Research on Short-term Residential Net Load Forecasting Method Considering Data Distribution Shift

doi:10.12204/j.issn.1000-7229.2024.02.008

Abstract

Abstract:

As new energy generation incurs more uncertainty, it leads to a more severe shift in net load data distribution. The data distribution shift means that the feature information learned by the model in historical data is no longer fully applicable to future data, thus posing a significant challenge to net load forecasting (NLF). Therefore, considering the data distribution shift problem in net load, this study proposes a short-term residential net load forecasting method based on IRM-UW-LSTM to improve net load forecasting accuracy. First, a dual-objective problem was established using invariant risk minimization (IRM), which includes accurate forecasting and learning of invariant features across different data distributions. Second, a long short-term memory neural network (LSTM) was used to deal with the nonlinear features of the time series data. Subsequently, an uncertainty weighting (UW)-based objective balancing mechanism was used to avoid overachieving either objective. In addition, a quantile regression method was introduced to extend this study to probabilistic forecasting. Finally, the effectiveness of the proposed method was verified using multiple dimensions of deterministic and probabilistic prediction results, different data distribution shift levels, and different PV penetration rates based on real residential meter data provided by Ausgrid, Australia.

Key words: short-term residential net load forecasting, data distribution shift, invariant risk minimization, long short-term memory neural network, uncertainty weighting

CLC Number:

TM714

WANG Ruilin, ZHAO Jian, SUN Zhiqing, XUAN Yi. Research on Short-term Residential Net Load Forecasting Method Considering Data Distribution Shift[J]. ELECTRIC POWER CONSTRUCTION, 2024, 45(2): 90-101.

Figures/Tables 16

Fig.1 The compare of the degree of the data distribution shift under the same season for different years

Fig.2 The compare of the degree of data distribution shift adjacent months under the same season

Fig.3 Invariant feature information in the net load forecasting

Fig.4 A unit of LSTM

Table 1 Feature establishment

Fig.5 The process of the net load forecasting method based on IRM-UW-LSTM

Table 2 Hyperparameter value

Table 3 Deterministic net load forecasting results

Fig.6 Deterministic net load forecasting results under different methods

Fig.7 Probabilistic net load forecasting results

Table 4 Probabilistic net load forecast evaluation results under different methods

Fig.8 MAAPE and RMSE results of proposed method and the second-best method under different degrees of data distribution shift

Fig.9 Relative improvement of proposed method compared to the second-best method under different degrees of data distribution shift

Fig.10 MAAPE and RMSE results of the proposed method and the suboptimal method under different photovoltaic permeability

