2DXformer: Dual Transformers for Wind Power Forecasting with Dual Exogenous Variables
Yajuan Zhang, Jiahai Jiang, Yule Yan, Liang Yang, Ping Zhang
TL;DR
This paper tackles wind power forecasting by introducing 2DXformer, a dual-transformer architecture that separately models endogenous and exogenous variables and their inter-variable relationships. It classifies inputs into exogenous static, exogenous dynamic, and endogenous, and uses dedicated blocks (ExTBlock and EnTBlock) along with variable embeddings and a ResidualMLP to capture exogenous influences on the endogenous stream. Empirical results on two real wind-farm datasets show strong improvements over a wide range of baselines, with ablations confirming the importance of dynamic exogenous variables and the ResidualMLP pathway. The approach advances multivariate spatiotemporal forecasting for renewable energy and can be applied to other domains, with code made publicly available.
Abstract
Accurate wind power forecasting can help formulate scientific dispatch plans, which is of great significance for maintaining the safety, stability, and efficient operation of the power system. In recent years, wind power forecasting methods based on deep learning have focused on extracting the spatiotemporal correlations among data, achieving significant improvements in forecasting accuracy. However, they exhibit two limitations. First, there is a lack of modeling for the inter-variable relationships, which limits the accuracy of the forecasts. Second, by treating endogenous and exogenous variables equally, it leads to unnecessary interactions between the endogenous and exogenous variables, increasing the complexity of the model. In this paper, we propose the 2DXformer, which, building upon the previous work's focus on spatiotemporal correlations, addresses the aforementioned two limitations. Specifically, we classify the inputs of the model into three types: exogenous static variables, exogenous dynamic variables, and endogenous variables. First, we embed these variables as variable tokens in a channel-independent manner. Then, we use the attention mechanism to capture the correlations among exogenous variables. Finally, we employ a multi-layer perceptron with residual connections to model the impact of exogenous variables on endogenous variables. Experimental results on two real-world large-scale datasets indicate that our proposed 2DXformer can further improve the performance of wind power forecasting. The code is available in this repository: \href{https://github.com/jseaj/2DXformer}{https://github.com/jseaj/2DXformer}.