基于CNN-BPNN的风光抽水蓄能电站联合优化

A Solar-pumping and Storage Power Station Based on CNNBPNN Joint Optimization

由于风电和光伏发电输出功率具有波动性和不稳定性,使其利用效率低、并网困难。本文利用抽水蓄能电站灵活性高的特点,将风力发电和光伏发电输出功率与抽水蓄能发电输出功率进行联合优化。首先,凭借卷积神经网络卷积核参数共享与短期信息提取强的优势,预测风力、光伏发电输出功率;然后,以环境和气象因素为基础,通过BP神经网络构建抽水蓄能电站发电功率预测模型,使最终并网功率为恒定值。本文建立的模型预测结果误差小,预测准确性高,最终优化后的输出功率趋于稳定,可有效提高风电、光伏发电并网的稳定性,减少电力系统中弃风、弃光现象。

Due to the volatility and instability of wind power and photovoltaic power generation,their utilization efficiency is low and grid connection is difficult.In this paper,the output power of wind power generation and pumped storage power generation and the output power of pumped storage power generation are jointly optimized.Firstly,the output power of wind and photovoltaic power generation is predicted by the advantages of convolution kernel parameter sharing and short-term information extraction;then the prediction model of pumped storage power station is constructed based on environmental and meteorological factors,so that the final grid-connected power is constant value.The prediction results of the model established in this paper have small error,high prediction accuracy,and the final optimized output power tends to be stable,which can effectively improve the grid connection stability of wind power and photovoltaic power generation,and reduce the phenomenon of wind abandonment and light abandonment in the power system.