Anna Samnioti

Figure 1 from the original publication

Flash calculations are essential in reservoir engineering applications, most notably in compositional flow simulation and separation processes, to provide phase distribution factors, known as k-values, at a given pressure and temperature. The calculation output is subsequently used to estimate composition-dependent properties of interest, such as the equilibrium phases’ molar fraction, composition, density, and compressibility. However, when the flash conditions approach criticality, minor inaccuracies in the computed k-values may lead to significant deviation in the dependent properties, which is eventually inherited to the simulator, leading to large errors in the simulation. Although several machine-learning-based regression approaches have emerged to drastically accelerate flash calculations, the criticality issue persists. To address this problem, a novel resampling technique of the ML models’ training data population is proposed, which aims to fine-tune the training dataset distribution and optimally exploit the models’ learning capacity across various flash conditions. The results demonstrate significantly improved accuracy in predicting phase behavior results near criticality, offering valuable contributions not only to the subsurface reservoir engineering industry but also to the broader field of thermodynamics. By understanding and optimizing the model’s training, this research enables more precise predictions and better-informed decision-making processes in domains involving phase separation phenomena. The proposed technique is applicable to every ML-dominated regression problem, where properties dependent on the machine output are of interest rather than the model output itself.

This study provides insights into the experience gained from the investigation of the dynamic behavior of a mature sour hydrocarbon reservoir modeling under an acid gas re-injection process production strategy. The primary objective was to analyze and evaluate the production behavior of proposed injection zones by assessing various injection scenarios and obtaining oil production over time. To achieve that, a workflow was developed to prioritize potential injection areas, select the optimal wells, determine the optimal operational parameters and optimize a pilot application design based on expected performance. Within this framework, the study encompasses diverse acid gas injection schemes on a pilot scale approach, including acid gas combined with waterflooding. The outcome of this analysis will eventually lead to the identification of the most promising and highest-performing injection scheme, elucidating the optimal range of operating parameters. This optimal combination forms the basis for the economic analysis of the venture and the subsequent detailed design of a full-scale application, where real-world implementation will validate the projected results.

In recent years, Machine Learning (ML) has become a buzzword in the petroleum industry, with numerous applications which guide engineers in better decision making. The most powerful tool that most production development decisions rely on is reservoir simulation with applications in multiple modeling procedures, such as individual simulation runs, history matching and production forecast and optimization. However, all of these applications lead to considerable computational time and computer resource-associated costs, rendering reservoir simulators as not fast and robust enough, and thus introducing the need for more time-efficient and intelligent tools, such as ML models which are able to adapt and provide fast and competent results that mimic the simulator’s performance within an acceptable error margin. In a recent paper, the developed ML applications in a subsurface reservoir simulation were reviewed, focusing on improving the speed and accuracy of individual reservoir simulation runs and history matching. This paper consists of the second part of that study, offering a detailed review of ML-based Production Forecast Optimization (PFO). This review can assist engineers as a complete source for applied ML techniques in reservoir simulation since, with the generation of large-scale data in everyday activities, ML is becoming a necessity for future and more efficient applications.

In recent years, machine learning (ML) has become a buzzword in the petroleum industry with numerous applications that guide engineers toward better decision making. The most powerful tool that most production development decisions rely on is reservoir simulation with applications in numerous modeling procedures, such as individual simulation runs, history matching and production forecast and optimization. However, all these applications lead to considerable computational time- and resource-associated costs, and rendering reservoir simulators is not fast or robust, thus introducing the need for more time-efficient and smart tools like ML models which can adapt and provide fast and competent results that mimic simulators’ performance within an acceptable error margin. The first part of the present study (Part I) offers a detailed review of ML techniques in the petroleum industry, specifically in subsurface reservoir simulation, for cases of individual simulation runs and history matching, whereas ML-based production forecast and optimization applications are presented in Part II. This review can assist engineers as a complete source for applied ML techniques since, with the generation of large-scale data in everyday activities, ML is becoming a necessity for future and more efficient applications.

