Improved Transformer-PSO Short-Term Electricity Price Prediction Method Considering Multidimensional Influencing Factors

doi:10.16183/j.cnki.jsjtu.2023.065

Abstract

Abstract:

With the construction of a diversified electricity market, the factors affecting electricity prices are gradually increasing, and the market environment has undergone more drastic changes. In order to improve the accuracy of short-term electricity price prediction, an improved Transformer-particle swarm optimization (PSO) short-term electricity price prediction method considering multiple factors affecting electricity prices is proposed. First, based on the consideration of historical electricity prices and historical loads, the relevant factors of electricity price formation are further analyzed. The autocorrelation function is used to analyze the multi-cycle characteristics of electricity price and adjust input sequence, which overcomes the problem of limited prediction accuracy caused by using historical data only and adjusting the input sequence by experience. Then, by combining long short-term memory (LSTM), self-attention mechanism, multi-layer attention mechanism, and adopting a multi-input structure, an improved Transformer model is established to further enhance the ability of the LSTM model to capture long short-term dependencies between different time step information, to overcome the information utilization bottleneck of LSTM, and to adapt to complex multiple sequence inputs including historical electricity prices and various electricity price causes. In addition, the PSO intelligent algorithm is utilized to search for the optimal learning rate of the model at different learning stages, overcoming the limitations of manually adjusting the learning rate. Finally, the PJM market electricity price is used for example analysis. The results show that the proposed short-term electricity price prediction model can be applied to the market environment where electricity prices are affected by various factors and drastic changes, and effectively improve the accuracy of short-term electricity price prediction.

Key words: short-term electricity price prediction, multidimensional influencing factors, autocorrelation analysis, improved Transformer model, particle swarm optimization (PSO)

CLC Number:

TM711

SUN Xin, WANG Simin, XIE Jingdong, JIANG Hailin, WANG Sen. Improved Transformer-PSO Short-Term Electricity Price Prediction Method Considering Multidimensional Influencing Factors[J]. Journal of Shanghai Jiao Tong University, 2024, 58(9): 1420-1431.

