Naturally fractured reservoirs are more difficult, complex and expensive to evaluate using numerical simulation when compared to conventional reservoirs. There are well known approaches, dual porosity and dual permeability system in which two grids - one for the fractures and another one for matrix – are used to model the behavior of fracture reservoirs characterized by initial high production followed by a steep decline and then low production for many years. However, most of the time these approaches require a large amount of input data in addition to being computationally too expensive and time consuming in field applications utilizing many grid blocks to model. This paper presents a new pseudo-approach in which a single porosity model can be used for modelling of naturally fractured reservoirs with some modification in the absolute permeability and relative permeabilities. The absolute permeability of the single porosity model is enhanced to capture the effect of permeability from the fracture. This can be a multiplier globally applied to all blocks or local enhancement around the wells having high fracture intensity. Initially the flow is mostly coming from the fracture network so the first "fracture dominant" relative permeability or combination of fracture/matrix relative permeability is used, and later in the lifetime of the reservoir, when the flow transitions to primarily matrix flow, a second "matrix dominated" relative permeability is used to control the fluid flow. The key in this approach is to find the time/date which flow diverted from fracture to matrix. This can be determined from the overall oil rate of the field. After finding the correct date, then the relative permeability is altered from "fracture dominant" to "matrix dominant" recurrently on that time. The new approach is applied to an onshore matured field in Indonesia. The numerical model has the total grid blocks of 1.2 million, 75 wells and around 700 thousand active grid blocks. The original single porosity model could not match the field historical data while dual porosity could captured it correctly. Numerical simulation is utilized along with the new method in a single porosity model for history matching of the field and the results are compared with the dual porosity model of the same model. The absolute permeability enhancement and the first relative permeability curves are used as matching parameters. The results of this study show that both models having same/similar production and pressure profile. Liquid rate, oil rate, water cut, GOR and average pressure are compared. Furthermore, the runtime for the field case improved by 75%. The total runtime of the new approach was 22 hours resulting in significant speed-up compared to the dual porosity runtime of about 4 days. This approach is going to be used for few other fractured reservoirs in the future where time and/or fracture data are limited.

A major Malaysian matured offshore oilfield which is currently under waterflooding has been seen declining in production in recent years. Among various enhanced oil recovery (EOR) techniques evaluated, this field appears to be amenable to chemical EOR implementation. Chemical EOR project requires high capital and operating expenditure (CAPEX and OPEX) and often involve complex logistical and operational challenges in an offshore environment. A comprehensive study and technical road map plan from laboratory to pilot and then to full field reservoir simulation model has been established to reduce the project risks prior to field-scale chemical EOR implementation. For this study, detailed EOR screening and ranking evolution is conducted and confirmed that chemical EOR is ranked high among other EOR techniques and stands for better chance of success techno-economically. Subsequently, all the relevant field examinations to verify the incremental oil recovery from chemical EOR including extensive laboratory experiments such as fluid-fluid and fluid-rock evaluations and pilot tests by single well chemical tracer method were designed and implemented. This paper mainly presents the challenges and the strategies to build a realistic full field chemical EOR numerical simulation model (using CMG's STARS), which include history matching and waterflooding optimization process stages. The work has been carried out to address the best practice workflow for chemical EOR simulation, lessons learned from how to properly prepare and incorporate the chemical input data, identify uncertainties relate to project risks and minimize or mitigate the impact of risks to the project economics. Numerical simulation was utilized along with assisted optimization methods that combine experimental design and artificial intelligence (AI) techniques (using CMG's CMOST) to determine injection chemical concentration and chemical slug size for optimal oil recovery factor and project net present value (NPV). Sensitivity studies were also performed with the reservoir simulation models to determine the impact of effects such as residual oil saturation (Sor) reduction, chemical loss through adsorption, dilution effects on capillary number, salinity and viscosity effects, cooling effects, chemical reactions, among others. The study results show that chemical EOR injection is a technically feasible and economically viable option for this oilfield from subsurface, incremental oil recovery, and facilities stand points. Furthermore, the results of this risk assessment will facilitate and expedite the full field project execution and investment plan in future.

