An Experimental Study and Numerical Simulation Analysis of Thermal Oxidation Characteristics Based on Kinetic Parameters in Heavy Oil Reservoirs

: In situ combustion (ISC), an efficient and economical method for enhancing heavy oil recovery in high-pressure, high-viscosity, and thermally challenged reservoirs, relies on the kinetics of crude oil oxidation. Despite an increased focus on kinetic models, there is a gap in understanding how oxidation kinetic parameters impact ISC effectiveness in heavy oil reservoirs. This study addresses this by selecting heavy oil samples from the G Block in the Liaohe oilfield and the M Block in the Huabei oilfield and conducting ramped temperature oxidation (RTO), pressure differential scanning calorimetry (PDSC), and thermogravimetric analysis (TGA) experiments. RTO detailed the thermal conversion process, categorizing oxidation into low-temperature oxidation (LTO), fuel deposition (FD), and high-temperature oxidation (HTO) stages. PDSC and TGA provided thermal characteristics and kinetic parameters. The feasibility of fire flooding was evaluated. Using CMG-STARS, an ISC model was established to analyze the impact of kinetic parameter changes. Activation energy significantly affected coke combustion, while the pre-exponential factor had a notable impact on cracking reactions. The recommended values for activation energy and the pre-exponential factor are provided. This study not only guides fire flooding experiments but also supports field engineering practices, particularly for in situ combustion in heavy oil reservoirs.


