摘要
针对电力负荷序列不稳定且传统的神经网络在电力负荷预测中预测精度较低等问题,提出了门控循环单元神经网络(Gated Recurrent Unit Neural Network, GRU)负荷预测的方法。利用python编程语言在Tensorflow框架在搭建门控循环单元神经网络,将影响电力负荷的特征数据输入模型中进行训练。通过和人工神经网络(Artificial Neural Network,ANN)模型对比,可得出门控循环单元神经网络模型的效果要优于传统神经网络。
Aiming at the problem of unstable power load sequence and low prediction accuracy of traditional Neural Network in power load prediction, a Gated cycle Unit Neural Network(GRU) load prediction method is proposed. Python programming language is used in Tensorflow framework to build a gated circulation unit neural network and train the characteristic data input model that affects power load. By comparing with Artificial Neural Network(ANN) model, it can be concluded that the effect of Neural Network model of out-door control circulation unit is better than that of traditional Neural Network.
作者
陈静
谭爱国
钟建伟
Chen Jing;Tan Ai-guo;Zhong Jian-wei(Hubei Minzu University,School of Information Engineering,Hubei Enshi 445000)
出处
《电子质量》
2022年第2期122-126,共5页
Electronics Quality
关键词
深度学习
数据预处理
神经网络
负荷预测
门控循环单元神经网络
时序特征
deep learning
data preprocessing
neural network
load forecasting
gated recurrent unit neural network
timing characteristics
作者简介
陈静(1995-),女,硕士研究生,研究方向为电力系统负荷预测。
