洪维强
个人信息
Personal information
教授 博士生导师
教师英文名称：Hong Wei-Chiang
教师拼音名称：hwq
所在单位：船舶工程学院
职务：Professor
性别：男
学位：管理学博士学位
在职信息：在职
主要任职：Doctoral supervisor
其他任职：Associate Editor for Applied Soft Computing
毕业院校：Da Yeh University
学科：船舶与海洋结构物设计制造曾获荣誉
2026 全球 2% 科学家(年度与终身)
2025 全球 2% 科学家(年度与终身)
2024 全球 2% 科学家(年度与终身)
2023 全球 2% 科学家(年度与终身)
2022 全球 2% 科学家(年度与终身)
2021 全球 2% 科学家(年度与终身)
2020 全球 2% 科学家(年度与终身)
2019 全球 2% 科学家(年度与终身)
2026 ScholarGPS®全球前 0.05%预测专业学者
2025 ScholarGPS®全球前 0.05%预测专业学者
2024 ScholarGPS®全球前 0.05%预测专业学者
2023 ScholarGPS®全球前 0.05%预测专业学者
2022 ScholarGPS®全球前 0.05%预测专业学者
2026 GlobalAuthorID.com全球前十万科学家
2023 GlobalAuthorID.com全球前十万科学家
2023 第21届徐有庠基金会杰出教授奖
2014 第12届徐有庠基金会杰出教授奖
手机版二维码
Mobile QR code
影响因子：5.7
DOI码：10.1007/s11071-023-08922-9
所属单位：哈尔滨工程大学船舶工程学院
发表刊物：Nonlinear Dynamics
刊物所在地：美国
关键字：Empirical wavelet transform (EWT); Support vector regression (SVR); Ant-lion optimization (ALO); Residential electricity consumption forecast
摘要：Accurate forecasting of residential power load provides key support for energy policymakers and power system managers. However, due to the strong volatility and nonlinearity of residential power load, the accuracy of using traditional algorithms to predict residential power load is not high. Given the volatility and nonlinearity of residential power consumption, this paper proposes a feature extraction and antlion hybrid intelligent power forecasting algorithm, namely EWT-S-ALOSVR. Empirical wavelet decomposition (EWT) can extract the features of multiple factors that affect residential electricity consumption to form multiple sub-columns with characteristics. Disassemble the sequence into multiple fluctuation sources, and analyze the characteristics of the fluctuation sources. Support vector regression (SVR) is a linear expression of nonlinear behavior, which can well solve the nonlinearity of residential power load, and the appropriate model parameters are selected by using antlion optimization (ALO) random walk and elite selection methods. The hybrid of the SVR algorithm and ALO optimization can not only effectively identify the fluctuation and nonlinearity of power load, but also predict more accurately. This paper selects the power load of a certain residential working day for analysis and obtains the correlation between the potential characteristics of residential electricity and the selection of parameters through statistical analysis. The model is optimized according to the relationship between parameters and data characteristics. The optimized model’s prediction outcomes are more accurate and robust when compared to those of other models.
合写作者：Ruo-Tong Zhang,Cen-Cen Cao,Yi-Hsuan Yeh
第一作者：范国峰
论文类型：期刊论文
通讯作者：洪维强
学科门类：工学
文献类型：J
卷号：111
期号：2
页面范围：20139-20163
ISSN号：0924-090X
是否译文：否
发表时间：2023-01-01
收录刊物：SCI
发布期刊链接：https://link.springer.com/article/10.1007/s11071-023-08922-9