Introduction
Petroleum resources hold strategic importance as a key global energy source.Over the past few decades, there has been a rapid increase in demand for crude oil [1,2].However, due to many traditional light-to-medium oil fields entering a high water-cut period, the supply of oil and gas is limited and even shows a declining trend.As a result, increasing attention is shifting toward abundant reserves of heavy oil resources [3][4][5][6].
Heavy oil reservoirs are characterized by high formation pressure, high crude oil viscosity, and significant heat loss, presenting production challenges such as sand production and difficulties in lifting.These factors make economically efficient production extremely challenging [7].After years of water injection development practice, it has been observed that due to the high viscosity of crude oil, significant differences in oil-water mobility, and severe reservoir heterogeneity, it is prone to forming preferential water flow channels, leading to limited water flooding effectiveness [8,9].Through the implementation of heavy oil cold production technologies such as foam flooding, chemical viscosity reduction, and gas injection cycles, the water flooding recovery rate has improved.However, challenges still exist, including issues with foam drive gas breakthroughs and a decrease in the effectiveness of multiple cycles of gas injection and production [10].
To address this issue, researchers began exploring the use of air injection, also known as in situ combustion (ISC), to enhance the recovery of heavy oil resources [11,12].This method involves injecting air into the reservoir to create a high-temperature combustion zone.The continuous injection of air maintains the stability of the combustion zone, generating combustion gases that push crude oil from the injection well to the production well [13,14].The ISC technology utilizes underground chemical reactions to thermally decompose heavy oil into light oil, thereby improving the quality of the extracted oil.This process encompasses various displacement mechanisms, including heating and viscosity reduction, chemical modification, steam distillation, non-miscible gas displacement, and carbon dioxide miscible displacement [15,16].ISC technology consumes only about 20% of the fuel compared to other methods, effectively displacing the remaining oil and leading to a theoretical recovery rate exceeding 70%.With its high displacement efficiency, it provides significant economic benefits [13].The technology also possesses various other advantages.Firstly, for reservoirs with high formation pressure and sensitivity to water, the implementation of air injection is more convenient, avoiding difficulties associated with water injection processes [17].Secondly, compared to traditional water flooding methods, air injection has lower costs as it does not require additional water resources [18,19].Currently, this technology has been widely applied in heavy oil reservoirs such as Hongqian, Shuguang, Shengli, Suplacu de Barcau, and Balol oilfields.Through field applications, researchers have actively explored the potential of heavy oil ISC development, mainly manifested in its ease of implementation, high displacement efficiency, and good economic benefits.
Crude oil oxidation kinetics is the fundamental theory behind air injection for enhanced oil recovery.With an increasing understanding of the theory of fire flooding, it has become apparent that the kinetic parameters of chemical reactions play a crucial role in the calculation of fire flooding engineering.Research on the kinetic parameters is primarily focused on laboratory studies involving both physical and numerical simulations [20,21].Commonly used devices for physical simulation and theoretical research include ramped temperature oxidation (RTO) instruments, pressure differential scanning calorimetry (PDSC) devices, thermogravimetric analysis (TGA) instruments, pressure reaction vessels, one-dimensional combustion tubes, and three-dimensional physical models.These methods can simulate the combustion process of crude oil in the reservoir, providing insights into the oxidation characteristics, transformation processes, and the impact of operating conditions under air injection.They are valuable for calculating the kinetic parameters of the reactions [22][23][24][25].Numerical simulations can guide research on in situ combustion mechanisms, assist in case design, and contribute to economic predictions [26,27].
As early as 1968, Bousaid and Ramey began to identify the types of reactions occurring in each stage of crude oil based on the parameters of the products [28].In Al-Saadoon's lowtemperature oxidation (LTO) experiments, it was found that LTO led to the generation of more fuel, which greatly aided in in situ combustion [29].Thomas et al. used thermocouples to measure tail gas under various temperature conditions, obtaining parameters such as fuel deposition rates [30].An ARC for the assessment of thermal hazards was developed by Townsend and Tou, capable of providing time-temperature-pressure data of chemical reactions occurring under adiabatic conditions in samples.Its accuracy and practicality were also discussed [31].Kok comprehensively applied PDSC and TGA to experimentally study the combustion characteristics of heavy oil, observing three stages: low-temperature oxidation, fuel deposition, and high-temperature oxidation [32].In laboratory research, researchers have conducted numerous experiments and numerical simulations.Early studies primarily focused on investigating the oxidation characteristics of heavy oil under single conditions, without comparing different pressure and heating rate conditions.
Vargas et al. successfully demonstrated the three main reaction zones in the combustion process of heavy oil by using an ARC.The results showed that the variation in the combustion process would affect the temperature interval of the oxidation reaction.The sensitivity analysis of the kinetic parameters obtained from the combustion reaction was carried out [33].Dos Santos et al. used an ARC to obtain data from the oxidation reactions of two Brazilian heavy oils and showed that oxygen addition reactions predominate in the temperature range of 200~300 • C, while bond-breaking reactions prevail above 350 • C.Moreover, sand and clay enhance the contact between oil and oxygen, accelerating the reaction and thereby reducing the activation energy [34].Xu and Huang explored the gases produced after the combustion of asphaltene using thermogravimetric-Fourier transform infrared (TG-FTIR) spectroscopy, observing changes in the content of gases such as CO and CO 2 [35].Padilla et al. obtained reaction kinetic models through different RTO experiments using isoconversional methods and validated these models using numerical simulation software [36].Zhao et al. conducted a comparative study on the thermal behavior and combustion kinetics of heavy crude oil and its saturate, aromatic, resin, and asphaltene (SARA) fractions using TG and DSC.It was found that saturates and aromatics experience significant mass loss during the LTO stage, while resins and asphaltenes encounter noticeable mass loss during the fuel deposition (FD) period, leading to the formation of a large amount of coke, which in turn causes high-temperature oxidation (HTO) reactions [37].Liu et al. investigated the activation energy coking characteristics of two heavy oils with different viscosities, densities, and compositions under pyrolysis and oxidative atmospheres.They found that the coke yield for the identical fractions of the two heavy oils did not differ significantly [38].Romel et al. conducted kinetic experiments on Colombian heavy oil using a self-designed one-dimensional model and validated the historical matching with a kinetic model [39].In recent years, there has been a greater emphasis on obtaining kinetic parameters and establishing kinetic models.It has been learned that kinetic parameters almost always depend on the composition of the oil.However, there is currently a lack of research on the impact of the magnitude of kinetic parameters for oil samples on the effectiveness of in situ combustion in heavy oil reservoirs, that is, the adaptability of oil samples to ISC.
To address these issues, it is crucial to take into account the oxidation characteristics of oil samples and incorporate changes in kinetic parameters into numerical simulation analysis.In this study, we chose two different heavy oil samples from the Liaohe oilfield and the Huabei oilfield, with variations in viscosity, density, and composition.Experimental analyses were conducted using RTO, PDSC, and TGA to examine the oxidation characteristics of the oil samples.The experimental equipment and procedures are described in detail.In the RTO experiments, online measurements of the exhaust gas composition were conducted to analyze the thermal conversion process of heavy oil and the influence of operating conditions.This allowed for the determination of the combustion state of the oil and the evaluation of the feasibility of in situ combustion.Additionally, PDSC and TGA experiments were employed to obtain the thermal flow and thermogravimetric characteristics of the oil samples under different pressure and heating rate conditions.Combining these results with the Arrhenius equation, the reaction kinetic parameters were calculated.
Finally, utilizing a CMG-STARS simulator, a combustion drive model was established to analyze the impact of changes in crude oil oxidation kinetic parameters on the combustion drive effect.Previous research primarily derived kinetic parameters from thermal analysis experiments to provide reaction models for reservoirs' ISC numerical simulations.Our study explores the impact of changing pre-exponential factors and activation energies on reservoir recovery rates through numerical simulation.This approach yields suitable crude oil reaction kinetic parameters for ISC, thereby enabling the selection of oil reservoirs for ISC adaptability from the perspective of their oil characteristics.This series of studies serves as a reference for combustion drive experiments under various conditions and provides a theoretical basis for field engineering practices.It holds certain guiding significance for the ISC development of heavy oil reservoirs.

