计及光伏电站功率预测的电力系统优化分析

光伏功率预测多采用间接预测法，由预测太阳辐照度数值结合光转电模型来预测光伏出力。为了解决传统BP算法在短期太阳辐照度预测中易陷入局部最优和收敛速度慢的问题，引入了自适应调节学习率和陡度因子建立太阳辐照度预测模型。在双极性Sigmoid函数中加入陡度因子以提高BP算法的收敛速度，为了便于数据处理将输入数据归一在-1, 1]，同时引入自适应调节学习率以调整网络权值，提高收敛性能。为了研究含光伏电站的电力系统优化问题，建立了系统日综合成本最小和日废气排放量最少的双目标优化模型，并采用双目标细菌群体趋药性算法进行优化。算例证明:改进BP神经网络算法能有效地提高预测精度，增强神经网络模型的泛化能力，具有较好的实用性；预测光伏出力能够统筹安排机组出力，合理消纳光伏资源。

Power system optimization analysis considering power prediction of PV power station

The indirect prediction method is used to forecast the PV power, and the solar irradiance forecast is combined with the light-to-electric model to predict the PV output. In order to solve the shortcomings that the traditional BP algorithm in the short-term solar irradiance prediction is easy to fall into the local optimum and has slow convergence rate, the adaptive adjustment learning rate and the steepness factor are introduced to establish the solar irradiance prediction model. In the bipolar Sigmoid function, the steepness factor is added to improve the convergence speed of the BP algorithm, the normalized input data is limited to -1, 1] to facilitate data processing, and the adaptive adjustment learning rate is introduced to adjust network weight and to improve convergence performance. In order to study the optimization of power system with PV power station, a double objective optimization model to minimize daily composite cost and daily waste gas emission of system is established. And the two-target bacterial population chemotaxis algorithm is used to optimize the system. Examples show that the improved BP neural network algorithm can effectively improve the prediction accuracy and enhance the generalization ability of the neural network model, and it has good practicability. Forecasting the PV output can arrange the unit output and rationally absorb the PV resources. This work is supported by National Natural Science Foundation of China (No. 61573303) and Natural Science Foundation of Hebei Province (No. E2016203092).