摘要
为了提高新型电力系统电量趋势预测的精度,文章提出了基于深度神经网络的新型电量趋势预测的方法。采用灰色关联分析深入研究了新能源接入、外部环境以及节假日对电量预测的影响,提出了用粒子群算法优化的Informer电量趋势预测方法。Informer模型采用概率稀疏自注意力机制和自注意力蒸馏机制,使其时间复杂度大大降低,有效地使用内存。通过建立Informer电量预测模型,并对模型的超参数进行优化,可输出最好的电量预测结果。通过算例分析对所提出的方法进行有效性验证,显示该方法具有较高的预测精度,适用于中长期电量趋势预测。
In order to improve the accuracy of electric quantity trend prediction in new power system,a new electric quantity trend prediction method based on deep neural network is proposed in this paper.The influence of new energy access,external environment and holidays on electricity forecast is studied by grey correlation analysis.The method of Informer electric quantity trend prediction optimized by particle swarm optimization algorithm is proposed.Informer model uses probabilistic sparse self-attention mechanism and self-attention distillation mechanism to greatly reduce its time complexity and effectively use memory.By establishing the Informer electric quantity prediction model and optimizing the hyperparameters of the model,the best electric quantity prediction result can be output.The effectiveness of the proposed method is verified by example analysis,which shows that the method has high prediction accuracy and is suitable for medium and long term electric quantity trend prediction.
作者
贺春光
韩光
赵阳
宋楠
He Chunguang;Han Guang;Zhao Yang;Song Nan(Economic and Technological Research Institute of State Grid Hebei Electric Power Co.,LTD.,Shijiazhuang 050000,China;State Grid Hebei Electric Power Co.,LTD.,Shijiazhuang 050000,China)
出处
《可再生能源》
CAS
CSCD
北大核心
2022年第11期1554-1562,共9页
Renewable Energy Resources
基金
国家重点研发计划项目(2016YFB0900100)
国网河北省电力有限公司科技项目(5204JY200001)。
关键词
新能源接入
电量趋势预测
粒子群算法
INFORMER
灰色关联度分析
new energy access
electric quantity trend prediction
particle swarm optimization algorithm
Informer
grey correlation analysis
作者简介
贺春光(1989-),男,高级工程师,研究方向为新能源电力系统规划尧分析尧运行与控制技术。E-mail:majin202205@163.com。