Oils
The crude oil samples used in the experiments were sourced from the G Block of the Liaohe oilfield and the M Block of the Huabei oilfield, labeled as "G" and "M".Table 1 lists the reservoir conditions.The samples were subjected to dehydration and impurity removal.The dehydrated oil samples had a water content of less than 0.5%.Under degassing conditions at 50 • C, the viscosity of G crude oil is 15,000 mPa•s, and the density is 0.947 g/cm 3 .For M crude oil, the viscosity is 119 mPa•s, and the density is 0.908 g/cm 3 .According to petroleum classification standards, both oil samples were classified as heavy oil.The characteristics of crude oil were qualitatively characterized based on four components: saturate, aromatic, resin, and asphaltene (SARA) fractions [40].We employed thin-layer chromatography-flame ionization detection (TLC-FID, Shandong Runyang Instrument Company, Sheng, China) to analyze the SARA fractions, and the steps were as follows: (1) Take a certain amount of crude oil sample, dissolve it in chloroform, and prepare a solution with a concentration of 10-20 mg/mL.(2) Place the chromatography stick on the spotting plate.Using a micro-syringe (Shandong Runyang Instrument Company, Sheng, China), extract 1 µL of the sample solution and spot it at the end of the thin-layer chromatography stick in five separate rounds.Place the chromatography stick in a humidity chamber (saturated NaOH solution) and keep it for 10 min until the solvent evaporates.(3) Place the chromatography stick into the first chromatography jar (n-heptane).Wait for the solvent to migrate 8 cm on the chromatography stick and then remove the stick and place it in a humidity chamber.Allow it to stand for 10 min to allow the eluent to evaporate.(4) Place the chromatography stick into the second chromatography jar (n-heptane mixed with dichloromethane in a volume ratio of 1:1).Wait for the solvent to migrate 5 cm on the chromatography stick and then remove the stick and place it in a humidity chamber.Allow it to stand for 10 min to allow the eluent to evaporate.(5) Place the chromatography stick into the third chromatography jar (n-heptane mixed with isopentanol in a volume ratio of 9:1).Wait for the solvent to migrate 2.5 cm on the chromatography stick and then remove the stick and let it stand for 2 min to allow the eluent to evaporate.(6) Place the chromatography stick with separated samples into a hydrogen flame ionization scanner for scanning to obtain the chromatogram of the oil sample components.
Table 2 lists the SARA information.As shown in Table 2, both oil samples had high resin and asphaltene content, classifying them as heavy oil.Oil sample M had higher saturate and aromatic content, indicating a relatively lower hydrogen-to-carbon ratio, closer to light oil.RTO experiments were used to simulate the thermal conversion process of crude oil under programmed heating conditions.These experiments were carried out to simulate the heating process of reservoir crude oil in ISC technology and, to some extent, mimic the displacement process.In these experiments, oil sand was used as the sample, and in the heating experiment, the thermal conversion process of crude oil was analyzed by the online measurement of exhaust gas components.We employed a self-developed RTO structure, with a maximum experimental pressure of 5.0 MPa and a maximum experimental temperature of 600 • C.
The RTO structure consisted of a gas circuit system, a tubular furnace, an output gas collection system, and a software control system.Figure 1 displays the schematic diagram of the RTO system.The core of the system was the tubular furnace, consisting of an outer insulation layer and a central reaction tube filled with sand.To better measure the temperature changes during the reaction process, the tubular furnace was divided into three equal-length sections, the entrance section, middle section, and exit section, each with a length of 10 cm.Thermocouple sleeves were placed between the outer insulation layer and the central reaction tube.Figure 2 displays the central reaction tube.Three thermocouples were welded at the central position of each section at intervals of 10 cm to measure the temperature in these three regions.During the experimental process, the injected gases were air and nitrogen and were controlled with two flowmeters.The outlet gas underwent continuous analysis after cooling and drying treatments by connecting the output pipeline to a gas analyzer.

Procedures
For the experiment, degassed crude oil samples were selected from both the G Block and the M Block, conducting three sets of experiments at different pressures with a heating rate of 5.0 K/min.The pressures were set at 1.0, 3.0, and 5.0 MPa, totaling six experimental groups.Table 3 lists the RTO experiment conditions, and the steps are described below.(1) Mix dehydrated crude oil with silicon dioxide at a mass ratio of 3% crude oil to 97% silicon dioxide to prepare the experimental oil sand.(2) Weigh approximately 4.0 g of oil sand and place it at the central position of the reaction tube.Fill the remaining space with densely packed pure silica.Connect the necessary thermocouples.(3) Add insulation to the reaction tube filled with oil sand and place it in the heating furnace.Connect the injection and output pipelines, gas analyzer, and experimental circuit.(4) Set the experimental pressures to 1.0, 3.0, and 5.0 MPa; the heating rate to 5.0 K/min; and the air injection rate to 400 mL/min.Use two injection channels during the experiment, with nitrogen as the protective gas and air as the reaction gas entering the reaction tube and reacting with the sample.While the experimental setup shared similarities with the one-dimensional combustion tube experiment on a system level, it was significantly smaller in terms of the physical model.In this physical model, the oil sand was distributed only at the central position of the reaction tube.It was analogous to taking a single grid from the one-dimensional combustion tube experiment, with the remaining space filled with pure silica.This arrangement was designed to ensure optimal contact between the crude oil sample and air, thereby providing a more accurate representation of the temperature elevation process of crude oil within a single grid.