The potential of nanoparticles, which are classified as advanced fluid material, have been unlocked for improved oil recovery in recent years such as nanoparticles-assisted waterflood process. However, there is no existing commercial reservoir simulation software that could properly model phase behaviour and transport phenomena of nanoparticles. This paper focuses on the development of a novel robust advanced simulation algorithms for nanoparticles that incorporate all the main mechanisms that have been observed for interpreting and predicting performance. The general algorithms were developed by incorporating important physico-chemical interactions that exist across nanoparticles along with the porous media and fluid: phase behaviour and flow characteristic of nanoparticles that includes aggregation, splitting and solid phase deposition. A new reaction stoichiometry was introduced to capture the aggregation process. The new algorithm was also incorporated to describe disproportionate permeability alteration and adsorption of nanoparticles, aqueous phase viscosities effect, interfacial tension reduction, and rock wettability alteration. Then, the model was tested and duly validated using several previously published experimental datasets that involved various types of nanoparticles, different chemical additives, hardness of water, wide range of water salinity and rock permeability and oil viscosity from ambient to reservoir temperature. A novel advanced simulation tool has successfully been developed to model advanced fluid material, particularly nanoparticles for improved/enhanced oil recovery. The main scripting of physics and mechanisms of nanoparticle injection are accomplished in the model and have acceptable match with various type of nanoparticles, concentration, initial wettability, solvent, stabilizer, water hardness and temperature. Reasonable matching for all experimental published data were achieved for pressure and production data. Critical parameters have been observed and should be considered as important input for laboratory experimental design. Sensitivity studies have been conducted on critical parameters and reported in the paper as the most sensitive for obtaining the matches of both pressure and production data. Observed matching parameters could be used as benchmarks for training and data validation. Prior to using in a 3D field-scale prediction in Malaysian oilfields, upscaling workflows must be established with critical parameters. For instance, some reaction rates at field-scale can be assumed to be instantaneous since the time scale for field-scale models is much larger than these reaction rates in the laboratory.

A novel developed fines stabilizer (FS) chemical was used to successfully treat an oil well by reducing the decline rate from 17% per month (pre-treatment) to 1.5%/month (post-treatment for a 6-month monitoring period). Following this success story, more qualified candidates are flocking to replicate it. As a result, the challenge now is to predict each candidate's post-treatment performance. The paper focuses on the development of numerical modelling of a fine stabilizer chemical using a chemical reaction approach to understand the mechanistic process and will serve as a benchmark for future technology replication performance prediction. The mechanisms of the novel developed FS chemical were observed in the laboratory through fine particle coagulation and flocculation. The methodology of this study is to translate the mechanisms and key features associated with observed laboratory data into a scripted chemical reaction program that was coupled with a numerical reservoir simulator. Prior to well modelling, laboratory coreflooding data were validated. To properly represent the mechanistic process with a chemical reaction approach and capture the near wellbore effect, a single well model with local grid refinement was developed. The chemical reaction feature provides a versatile toolbox for modeling complex processes involving chemical and physical interactions. The actual production data history matching process with was carried out to investigate key important parameters. The FS chemical reaction was divided into two stages: damage and treatment. The damage was defined as fines deposited in the pore throat plugging and reducing permeability near the wellbore. Fines migration frequently indicates a build-up of fines in the near-wellbore region over time. As a result, the damage caused by these deposited fines reduces permeability. The novel FS chemical will remove deposited fines through micro-flocs and coagulating fines as solid. The history matching process was completed for core data and the mechanistic model with production and pressure data, and acceptable matches were obtained using rate control. The mechanistic model was tested with constrain at bottomhole pressure for few months historical data. It can predict the production performance with good accuracy compared to actual welltest data. Key parameters of FS injection were observed because of this research and can be used as a benchmark in the future to demonstrate the concept of extending oil well productivity and predicting field replication to recurrence the success story.

The deployment at the field-scale of a novel technique to improve oil recovery using nanoparticles injection is challenging. It requires a comprehensive evaluation of a series of laboratory experiments, to translate and validate the mechanisms into a numerical model to predict accurately and reduce uncertainty parameters. This paper describes the application of novel advanced reservoir modeling for nanoparticles from pore-scale to field-scale, using an offshore Malaysian oilfield as a pilot field case. A series of laboratory experiments (fluid-fluid and fluid-rock) and numerical studies: nanofluid formulation, pore-scale studies, validation, and upscaling process into the field-scale model were carried out. The development of nanofluids was formulated to meet key criteria such as compatibility and thermal stability at the intended field condition. Prior to coreflooding tests with native core, a series of experiments to observe mechanisms were carried out. The results of the laboratory experiments were then validated in the 1D coreflooding model. The procedure was continued with observed critical parameters being scaled-up into 3D field-scale model before running the prediction scenarios. The newly developed nanofluids for the intended field performed well in compatibility and thermal stability tests at reservoir temperature. Precipitation and sedimentation were not observed in this solution. The wettability alteration to more water-wet was observed with consistent results through interfacial tension measurements, contact angle measurements, and relative permeability measurements. Coreflooding was performed using native core, and the reduction of residual oil saturation was approximately 25% between pre- and post-nanoflooding. The adsorption of nanofluids was measured to be around 1.12 mg/g of rock. All these results were input into the model and the history match quality index achieved an acceptable match of ~95%. Several critical parameters for the upscaling process were investigated such as reaction rate of particle aggregation, adsorption, and retention factor. During the scale-up process, the velocity of the fluids and pressure drop were conserved because the recovery is sensitive to flooding rate and the viscosity of the fluids are pressure dependent. The field-scale model was run for the intended field location. The potential of using nanoparticles was evaluated and compared to the no further activity scenario giving an additional recovery factor of approximately 1% per year. The developed method of novel robust advanced reservoir modeling for nanoparticles creates a new reference as the first application in the world of novel advanced numerical modeling at field-scale.