Table 5 Net load forecasting results for different RNN networks

Table 6 Computational time for different RNN network

References 27

[1]	SUN M Y, ZHANG T Q, WANG Y, et al. Using Bayesian deep learning to capture uncertainty for residential net load forecasting[J]. IEEE Transactions on Power Systems, 2020, 35(1): 188-201. doi: 10.1109/TPWRS.59 URL
[2]	WANG X, XU Y, FU Z H, et al. A dynamic interactive optimization model of CCHP system involving demand-side and supply-side impacts of climate change. Part II: application to a hospital in Shanghai, China[J]. Energy Conversion and Management, 2022, 252: 115139. doi: 10.1016/j.enconman.2021.115139 URL
[3]	BIJNENS G, KONINGS J, VANORMELINGEN S. The impact of electricity prices on European manufacturing jobs[J]. Applied Economics, 2022, 54(1): 38-56. doi: 10.1080/00036846.2021.1951647 URL
[4]	王晓东, 张一龙, 刘颖明, 等. 基于双时间尺度净负荷预测的储能调峰控制策略[J]. 太阳能学报, 2021, 42(7): 58-64.
	WANG Xiaodong, ZHANG Yilong, LIU Yingming, et al. Double time scale power prediction based peaking shaving control strategy of energy storage system[J]. Acta Energiae Solaris Sinica, 2021, 42(7): 58-64.
[5]	WANG Y, ZHANG N, CHEN Q X, et al. Data-driven probabilistic net load forecasting with high penetration of behind-the-meter PV[J]. IEEE Transactions on Power Systems, 2018, 33(3): 3255-3264. doi: 10.1109/TPWRS.2017.2762599 URL
[6]	LI K P, YAN J C, HU L, et al. Two-stage decoupled estimation approach of aggregated baseline load under high penetration of behind-the-meter PV system[J]. IEEE Transactions on Smart Grid, 2021, 12(6): 4876-4885. doi: 10.1109/TSG.2021.3105747 URL
[7]	FAUSTINE A, PEREIRA L. FPSeq2Q: fully parameterized sequence to quantile regression for net-load forecasting with uncertainty estimates[J]. IEEE Transactions on Smart Grid, 2022, 13(3): 2440-2451. doi: 10.1109/TSG.2022.3148699 URL
[8]	鲍海波, 吴阳晨, 张国应, 等. 基于特征加权Stacking集成学习的净负荷预测方法[J]. 电力建设, 2022, 43(9): 104-116. doi: 10.12204/j.issn.1000-7229.2022.09.011
	BAO Haibo, WU Yangchen, ZHANG Guoying, et al. Net load forecasting method based on feature-weighted stacking ensemble learning[J]. Electric Power Construction, 2022, 43(9): 104-116. doi: 10.12204/j.issn.1000-7229.2022.09.011
[9]	王炜, 冯斌, 黄刚, 等. 基于自注意力编码器和深度神经网络的短期净负荷预测[J/OL]. 中国电机工程学报, 2023(2023-02-24)[2023-07-18]. https://doi.org/10.13334/j.0258-8013.pcsee.221835.
	WANG Wei, FENG Bin, HUANG Gang et al. Short-term net load forecasting based on self-attentive encoder and deep neural network[J/OL]. Proceedings of the CSEE, 2023(2023-02-24)[2023-07-18]. https://doi.org/10.13334/j.0258-8013.pcsee.221835.
[10]	ARJOVSKY M, BOTTOU L, GULRAJANI I et al. Invariant risk minimization[DB/OL]. (2020-03-27)[2023-07-18]. https://arxiv.org/abs/1907.02893.
[11]	RATNAM E L, WELLER S R, KELLETT C M, et al. Residential load and rooftop PV generation: an Australian distribution network dataset[J]. International Journal of Sustainable Energy, 2017, 36(8): 787-806. doi: 10.1080/14786451.2015.1100196 URL
[12]	GHATAK G, MOHANTY H, RAHMAN A U. Kolmogorov-smirnov test-based actively-adaptive Thompson sampling for non-stationary bandits[J]. IEEE Transactions on Artificial Intelligence, 2022, 3(1): 11-19. doi: 10.1109/TAI.2021.3121653 URL
[13]	李旭霞, 张琳娜, 郑晓明, 等. 基于KL散度的储能电站分布鲁棒规划方法[J]. 太阳能学报, 2022, 43(4): 46-55. doi: 10.19912/j.0254-0096.tynxb.2021-0049
	LI Xuxia, ZHANG Linna, ZHENG Xiaoming, et al. KL divergence-based distributionally robust planning method for energy storage plants[J]. Acta Energiae Solaris Sinica, 2022, 43(4): 46-55. doi: 10.19912/j.0254-0096.tynxb.2021-0049