Procedures
For the experiment, degassed crude oil samples were selected from both the G Block and the M Block, conducting three sets of experiments at different pressures with a heating rate of 5.0 K/min.The pressures were set at 1.0, 3.0, and 5.0 MPa, totaling six experimental groups.Table 3 lists the RTO experiment conditions, and the steps are described below.(1) Mix dehydrated crude oil with silicon dioxide at a mass ratio of 3% crude oil to 97% silicon dioxide to prepare the experimental oil sand.(2) Weigh approximately 4.0 g of oil sand and place it at the central position of the reaction tube.Fill the remaining space with densely packed pure silica.Connect the necessary thermocouples.
(3) Add insulation to the reaction tube filled with oil sand and place it in the heating furnace.Connect the injection and output pipelines, gas analyzer, and experimental circuit.(4) Set the experimental pressures to 1.0, 3.0, and 5.0 MPa; the heating rate to 5.0 K/min; and the air injection rate to 400 mL/min.Use two injection channels during the experiment, with nitrogen as the protective gas and air as the reaction gas entering the reaction tube and reacting with the sample.(5) Start the experiment and record the changes in O 2 , CO 2 , and CO component concentrations.(6) End the experiment when the temperature in the RTO experiment reaches 550 • C.

Equipment
PDSC experiments were used to measure the exothermic capacity of crude oil at low and high temperatures under different pressure conditions.The instrument took the oil sand as the sample, and temperature elevation experiments were conducted.The thermal transformation process of crude oil was analyzed by monitoring the relationship between temperature and the heat signal.The thermal flux difference was monitored using NiCr Constantan thermocouples.For this experiment, we used DSC-Q20P DSC from TA Instruments (New Castle, DE, USA), with a maximum experimental pressure of 7.0 MPa and a maximum experimental temperature of 550 • C.
The DSC structure consisted of the computer operating system, the reaction furnace system, and the gas circuit system.The reaction furnace system included two types: heat flow and heat compensation.Figure 3 displays the schematic diagram of a heat-flow PDSC including a balance, a gas circuit, wires, and a high-pressure chamber.The heat was transferred to the sample through a copper plate, which also served as part of the thermocouple nodes for temperature measurement.The equipment was used to conduct tests on the sample together with the computer control system.

Equipment
PDSC experiments were used to measure the exothermic capacity of crude oil at low and high temperatures under different pressure conditions.The instrument took the oil sand as the sample, and temperature elevation experiments were conducted.The thermal transformation process of crude oil was analyzed by monitoring the relationship between temperature and the heat signal.The thermal flux difference was monitored using NiCr Constantan thermocouples.For this experiment, we used DSC-Q20P DSC from TA Instruments (New Castle, DE, USA), with a maximum experimental pressure of 7.0 MPa and a maximum experimental temperature of 550 °C.
The DSC structure consisted of the computer operating system, the reaction furnace system, and the gas circuit system.The reaction furnace system included two types: heat flow and heat compensation.Figure 3 displays the schematic diagram of a heat-flow PDSC including a balance, a gas circuit, wires, and a high-pressure chamber.The heat was transferred to the sample through a copper plate, which also served as part of the thermocouple nodes for temperature measurement.The equipment was used to conduct tests on the sample together with the computer control system.

Procedures
For these experiments, degassed crude oil samples were selected from both the G Block and M Block, conducting three sets of experiments at different pressures with a heating rate of 10.0 K/min.The pressures were set at 1.0, 3.0, and 5.0 MPa, totaling six experimental groups.Table 4 lists the PDSC experiment conditions, and the steps are explained below.(3) Set the experimental pressure to 1.0, 3.0, and 5.0 MPa; the heating rate to 10.0 K/min; and the airflow rate to 50 mL/min.During the experiment, divide nitrogen and air into two injection channels, with nitrogen as the protective gas and air as the reaction

Procedures
For these experiments, degassed crude oil samples were selected from both the G Block and M Block, conducting three sets of experiments at different pressures with a heating rate of 10.0 K/min.The pressures were set at 1.0, 3.0, and 5.0 MPa, totaling six experimental groups.Table 4 lists the PDSC experiment conditions, and the steps are explained below.(1) Mix dehydrated crude oil with silicon dioxide at a mass ratio of 10% crude oil to 90% silicon dioxide to prepare experimental oil sand.Weigh approximately 30.0 mg for each experiment and place it in a crucible.(2) Start the DSC power supply.After the instrument is fully started, open the pressure chamber cover, place the sample crucible in the reaction chamber, and cover it successively.(3) Set the experimental pressure to 1.0, 3.0, and 5.0 MPa; the heating rate to 10.0 K/min; and the airflow rate to 50 mL/min.During the experiment, divide nitrogen and air into two injection channels, with nitrogen as the protective gas and air as the reaction gas entering the reaction chamber and reacting with the sample on the crucible surface.(4) End the experiment when the temperature in the DSC experiment reaches 550 • C. TGA experiments were used to determine the relationship between the mass of crude oil samples and temperature changes under various heating rates.This equipment took the oil sand as the sample, conducted temperature ramp experiments, measured the mass changes in the test sample with temperature, and calculated the kinetic parameters of the thermal conversion reactions of crude oil.It was employed to study the thermal stability of the materials.For these experiments, a TGA55 Thermal Gravimetric Analyzer from TA Instruments was used, with a temperature ramp range of 0.1-100.0K/min and a maximum experimental temperature of 1000 • C.
The TGA structure consisted of the computer operating system, the reaction furnace system, and the gas circuit system.The most commonly used measurement principles for TGA include two methods: the displacement method and the zero-point method.Figure 4 displays the schematic diagram of a zero-point TGA.The basic principle of the zero-point method is to detect the balance displacement caused by the change in sample weight and convert this small electrical quantity into an electromagnetic quantity.This small electrical quantity is amplified and sent to a recorder for recording.The magnitude of the electrical quantity is proportional to the change in sample weight, ultimately yielding the relationship between the weight signal and temperature.With these data, the thermal conversion process of crude oil was analyzed, and the kinetic parameters of the thermal conversion reaction were calculated.The structure of the reaction furnace included a balance, a gas path, an electrical wire, and a high-pressure chamber.The equipment was used in combination with the computer control system to complete sample testing.