Sour gas in hydrocarbon reservoirs contains significant amounts of H₂S and smaller amounts of CO₂. To minimize operational costs, meet air emission standards and increase oil recovery, operators revert to acid gas (re-)injection into the reservoir rather than treating H₂S in Claus units. This process requires the pressurization of the acid gas, which, when combined with low-temperature conditions prevailing in subsurface pipelines, often leads to the formation of hydrates that can potentially block the fluid flow. Therefore, hydrates formation must be checked at each pipeline segment and for each timestep during a flow simulation, for any varying composition, pressure and temperature, leading to millions of calculations that become more intense when transience is considered. Such calculations are time-consuming as they incorporate the van der Walls–Platteeuw and Langmuir adsorption theory, combined with complex EoS models to account for the polarity of the fluid phases (water, inhibitors). The formation pressure is obtained by solving an iterative multiphase equilibrium problem, which takes a considerable amount of CPU time only to provide a binary answer (hydrates/no hydrates). To accelerate such calculations, a set of classifiers is developed to answer whether the prevailing conditions lie to the left (hydrates) or the right-hand (no hydrates) side of the P-T phase envelope. Results are provided in a fast, direct, non-iterative way, for any possible conditions. A set of hydrate formation “yes/no” points, generated offline using conventional approaches, are utilized for the classifier’s training. The model is applicable to any acid gas flow problem and for any prevailing conditions to eliminate the CPU time of multiphase equilibrium calculations.

In recent years, Machine Learning (ML) has become a buzzword in the petroleum industry with numerous applications which guide engineers in better decision-making. The most powerful tool that most production development decisions rely on is reservoir simulation with applications in numerous modeling procedures, such as individual simulation runs, history matching and production forecast and optimization. However, all these applications lead to considerable computational time and computer resources associated costs, rendering reservoir simulators as not fast and robust enough, thus introducing the need for more time-efficient and smart tools, like ML models which are able to adapt and provide fast and competent results that mimic the simulator’s performance within an acceptable error margin. The first part of the present study (Part I) offers a detailed review of ML techniques in the petroleum industry, specifically in subsurface reservoir simulation, for the cases of individual simulation runs and history matching, whereas the ML-based Production Forecast Optimization applications will be presented in Part II. This review can assist engineers as a complete source for applied ML techniques since, with the generation of large-scale data in everyday activities, ML is becoming a necessity for future and more efficient applications.

This study provides insights into the experience gained from investigating the thermodynamic behavior of well and reservoir fluids during acid gas injection (AGI) in a hydrocarbon field to enhance oil recovery (EOR) and to reduce greenhouse gas emissions. Unlike conventional water and natural gas injection, AGI involves complicated phase changes and physical property variations of the acid gas and reservoir fluids at various pressure-temperature (P-T) conditions and compositions, and both constitute crucial parts of the EOR chain. A workflow is developed to deal with the reservoir fluid and acid gas thermodynamics, which is a key requirement for a successful design and operation. The workflow focuses firstly on the development of the thermodynamic models (EoS) to simulate the behavior of the reservoir fluids and of the injected acid gas and their integration in the field and in well dynamic models. Subsequently, the workflow proposes the thermodynamic simulation of the fluids’ interaction to determine the Minimum Miscibility Pressure (MMP), yielding the dynamic evolution of the fluids’ miscibility that may appear within the reservoir. Flow assurance in the acid gas transportation lines and in the wellbore is also considered by estimating the hydrate formation conditions.

An “energy evolution” is necessary to manifest an environmentally sustainable world while meeting global energy requirements, with natural gas being the most suitable transition fuel. Covering the ever-increasing demand requires exploiting lower value sour gas accumulations, which involves an acid gas treatment issue due to the greenhouse gas nature and toxicity of its constituents. Successful design of the process requires avoiding the formation of acid gas vapor which, in turn, requires time-consuming and complex phase behavior calculations to be repeated over the whole operating range. In this work, we propose classification models from the Machine Learning field, able to rapidly identify the problematic vapor/liquid encounters, as a tool to accelerate phase behavior calculations. To set up this model, a big number of acid gas instances are generated by perturbing pressure, temperature, and acid gas composition and offline solving the stability problem. The generated data are introduced to various classification models, selected based on their ability to provide rapid answers when trained. Results show that by integrating the resulting trained model into the gas reinjection process simulator, the simulation process is substantially accelerated, indicating that the proposed methodology can be readily utilized in all kinds of acid gas flow simulations.