[14]	WANG W, LI H J, DING Z M, et al. Rethinking maximum mean discrepancy for visual domain adaptation[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(1): 264-277. doi: 10.1109/TNNLS.2021.3093468 URL
[15]	孔祥玉, 马玉莹, 艾芊, 等. 新型电力系统多元用户的用电特征建模与用电负荷预测综述[J]. 电力系统自动化, 2023, 47(13): 2-17.
	KONG Xiangyu, MA Yuying, AI Qian, et al. Review on electricity consumption characteristic modeling and load forecasting for diverse users in new power system[J]. Automation of Electric Power Systems, 2023, 47(13): 2-17.
[16]	PETERS J, BÜHLMANN P, MEINSHAUSEN N. Causal inference by using invariant prediction: identification and confidence intervals[J]. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2016, 78(5): 947-1012. doi: 10.1111/rssb.12167 URL
[17]	CIPOLLA R, GAL Y, KENDALL A. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE, 2018: 7482-7491.
[18]	邓带雨, 李坚, 张真源, 等. 基于EEMD-GRU-MLR的短期电力负荷预测[J]. 电网技术, 2020, 44(2): 593-602.
	DENG Daiyu, LI Jian, ZHANG Zhenyuan, et al. Short-term electric load forecasting based on EEMD-GRU-MLR[J]. Power System Technology, 2020, 44(2): 593-602.
[19]	李焱, 贾雅君, 李磊, 等. 基于随机森林算法的短期电力负荷预测[J]. 电力系统保护与控制, 2020, 48(21): 117-124.
	LI Yan, JIA Yajun, LI Lei, et al. Short term power load forecasting based on a stochastic forest algorithm[J]. Power System Protection and Control, 2020, 48(21): 117-124.
[20]	沈兆轩, 袁三男. 利用卷积神经网络支持向量回归机的地区负荷聚类集成预测[J]. 电网技术, 2020, 44(6): 2237-2244.
	SHEN Zhaoxuan, YUAN Sannan. Regional load clustering integration forecasting based on convolutional neural network support vector regression machine[J]. Power System Technology, 2020, 44(6): 2237-2244.
[21]	钟劲松, 王少林, 冉懿, 等. 基于互信息和LSTM的用户负荷短期预测[J]. 电力建设, 2022, 43(7): 96-102. doi: 10.12204/j.issn.1000-7229.2022.07.011
	ZHONG Jinsong, WANG Shaolin, RAN Yi, et al. Short-term consumer load forecasting based on mutual information and LSTM[J]. Electric Power Construction, 2022, 43(7): 96-102. doi: 10.12204/j.issn.1000-7229.2022.07.011
[22]	周思思, 李勇, 郭钇秀, 等. 考虑时序特征提取与双重注意力融合的TCN超短期负荷预测[J]. 电力系统自动化, 2023, 47(18): 193-205.
	ZHOU Sisi, LI Yong, GUO Yixiu, et al. Ultra-short-term load forecasting based on temporal convolutional network considering temporal feature extraction and dual attention fusion[J]. Automation of Electric Power Systems, 2023, 47(18): 193-205.
[23]	庄家懿, 杨国华, 郑豪丰, 等. 基于多模型融合的CNN-LSTM-XGBoost短期电力负荷预测方法[J]. 中国电力, 2021, 54(5): 46-55.
	ZHUANG Jiayi, YANG Guohua, ZHENG Haofeng, et al. Short-term load forecasting method based on multi-model fusion using CNN-LSTM-XGBoost framework[J]. Electric Power, 2021, 54(5): 46-55.
[24]	RAZAVI S E, AREFI A, LEDWICH G, et al. From load to net energy forecasting: short-term residential forecasting for the blend of load and PV behind the meter[J]. IEEE Access, 2020, 8: 224343-224353. doi: 10.1109/Access.6287639 URL
[25]	朱卫涛, 邹文文, 贾钦, 等. 基于DWT-SOM-HFS的配电台区短期负荷预测研究与应用[J]. 智慧电力, 2023, 51(6): 78-85.
	ZHU Weitao, ZOU Wenwen, JIA Qin, et al. Research and application of short-term load forecasting in distribution substation based on DWT-SOM-HFS[J]. Smart Grid, 2023, 51(6): 78-85.
[26]	沈建良, 来骏, 张翼, 等. 基于ACMD和BiGRU-Attention的负荷预测模型研究[J]. 浙江电力, 2023, 42(6): 70-77.
	SHEN Jianliang, LAI Jun, ZHANG Yi, et al. Research on a load forecasting model based on ACMD and BiGRU-Attention[J]. Zhejiang Electric Power, 2023, 42(6): 70-77.
[27]	刘辉, 凌宁青, 罗志强, 等. 基于TCN-LSTM和气象相似日集的电网短期负荷预测方法[J]. 智慧电力, 2022, 50(8): 30-37.
	LIU Hui, LING Ningqing, LUO Zhiqiang, et al. Short-term load forecasting method for power grid based on TCN-LSTM and similar meteorological days set[J]. Smart Grid, 2022, 50(8): 30-37.