An Interpretable Interflow Simulated Graph Neural Network for Reservoir Connectivity Analysis

Abstract

Summary Reservoir connectivity analysis plays an essential role in controlling water cut in the middle and later stages of reservoir development. The traditional analysis methods, such as well test and tracer, may result in interruption and high reservoir development costs. Analyzing connectivity through history data is an advisable alternative method because the fluctuation of data reflects interwell interference. However, most of the former data-driven methods, such as capacitance and resistance model (CRM), estimate connectivity using formulas in relatively simple forms, leading to inadequate expression for underground interwell flow. In this paper, an interpretable recurrent graph neural network (GNN) is proposed to construct an interacting process imitating the real interwell flow regularity and overcoming the weakness in previous methods. In contrast, it is formed by a deep enough neural network structure with a relatively larger number of parameters when compared with the CRM model. In detail, this method makes the first use of both rate information and bottomhole pressure (BHP) to completely describe the hidden state of wells and the energy information exchanged among them, which are then continually updated in spatial and temporal ways. Meanwhile, a self-defined recurrent structure deals with the time lag and attenuation phenomenon as it records the residual energy from past timestamps. Finally, it calculates BHP for each production well with the manually specified production rate as extra input data. Detailed results are presented in two examples. Our proposed method shows significant advantages to other methods due to its reasonable structure and great ability to fit nonlinear mapping.

Fault characterization and flow barrier detection using capacitance-resistance model and diagnostic plots

“…Arunthavanathan et al (2021) applied a hybrid deep learning approach to successfully detect and diagnose faults in a process system. This showcases the application of domain knowledge transferrable to other disciplines and as utilized in the hybrid networks by Wang et al (2021) and Yu et al (2021).…”

“…Using an interpretable recurrent graph neural network (GNN) as well as historical rate and BHP data, Wang et al (2021) developed a model that simulates the energy exchange or "real interwell flow regularity" between wells, while incorporating the time-lag and attenuation phenomena. Yu et al (2021) applied two-layer neural networks with a sparsity-promoting regularization function for interwell connectivity pattern estimation based on the weight of the network.…”

An Interwell Connectivity Assessment Model for Polymer Flooding Short-Term Development Data Based on A-LSTM and EFAST Methods

“…The method decomposes the wave function of daily oil production data over time from the perspective of improving model data quality, so that the data signal, which is inherently nonlinear and nonsmooth, is transformed into multiple smooth wave functions. Wang et al [16] preprocess the raw data into a custom form so that each sample contains additional local wave information and historical residual energy information, and in predicting long-term production data of bottomhole pressure (BHP), data performed better. However, it is difficult to obtain production data continuously and for a long period of time in the actual production process of oilfields, and there is an urgent need to further optimize and improve the traditional LSTM algorithm so as to adapt it to the prediction work of short-term production data.…”