters for improving the prediction accuracy of the LSSVM model. According to the actual measurement data of foreign PV plants, the established prediction model was trained. The simulation results show that the proposed prediction model has high prediction accuracy and good adaptability.更多还原