摘要
为了控制电力系统的稳定运行,需要对电力负荷短期预测方法展开研究。提出融合多传感器数据的电力负荷短期预测方法,首先采用二维小波阈值方法对电力负荷数据进行去噪处理,以提高数据质量;然后选择自编码神经网络对多传感器采集的电力负荷数据进行融合处理;最后,依据选择和融合处理后的数据建立神经网络电力负荷预测模型,完成对电力负荷的短期预测。实验结果表明,融合多传感器数据的预测方法可以有效地处理数据,提高数据质量,并且具有较高的预测精度。
In order to control the stable operation of the power system,it is necessary to conduct research on short-term load forecasting methods.Propose a short-term power load prediction method that integrates multi-sensor data.Firstly,use a two-dimensional wavelet threshold method to denoise the power load data to improve data quality;Then select a self coding neural network to fuse and process the power load data collected by multiple sensors;Finally,based on the selected and fused data,a neural network power load prediction model is established to achieve short-term prediction of power loads.The experimental results show that the prediction method integrating multi-sensor data can effectively process data,improve data quality,and have high prediction accuracy.
作者
周俊宏
李倩
王骅
胡晖
邹文峰
ZHOU Junhong;LI Qian;WANG Hua;HU Hui;ZOU Wenfeng(Huizhou Power Supply Bureau,Guangdong Power Grid Co.,Ltd.,Huizhou,Guangdong,516000,China;Energy Development Research Institute,CSG,Guangzhou 510000,China)
出处
《自动化与仪器仪表》
2024年第8期86-89,94,共5页
Automation & Instrumentation
基金
南方电网公司广东电网有限责任公司基建技术创新专题(0313002022030103XG00002)。
关键词
数据融合
二维小波阈值去噪
数据质量
预测精度
data fusion
two-dimensional wavelet threshold denoising
data quality
prediction accuracy