Equipment
TGA experiments were used to determine the relationship between the mass of crude oil samples and temperature changes under various heating rates.This equipment took the oil sand as the sample, conducted temperature ramp experiments, measured the mass changes in the test sample with temperature, and calculated the kinetic parameters of the thermal conversion reactions of crude oil.It was employed to study the thermal stability of the materials.For these experiments, a TGA55 Thermal Gravimetric Analyzer from TA Instruments was used, with a temperature ramp range of 0.1-100.0K/min and a maximum experimental temperature of 1000 °C.
The TGA structure consisted of the computer operating system, the reaction furnace system, and the gas circuit system.The most commonly used measurement principles for TGA include two methods: the displacement method and the zero-point method.Figure 4 displays the schematic diagram of a zero-point TGA.The basic principle of the zeropoint method is to detect the balance displacement caused by the change in sample weight and convert this small electrical quantity into an electromagnetic quantity.This small electrical quantity is amplified and sent to a recorder for recording.The magnitude of the electrical quantity is proportional to the change in sample weight, ultimately yielding the relationship between the weight signal and temperature.With these data, the thermal conversion process of crude oil was analyzed, and the kinetic parameters of the thermal conversion reaction were calculated.The structure of the reaction furnace included a balance, a gas path, an electrical wire, and a high-pressure chamber.The equipment was used in combination with the computer control system to complete sample testing.

Procedures
For these experiments, degassed crude oil samples were selected from both the G Block and the M Block.Under a nitrogen atmosphere, three sets of experiments were conducted at different heating rates, specifically set at 3.0, 5.0, and 7.0 K/min, totaling six experimental groups.Table 5 lists the TGA experiment conditions, and the steps are described below.

Procedures
For these experiments, degassed crude oil samples were selected from both the G Block and the M Block.Under a nitrogen atmosphere, three sets of experiments were conducted at different heating rates, specifically set at 3.0, 5.0, and 7.0 K/min, totaling six experimental groups.Table 5 lists the TGA experiment conditions, and the steps are described below.(1) Mix dehydrated crude oil with silicon dioxide at a mass ratio of 10% crude oil to 90% silicon dioxide to prepare experimental oil sand.Weigh approximately 50.0 mg for each experiment and place it in a crucible.(2) Open the nitrogen chamber, adjust the pressure to 0.1 MPa, and continuously inject nitrogen as a protective gas to maintain atmospheric pressure.(3) Start the TGA power supply.After the instrument is fully started, open the pressure chamber cover, place the sample crucible on the balance in the reaction chamber, and then cover it successively.
(4) Set the experimental heating rates to 3.0, 5.0, and 7.0K/min.End the experiment when the temperature in the TGA experiment rises from room temperature (25 • C) to 650 • C.

Results and Discussion
3.1.Experiment 1: The Thermal Reaction Process Figure 5 shows the RTO results of air composition variation.As shown in Figure 5, the main components of the oxidation reaction off-gas for both heavy oils were CO 2 and CO. Figure 5 shows that when the reaction temperature reaches 200-300 • C, the curve exhibits the first peaks of O 2 consumption and the production of CO and CO 2 .This indicates that, at this temperature range, low-temperature oxidation reactions were most intense.As the reaction temperature increases to 400-500 • C, the curve shows the second peaks of the reaction.However, compared to the first set of peaks, the changes in the components in the air were relatively small.This indicates that, under high-temperature conditions, another round of oxidation reactions occurred but with lower intensity.
each experiment and place it in a crucible.
(2) Open the nitrogen chamber, adjust the pressure to 0.1 MPa, and continuously inject nitrogen as a protective gas to maintain atmospheric pressure.(3) Start the TGA power supply.After the instrument is fully started, open the pressure chamber cover, place the sample crucible on the balance in the reaction chamber, and then cover it successively.(4) Set the experimental heating rates to 3.0, 5.0, and 7.0K/min.End the experiment when the temperature in the TGA experiment rises from room temperature (25 °C) to 650 °C.

