No abstract

Summary Prediction of well production from unconventional reservoirs is often a complex problem with an incomplete understanding of physics and a considerable amount of data. The most effective way for dealing with it is to use the gray-box approach that combines the strengths of physics-based models and machine learning (ML) used for dealing with certain components of the prediction where physical understanding is poor or difficult. However, the development of methodologies for the incorporation of physics into ML is still in its infancy, not only in the oil and gas industry, but also in other scientific and engineering communities, including the physics community. To set the stage for further advancing the use of combining physics-based models with ML for predicting well production, in this paper we present a brief review of the current developments in this area in the industry, including ML representation of numerical simulation results, determination of parameters for decline curve analysis (DCA) models with ML, physics-informed ML (PIML) that provides an efficient and gridless method for solving differential equations and for discovering governing equations from observations, and physics-constrained ML (PCML) that directly embeds a physics-based model into a neural network. The advantages and potential limitations of the methods are discussed. The future research directions in this area include, but are not limited to, further developing and refining methodologies, including algorithm development, to directly embed physics-based models into ML; exploring the usefulness of PIML for reservoir simulations; and adapting the new developments of how the physics and ML are incorporated in other communities to the well-production prediction. Finally, the methodologies we discuss in the paper can be generally applied to conventional reservoirs as well, although the focus here is on unconventional reservoirs.

Hydraulically fracturing long horizontal wells is the key technology for economically producing hydrocarbon from unconventional reservoirs. A reservoir's fracability (the ease by which it can be hydraulically fractured) has often been used as an important parameter for identifying the sweet spots for production. Several fracability indicators, based on different types of rock properties, including mechanical, geochemical, and mineralogical properties, have been developed and used in industry. This study, based on observations from a source rock reservoir, proposes the use of reservoir water saturation as a new fracability indicator for organic‐rich tight carbonate source rocks that are not clay rich. The results from a machine learning model trained with the observations clearly show the strong and positive correlation between the linear flow parameter (that is obtained based on the newly proposed equivalent‐state approximation and characterizes the effectiveness of hydraulic fracturing) and the water saturation for oil wells, but not for gas wells. While further investigation is needed, the results may be due to the dual wettability of the carbonate source rock. Since minerals are more water‐wet than the organic matter, reservoir water tends to occupy pore spaces in the mineral matrix. Thus, water saturation reflects the relative portion of mineral matrix pore spaces. Given that the low‐clay content mineral matrix contains all the brittle components, the pore‐space development in the mineral matrix may have important implications for the fracability in the hydraulic fracturing process.

DTS/DAS applications provide key advantages in surveillance and better understanding of both unconventional and thermal operations in terms of key attributes including but not limited to conformance, wellbore integrity in better spatial and temporal terms. This study investigates the effects of CO2 in enhancing the steamflood process while incremental benefits are achieved through improved monitoring of the steamflood injection process using DTS/DAS applications using a completely synthetic but realistic reservoir model. A full-physics reservoir simulator is used to model the process. The technical and economic details of deployment of DTS/DAS as well as the steam-additive process are outlined in detail. Sensitivity study carried out on the model indicates the key attributes along with their significance. Athabasca bitumen properties are used. CO2 additive increases the steam chamber size but lowers the steam temperature while naptha/CO2 additives lower the viscosity, thus optimization study carried out the optimum operating levels of the additives not only in physical production/injection terms but also in terms of economics. The results indicate better reservoir management with DTS/DAS applications compared to the base case and injection can be monitored and adjusted better with such tools. DTS/DAS applications prove useful not only in terms of production performance but also in terms of economics. Physical properties of CO2 and naptha outline that the two have different dominant modes of improving recovery with steam-only injection. CO2 increases the extent of the steam chamber while lowering the steam temperature significantly. This study approaches the delicate process of additive use in steam processes while coupling the additional benefits of use of DTS/DAS applications in optimizing the recovery and the economics outlining the key attributes and the challenges and best practices in operations serving as a thorough reference for future applications.

No abstract