基于1DCNN-DACLSTM模型的风电超短期功率预测方法

Ultra-Short-Term Forecasting of Wind Power Based on 1DCNN-DACLSTM Model

风电超短期功率预测是电网运行态势感知的重要基础。针对风电随机波动性将给电网带来高动态扰动的问题,利用同步相量测量装置(PMU)的高频采样能力,提出一种融合一维卷积(1DCNN)和双重注意力卷积长短时记忆网络(DACLSTM)模型的风电超短期功率预测方法。首先,基于具有高精度高密集采样的PMU装置对风电超短期功率进行实时量测。然后,利用1DCNN在特征提取和时间卷积减少计算复杂度方面的优势,充分挖掘由PMU采样得到的风电功率及相关因素量测数据关键特征,进而结合DACLSTM模型自主分析风电功率数据与输入特征间的关联关系,实现基于1DCNN-DACLSTM组合模型的风电超短期功率高动态变化趋势预测。最后以已配置PMU的某实际风电场为例,验证所提方法的可行性和有效性。

Ultra-short-term prediction of wind power is an important basis for the situational awareness of power grid operation.Taking advantage of the high-frequency sampling of PMU,a novel ultra-short-term power prediction method based on the combination of 1DCNN and DACLSTM is proposed to handle the high dynamic disturbance problem of wind power with random fluctuation on power grid operation.Firstly,PMU with high precision and high-frequency sampling is introduced to measure the ultra-short-term wind power in real time.Secondly,by utilizing the advantages of 1DCNN in feature extraction and temporal convolution reducing computation com-plexity,the key features for wind power and its related factors sampled by PMU can be extracted,and then DACLSTM model is used to analyze the relationship between wind power and the input features.Wind power prediction based on the 1DCNN-DACLSTM model can realize the high dynamic trend prediction of wind power.Finally,the validity and feasibility of the proposed method are verified by tak-ing an actual wind farm with the configured PMU as an example.