Experiment 1: The Thermal Reaction Process
Figure 5 shows the RTO results of air composition variation.As shown in Figure 5, the main components of the oxidation reaction off-gas for both heavy oils were CO2 and CO. Figure 5 shows that when the reaction temperature reaches 200-300 °C, the curve exhibits the first peaks of O2 consumption and the production of CO and CO2.This indicates that, at this temperature range, low-temperature oxidation reactions were most intense.As the reaction temperature increases to 400-500 °C, the curve shows the second peaks of the reaction.However, compared to the first set of peaks, the changes in the components in the air were relatively small.This indicates that, under high-temperature conditions, another round of oxidation reactions occurred but with lower intensity.As shown in Figure 5(a1-c1), with increasing pressure, the partial pressure of oxygen increased, resulting in a larger contact area between oxygen and crude oil.Intense oxidation reactions occurred at the gas-liquid and gas-solid interfaces, accelerating the rate of crude oil oxidation.This led to a more thorough oxidation of crude oil.Oxygen consumption during the oxidation stage increased, and the reaction time decreased.As shown in Figure 5(a1,a2,c1,c2), under the same experimental conditions, the low-temperature oxygen consumption peak for the oil sample G was in the range of 1.2% to 2.2%.Oil sample M had an oxygen consumption rate close to 8% under 5.0 MPa conditions.It had a stronger low-temperature oxygen consumption capacity, and the higher the pressure, the greater the difference in oxygen consumption.
According to the variation curve of air injection components under different pressures, the experimental data for the G Block crude oil at 3.0 MPa were extracted, and the integration of these data across the entire reaction range yielded an atomic ratio of H/C at 1.802.The atomic ratio of H/C for low-temperature oxidation was determined as 3.125, while for high-temperature oxidation, it was 0.733.Using these values, the key reaction equations characterizing the combustion process were determined as follows: Figure 6 shows the oxidation stages division of air composition variation at 3.0 MPa with different oil samples.The oxidation stages are delineated at the inflection points of the curves, and based on the results shown in Figure 6, the oxidation process of crude oil can be summarized in three stages as follows: (1) Low-temperature oxidation (LTO) stage: In this stage, the dominant reaction in the oxidation process is light oil burning.The light components in crude oil undergo intense oxidation reactions, and hydrocarbons undergo oxygenation reactions, leading to the formation of alcohols, aldehydes, ketones, esters, etc.This is typically accompanied by the formation and breaking of carbon-oxygen and carbon-hydrogen bonds.Due to the higher content of light components in the M oil sample, the reaction intensity of the M oil sample was higher than that of the G oil sample in this stage.As the temperature increased, the oxidation reaction rate gradually decreased, laying the foundation for the subsequent stages.(2) Fuel deposition (FD) stage: In this stage, the dominant reaction is heavy oil burning, accompanied by heavy oil cracking.The consumption of O 2 slows down.This stage is characterized by the oxidative cracking of heavy oil, leading to the formation of coke and light hydrocarbons as fuels for the HTO phase.Typically, carbon-carbon and carbon-hydrogen bonds break, resulting in the formation of radicals and carboncarbon double bonds.This stage provides the basis for high-temperature oxidation.(3) High-temperature oxidation (HTO) stage: In this stage, the dominant reaction is coke oil burning, accompanied by heavy oil cracking.As crude oil undergoes thermal cracking, coke and light hydrocarbons are formed as fuels for combustion, leading to a significant high-temperature oxidation reaction.This process involves the breaking of carbon-carbon and carbon-hydrogen bonds, as well as the formation of carbonoxygen and hydrogen-oxygen bonds.Due to the higher content of heavy components in the G oil sample, the reaction intensity of the G oil sample was higher than that of the M oil sample, which had components closer to light oil in this stage.

Experiment 2: The Heat-Flow Process
Figure 7 shows the PDSC results of the heat-flow experiment.As shown in Figure 7, before 200 °C, there was no significant exothermic change in either oil sample.With the increase in temperature, the crude oil sample underwent an evaporation phase change.When the reaction temperature was below 200 °C, the evaporation phase change exceeded low-temperature oxidation, resulting in a constant zero heat-flow curve without any changes.As the temperature increased between 200 °C and 300 °C, the samples began to

Experiment 2: The Heat-Flow Process
Figure 7 shows the PDSC results of the heat-flow experiment.As shown in Figure 7, before 200 °C, there was no significant exothermic change in either oil sample.With the increase in temperature, the crude oil sample underwent an evaporation phase change.When the reaction temperature was below 200 °C, the evaporation phase change exceeded low-temperature oxidation, resulting in a constant zero heat-flow curve without any changes.As the temperature increased between 200 °C and 300 °C, the samples began to show changes in heat flow, marked by the appearance of the first heat-flow peak.In this temperature range, the exothermic heat released during the low-temperature oxidation of crude oil exceeded the endothermic heat associated with the evaporation process.When the temperature reached between 400 °C and 500 °C, a second heat-flow peak appeared.It was lower compared to the first peak, with a more pronounced decrease, especially in the M oil sample.With increasing pressure, under the condition of sufficient contact between the oil sample and air, the entire heat-flow curve shifts to the left, indicating progressively lower reaction temperature thresholds for each stage.The heat release in the LTO stage increased significantly, while there was almost no change in the heat release in the HTO stage.From Figure 7b, it can be observed that, at 5.0 MPa pressure, oil sample M shows a sharp increase in heat release during the low-temperature oxidation stage, rising dramatically around 270 °C, where the heat flux becomes a straight line.Subsequently, the heat flow rapidly changes to zero.It is speculated that due to the higher proportion of light With increasing pressure, under the condition of sufficient contact between the oil sample and air, the entire heat-flow curve shifts to the left, indicating progressively lower reaction temperature thresholds for each stage.The heat release in the LTO stage increased significantly, while there was almost no change in the heat release in the HTO stage.From Figure 7b, it can be observed that, at 5.0 MPa pressure, oil sample M shows a sharp increase in heat release during the low-temperature oxidation stage, rising dramatically around 270 • C, where the heat flux becomes a straight line.Subsequently, the heat flow rapidly changes to zero.It is speculated that due to the higher proportion of light components in this oil sample, under high pressure, combustion may have occurred at a certain temperature, leading to the significant consumption of crude oil.

