Water Front Proxy simulator

Published: 16 August 2022| Version 1 | DOI: 10.17632/vnbt63r6ny.1
Behzad Saberali, zhang kai, Naser Golsanami


- This source includes the main App code of the LSTM and BiLSTM-based proxy models for determining waterfront location in an injection process. The main program needs several sub-codes and functions for applying, so it can not be run independently. But the proxy models' structure and features can be reused from this source. - The developed proxy models are trained with a mixture of numerical simulation data, common well data, and some pre-defined geometrical data. The trained models can be used for front location simulation independent from any numerical simulation data. - Additional instructions for use are commented through the code's sections. - Some examples of the training performance are attached too.



China University of Petroleum Huadong - Qingdao Campus


Reservoir Simulation, Waterflooding, Deep Learning, Long Short-Term Memory Network