Time-series production forecasting method based on the integration of Bidirectional Gated Recurrent Unit (Bi-GRU) network and Sparrow Search Algorithm (SSA)

With the gowning demand of improving quality and benefit of unconventional resources, time-series production prediction plays an increasingly essential role in economic investment, stimulation scale and decision making. However, due to the limitations of traditional model-driven methods and data-driven methods, production forecasting is still challenging. In this paper, we propose a novel framework using Bidirectional Gated Recurrent Unit (Bi-GRU) and Sparrow Search Algorithm (SSA) that improves the accuracy of oil rate forecasting. The Bi-GRU could make full use of both past and future information inside production sequences and related features. SSA is employed for hyperparameter tuning of the Bi-GRU model. To validate the feasibility, robustness and efficiency of the proposed method, three cases are carried out from the perspective of an ideal single well from the simulation model, an actual single well under variable production constraints and actual multiple wells. Model performance is compared with traditional decline curve analysis, conventional time series methods and one-way recurrent neural networks. The observations show that the proposed method performs better than the others in terms of accuracy and robustness.