Multi-Convolution Feature Extraction and Recurrent Neural Network Dependent Model for Short-Term Load Forecasting

被引:30
作者
Goh, Hui Hwang [1 ]
He, Biliang [1 ]
Liu, Hui [1 ]
Zhang, Dongdong [1 ]
Dai, Wei [1 ]
Kurniawan, Tonni Agustiono [2 ]
Goh, Kai Chen [3 ]
机构
[1] Guangxi Univ, Coll Elect Engn, Nanning 53000, Peoples R China
[2] Xiamen Univ, Coll Environm & Ecol, Xiamen 361102, Fujian, Peoples R China
[3] Univ Tun Hussein Onn Malaysia, Fac Construct Management & Business, Dept Technol Management, Parit Raja 86400, Johor, Malaysia
关键词
Load modeling; Convolutional neural networks; Predictive models; Feature extraction; Load forecasting; Deep learning; Biological system modeling; Short-term load forecast; deep learning; multi-head CNN-LSTM; multi-step load prediction;
D O I
10.1109/ACCESS.2021.3107954
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Load forecasting is critical for power system operation and market planning. With the increased penetration of renewable energy and the massive consumption of electric energy, improving load forecasting accuracy has become a difficult task. Recently, it was demonstrated that deep learning models perform well for short-term load forecasting (STLF). However, prior research has demonstrated that the hybrid deep learning model outperforms the single model. We propose a hybrid neural network in this article that combines elements of a convolutional neural network (1D-CNN) and a long short memory network (LSTM) in novel ways. Multiple independent 1D-CNNs are used to extract load, calendar, and weather features from the proposed hybrid model, while LSTM is used to learn time patterns. This architecture is referred to as a CNN-LSTM network with multiple heads (MCNN-LSTM). To demonstrate the proposed hybrid deep learning model's superior performance, the proposed method is applied to Ireland's load data for single-step and multi-step load forecasting. In comparison to the widely used CNN-LSTM hybrid model, the proposed model improved single-step prediction by 16.73% and 24-step load prediction by 20.33%. Additionally, we use the Maine dataset to verify the proposed model's generalizability.
引用
收藏
页码:118528 / 118540
页数:13
相关论文
共 34 条
[21]   Short-term load forecasting using lifting scheme and ARIMA models [J].
Lee, Cheng-Ming ;
Ko, Chia-Nan .
EXPERT SYSTEMS WITH APPLICATIONS, 2011, 38 (05) :5902-5911
[22]   A Hybrid CNN-LSTM Model for Forecasting Particulate Matter (PM2.5) [J].
Li, Taoying ;
Hua, Miao ;
Wu, Xu .
IEEE ACCESS, 2020, 8 :26933-26940
[23]   A comparative study of the data-driven day-ahead hourly provincial load forecasting methods: From classical data mining to deep learning [J].
Liu, Xin ;
Zhang, Zijun ;
Song, Zhe .
RENEWABLE & SUSTAINABLE ENERGY REVIEWS, 2020, 119
[24]   Short-term residential load forecasting: Impact of calendar effects and forecast granularity [J].
Lusis, Peter ;
Khalilpour, Kaveh Rajab ;
Andrew, Lachlan ;
Liebman, Ariel .
APPLIED ENERGY, 2017, 205 :654-669
[25]  
Mukhopadhyay P, 2017, 2017 7TH INTERNATIONAL CONFERENCE ON POWER SYSTEMS (ICPS), P812, DOI 10.1109/ICPES.2017.8387401
[26]   Domain Fusion CNN-LSTM for Short-Term Power Consumption Forecasting [J].
Shao, Xiaorui ;
Pu, Chen ;
Zhang, Yuxin ;
Kim, Chang Soo .
IEEE ACCESS, 2020, 8 :188352-188362
[27]  
Sheikh SamsherKadir., 2012, International Journal of Engineering Sciences Emerging Technologies, V1, P97
[28]   A hybrid method based on neural network and improved environmental adaptation method using Controlled Gaussian Mutation with real parameter for short-term load forecasting [J].
Singh, Priyanka ;
Dwivedi, Pragya ;
Kant, Vibhor .
ENERGY, 2019, 174 :460-477
[29]   Integration of new evolutionary approach with artificial neural network for solving short term load forecast problem [J].
Singh, Priyanka ;
Dwivedi, Pragya .
APPLIED ENERGY, 2018, 217 :537-549
[30]  
Sun X., 2017, Proceedings of 2017 Computer Vision and Pattern Recognition (CVPR), P1, DOI [10.1109/EI2.2017.8245401, DOI 10.1109/EI2.2017.8245401]