Figures/Tables 19

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

Tab.1

Tab.2

Fig.11

Fig.12

Fig.13

Fig.14

Fig.15

Tab.3

Tab.4

References 25

[1]	蒋锋, 何佳琪, 曾志刚, 等. 基于分解-优化-集成学习方法的电价预测[J]. 中国科学: 信息科学, 2018, 48(10): 1300-1315.
	JIANG Feng, HE Jiaqi, ZENG Zhigang, et al. Decomposition-optimization-ensemble learning approach for electricity price forecasting[J]. Scientia Sinica (Informationis), 2018, 48(10): 1300-1315.
[2]	HUANG L, YANG Y B, ZHAO H L, et al. Time series modeling and filtering method of electric power load stochastic noise[J]. Protection & Control of Modern Power Systems, 2017, 2(1): 25.
[3]	CIFTER A. Forecasting electricity price volatility with the Markov-switching GARCH model: Evidence from the Nordic electric power market[J]. Electric Power Systems Research, 2013, 102: 61-67.
[4]	谭忠富, 张金良. 利用多因素小波变换和多变量时间序列模型的日前电价预测[J]. 中国电机工程学报, 2010, 30(1): 103-110.
	TAN Zhongfu, ZHANG Jinliang. Day-ahead electricity price forecasting based on multi-factor wavelet analysis and multivariate time series models[J]. Proceedings of the CSEE, 2010, 30(1): 103-110.
[5]	焦李成, 杨淑媛, 刘芳, 等. 神经网络七十年: 回顾与展望[J]. 计算机学报, 2016, 39(8): 1697-1716.
	JIAO Licheng, YANG Shuyuan, LIU Fang, et al. Seventy years beyond neural networks: Retrospect and prospect[J]. Chinese Journal of Computers, 2016, 39(8): 1697-1716.
[6]	PENG L, LIU S, LIU R, et al. Effective long short-term memory with differential evolution algorithm for electricity price prediction[J]. Energy, 2018, 162: 1301-1314.
[7]	姚子麟, 张亮, 邹斌, 等. 含高比例风电的电力市场电价预测[J]. 电力系统自动化, 2020, 44(12): 49-55.
	YAO Zilin, ZHANG Liang, ZOU Bin, et al. Electricity price prediction for electricity market with high proportion of wind power[J]. Automation of Electric Power Systems, 2020, 44(12): 49-55.
[8]	WANG D Y, LUO H Y, GRUNDER O, et al. Multi-step ahead electricity price forecasting using a hybrid model based on two-layer decomposition technique and BP neural network optimized by firefly algorithm[J]. Applied Energy, 2017, 190: 390-407.
[9]	赵雅雪, 王旭, 蒋传文, 等. 基于最大信息系数相关性分析和改进多层级门控LSTM的短期电价预测方法[J]. 中国电机工程学报, 2021, 41(1): 135-146.
	ZHAO Yaxue, WANG Xu, JIANG Chuanwen, et al. A novel short-term electricity price forecasting method based on correlation analysis with the maximal information coefficient and modified multi-hierachy gated LSTM[J]. Proceedings of the CSEE, 2021, 41(1): 135-146.
[10]	YU L A, MA Y M, MA M Y. An effective rolling decomposition-ensemble model for gasoline consumption forecasting[J]. Energy, 2021, 222: 119869.
[11]	韩升科, 胡飞虎, 陈之腾, 等. 基于GCN-LSTM的日前市场边际电价预测[J]. 中国电机工程学报, 2022, 42(9): 3276-3286.
	HAN Shengke, HU Feihu, CHEN Zhiteng, et al. Day ahead market marginal price forecasting based on GCN-LSTM[J]. Proceedings of the CSEE, 2022, 42(9): 3276-3286.
[12]	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: ACM, 2017: 6000-6010.
[13]	LIM B, ARıK S Ö, LOEFF N, et al. Temporal fusion Transformers for interpretable multi-horizon time series forecasting[J]. International Journal of Forecasting, 2021, 37(4): 1748-1764.
[14]	殷豪, 丁伟锋, 陈顺, 等. 基于长短时记忆网络-纵横交叉算法的含高比例新能源电力市场日前电价预测[J]. 电网技术, 2022, 46(2): 472-480.
	YIN Hao, DING Weifeng, CHEN Shun, et al. Day-ahead electricity price forecasting of electricity market with high proportion of new energy based on LSTM-CSO model[J]. Power System Technology, 2022, 46(2): 472-480.
[15]	张淑清, 李君, 姜安琦, 等. 基于FPA-VMD和Bi-LSTM神经网络的新型两阶段短期电力负荷预测[J]. 电网技术, 2022, 46(8): 3269-3279.
	ZHANG Shuqing, LI Jun, JIANG Anqi, et al. A novel two-stage model based on FPA-VMD and Bi-LSTM neural network for short-term power load forecasting[J]. Power System Technology, 2022, 46(8): 3269-3279.
[16]	梁智, 孙国强, 李虎成, 等. 基于VMD与PSO优化深度信念网络的短期负荷预测[J]. 电网技术, 2018, 42(2): 598-606.
	LIANG Zhi, SUN Guoqiang, LI Hucheng, et al. Short-term load forecasting based on VMD and PSO optimized deep belief network[J]. Power System Technology, 2018, 42(2): 598-606.
[17]	GOODFELLOW I, BENGIO Y, COURVILLE A. Deep learning[M]. Cambridge, Massachusetts: The MIT Press, 2016.
[18]	MAKRIDAKIS S, SPILIOTIS E, ASSIMAKOPOULOS V. The M4 Competition: 100, 000 time series and 61 forecasting methods[J]. International Journal of Forecasting, 2020, 36(1): 54-74.
[19]	丁文娇, 王帮灿, 陈清贵. 基于纳什均衡的PJM日前市场虚拟交易机制探析[C]//中国电机工程学会电力市场专业委员会2019年学术年会暨全国电力交易机构联盟论坛. 成都: 中国电机工程学会电力市场专业委员会, 2019: 329-335.
	DING Wenjiao, WANG Bangcan, CHEN Qinggui. Research of virtual trading mechanism of PJM day-ahead market based on Nash equilibrium[C]//Proceedings of the 2019 Academic Annual Meeting of the Power Market Professional Committee of the CSEE and the National Alliance of Power Trading Institutions Forum. Chengdu, China: Power Market Professional Committee of the CSEE, 2019: 329-335.
[20]	王永茹. 基于TGARCH模型的电价波动性分析[D]. 山东: 山东大学, 2020.
	WANG Yongru. Analysis of electricity price volatility based on TGARCH model[D]. Shandong: Shandong University, 2020.
[21]	彭春华, 刘刚, 相龙阳. 基于Relief相关性特征提取和微分进化支持向量机的短期电价预测[J]. 电工技术学报, 2013, 28(1): 277-284.
	PENG Chunhua, LIU Gang, XIANG Longyang. Short-term electricity price forecasting using relief-correlation analysis based on feature selection and differential evolution support vector machine[J]. Transactions of China Electrotechnical Society, 2013, 28(1): 277-284.
[22]	CHOLLET F. Deep learning with Python[M]. Greenwich: Manning Publications, 2017: 67.
[23]	谢谦, 董立红, 厍向阳. 基于Attention-GRU的短期电价预测[J]. 电力系统保护与控制, 2020, 48(23): 154-160.
	XIE Qian, DONG Lihong, SHE Xiangyang. Short-term electricity price forecasting based on Attention-GRU[J]. Power System Protection & Control, 2020, 48(23): 154-160.
[24]	金海东, 刘全, 陈冬火. 一种带自适应学习率的综合随机梯度下降Q-学习方法[J]. 计算机学报, 2019, 42(10): 2203-2215.
	JIN Haidong, LIU Quan, CHEN Donghuo. Adaptive learning-rate on integrated stochastic gradient decreasing Q-learning[J]. Chinese Journal of Computers, 2019, 42(10): 2203-2215.
[25]	HARVEY D, LEYBOURNE S, NEWBOLD P. Testing the equality of prediction mean squared errors[J]. International Journal of Forecasting, 1997, 13(2): 281-291.

模型	y_RMSE	y_MAE	y_SMAPE
A	2.399	1.760	7.014
B	3.072	2.412	9.569
C	4.592	3.534	14.366
D	3.345	2.648	10.617
E	4.131	3.138	12.796

模型	检验值	P
A、B	-3.813	1.9×10^-4
B、C	-6.755	2.3×10^-10
B、D	-5.020	1.3×10^-6
D、E	-5.595	8.9×10^-8

模型	y_RMSE	y_MAE	y_SMAPE
A	10.510	8.333	11.083
B	13.053	9.944	12.880
C	19.820	14.664	18.130
D	16.578	11.772	14.635
E	19.895	14.462	17.771

模型	检验值	P
A、B	-4.999	9.328×10^-7
B、C	-6.635	1.300×10^-10
B、D	-5.125	5.000×10^-7
D、E	-8.959	2.300×10^-17