Experiment 3: The Weight Loss Process and Dynamic Parameter Analysis
Figure 8 shows the weight loss rates with different heating rates.Figure 9 shows the conversion rates with different heating rates.The ordinate α in the conversion rate curve represents the conversion rate, and its equation is as follows: where m, m 0 , and m f are the mass of the sample during the reaction, the initial mass of the sample, and the final mass of the sample, respectively.The ordinate of the reaction rate curve is dα/dt, with the first derivative of the conversion rate α with respect to reaction time representing the rate of oxidation reaction.
loss rates for both stages increased, and the peak temperatures shifted to the right.It is worth noting that the oil sample exhibited weight loss around 50 °C, attributed to the light oil content starts evaporating at very low temperatures, possibly even at room temperature.Additionally, the lighter components inherent in crude oil contributed to noticeable weight loss changes, making it challenging to perform a definitive assessment of the weight loss phenomenon.
(a) (b) Figure 9 shows the conversion rates with different heating rates.In Figure 9, the conversion rate curves of the two oil samples exhibit a high degree of similarity throughout the entire range.With the increase in heating rate, the time required for crude oil to reach a 100% conversion rate also increased correspondingly.This is attributed to the relatively longer reaction time of crude oil at lower heating rates, resulting in more thorough reactions.

Determination of Kinetic Parameters
The kinetic parameters of crude oil oxidation include activation energy and the preexponential factor [41].In this study, the activation energy and pre-exponential factor were calculated based on the Flynn-Wall-Ozawa method and the Arrhenius equation by using the TGA experimental results.The activation energy and pre-exponential factor are determined using Equations ( 2) and ( 3) with characteristic temperatures at the same conversion rate under different heating rates [42,43].The equations are as follows: exp / where E is the activation energy, kJ/mol; R is the general gas constant, and its value is 8.314, J/(mol•K); β is the heating rate, K/min; Tα is the temperature at a specific conversion, K; A is the pre-exponential factor, (kPa•s)⁻¹; and PO2 is the partial pressure of oxygen, Pa.
For the TGA experiments on the two different oil samples, a median heating rate of β = 5.0 K/min was determined.The characteristic temperatures for oil sample G were obtained from the TGA curve at the rate of 5.0 K/min as Tα1 = 765.75K, and for oil sample M, it was Tα2 = 737.95K. Subsequently, the specified conversion rates corresponding to these temperatures were determined based on the conversion rate and temperature data as α1 = 0.965 for oil sample G and α2 = 0.940 for oil sample M.
According to the characteristic conversion rates, Table 6 presents the comparison of oxidation kinetic parameters with equal conversion.Figure 8 shows that the peaks in the weight loss rate of oil sample G in the LTO and HTO stages are similar.The peak weight loss rate of oil sample M in the HTO stage was slightly higher than that in the LTO stage.With an increase in heating rate, the peak weight loss rates for both stages increased, and the peak temperatures shifted to the right.It is worth noting that the oil sample exhibited weight loss around 50 • C, attributed to the light oil content starts evaporating at very low temperatures, possibly even at room temperature.Additionally, the lighter components inherent in crude oil contributed to noticeable weight loss changes, making it challenging to perform a definitive assessment of the weight loss phenomenon.
Figure 9 shows the conversion rates with different heating rates.In Figure 9, the conversion rate curves of the two oil samples exhibit a high degree of similarity throughout the entire range.With the increase in heating rate, the time required for crude oil to reach a 100% conversion rate also increased correspondingly.This is attributed to the relatively longer reaction time of crude oil at lower heating rates, resulting in more thorough reactions.

