Efficient tracking and estimation of solvent chamber development during warm solvent injection in heterogeneous reservoirs via machine learning

Abstract

Warm solvent injection (WSI) has been proposed as a promising alternative to steam-based methods for bitumen recovery, due to its potential to reduce greenhouse gas emissions and environmental footprint. It involves injecting heated vaporized solvent to reduce the viscosity of bitumen via solvent diffusion and latent heat transfer. The presence of reservoir heterogeneity caused by shale barriers is a severe concern for the success of the WSI because the conformance of solvent chamber advancement can be compromised. However, the efficient estimation and tracking methods for solvent chamber growth and propagation in heterogeneous reservoirs have not been widely investigated in the past. To fill this gap, this work proposes a novel machine learning-based approach to efficiently track solvent chamber positions in heterogeneous reservoirs for the WSI process. From a large training dataset consisting of numerous synthetic heterogeneous models and their simulation results, the input features and output parameters are extracted from oil production time-series data and the dynamic evolution of solvent chamber, respectively. A convolutional neural network (CNN) is implemented to dynamically track solvent chamber positions by correlating the extracted inputs and outputs. The estimation results are reliable and accurate for both scenarios where the shale barriers are either regularly or irregularly shaped with high conformance index (CI). Only production data is used to assess the conformance of solvent chamber advancement, which is an important consideration in operations design and real-time optimization. The presented workflow offers a novel alternative to infer the development of solvent chambers in heterogeneous reservoirs from production time-series data directly. This type of analysis could complement many existing monitoring techniques to deliver a more comprehensive inference of the distribution of shale heterogeneities in solvent-based bitumen recovery operations.