Determination of Kinetic Parameters
The kinetic parameters of crude oil oxidation include activation energy and the preexponential factor [41].In this study, the activation energy and pre-exponential factor were calculated based on the Flynn-Wall-Ozawa method and the Arrhenius equation by using the TGA experimental results.The activation energy and pre-exponential factor are determined using Equations ( 2) and (3) with characteristic temperatures at the same conversion rate under different heating rates [42,43].The equations are as follows: where E is the activation energy, kJ/mol; R is the general gas constant, and its value is 8.314, J/(mol•K); β is the heating rate, K/min; Tα is the temperature at a specific conversion, K; A is the pre-exponential factor, (kPa•s) −1 ; and P O2 is the partial pressure of oxygen, Pa.
For the TGA experiments on the two different oil samples, a median heating rate of β = 5.0 K/min was determined.The characteristic temperatures for oil sample G were obtained from the TGA curve at the rate of 5.0 K/min as Tα 1 = 765.75K, and for oil sample M, it was Tα 2 = 737.95K. Subsequently, the specified conversion rates corresponding to these temperatures were determined based on the conversion rate and temperature data as α 1 = 0.965 for oil sample G and α 2 = 0.940 for oil sample M.
According to the characteristic conversion rates, Table 6 presents the comparison of oxidation kinetic parameters with equal conversion.We plotted lgβ against 1/Tα.Figure 10 shows the relationships between lgβ and 1/Tα.
Appl.Sci.2024, 14, x FOR PEER REVIEW 14 of 20 We plotted lgβ against 1/Tα.Figure 10 shows the relationships between lgβ and 1/Tα.As shown in Figure 10, the linear fit equation for lgβ against 1/Tα for oil sample G is y = −7.2078x+ 10.108, with a correlation coefficient R 2 = 0.9997.For oil sample M, the linear fit equation is y = −7.0261x+ 10.224, with a correlation coefficient R 2 = 0.9997.By using Equation (2) to calculate the activation energy and Equation (3) to calculate the pre-exponential factor, the results for oil sample G were an activation energy E = 131.2kJ/mol and a pre-exponential factor A = 9.46 × 10 4 (kPa•s) −1 .Similarly, for oil sample M, the activation energy was E = 127.9kJ/mol, and the pre-exponential factor was A = 1.26 × 10 5 (kPa•s) −1 .Compared to the activation energy of 147 kJ/mol for coke combustion, the activation energies for the HTO stage of both oil samples were similar, indicating that high-temperature combustion could continue.This suggests that the oxidation kinetics of the two different heavy oil samples in terms of their viscosity, density, and composition met the requirements for ISC.

Model
To accurately understand the relationship between various parameters and the ISC As shown in Figure 10, the linear fit equation for lgβ against 1/Tα for oil sample G is y = −7.2078x+ 10.108, with a correlation coefficient R 2 = 0.9997.For oil sample M, the linear fit equation is y = −7.0261x+ 10.224, with a correlation coefficient R 2 = 0.9997.By using Equation (2) to calculate the activation energy and Equation (3) to calculate the preexponential factor, the results for oil sample G were an activation energy E = 131.2kJ/mol and a pre-exponential factor A = 9.46 × 10 4 (kPa•s) −1 .Similarly, for oil sample M, the activation energy was E = 127.9kJ/mol, and the pre-exponential factor was A = 1.26 × 10 5 (kPa•s) −1 .Compared to the activation energy of 147 kJ/mol for coke combustion, the activation energies for the HTO stage of both oil samples were similar, indicating that high-temperature combustion could continue.This suggests that the oxidation kinetics of the two different heavy oil samples in terms of their viscosity, density, and composition met the requirements for ISC.

Model
To accurately understand the relationship between various parameters and the ISC effectiveness in practical production, we used the results from the crude oil oxidation experiments as a foundation and combined this with field conditions.We employed the CMG-STARS software, version 2020 to establish an in situ combustion model.In this model, we took into account the oxidation characteristics of the oil samples and incorporated changes in the kinetic parameters for a more in-depth numerical simulation analysis.This approach allowed us to gain a deeper understanding of how these parameters impact the effectiveness of ISC. Figure 11 shows the numerical model grid diagram.
energy was E = 127.9kJ/mol, and the pre-exponential factor was A = 1.26 × 10 5 ( Compared to the activation energy of 147 kJ/mol for coke combustion, the activa ergies for the HTO stage of both oil samples were similar, indicating that high-te ture combustion could continue.This suggests that the oxidation kinetics of the ferent heavy oil samples in terms of their viscosity, density, and composition me quirements for ISC.

Model
To accurately understand the relationship between various parameters and effectiveness in practical production, we used the results from the crude oil oxida periments as a foundation and combined this with field conditions.We emplo CMG-STARS software, version 2020 to establish an in situ combustion model.model, we took into account the oxidation characteristics of the oil samples and i rated changes in the kinetic parameters for a more in-depth numerical simulation a This approach allowed us to gain a deeper understanding of how these parameters the effectiveness of ISC. Figure 11  The model was composed of 2205 effective grids with a size of 21 × 21 × 5, where the grid size in the I and J directions was 10 m with a permeability of 1200 mD, and the grid size in the K direction was 2 m with a permeability of 600 mD.The reservoir depth was set at 1000 m, the temperature was set at 50 • C, the pressure was 10 MPa, the reservoir porosity was 20%, the oil saturation rate was 40%, the reservoir crude oil viscosity was 4440 mPa•s, and the reservoir dip angle was set at 0 • .The maximum development time for each simulation was set to 20 years.Figure 12 shows the relative permeability curves.The model was composed of 2205 effective grids with a size of 21 × 21 × 5, where the grid size in the I and J directions was 10 m with a permeability of 1200 mD, and the grid size in the K direction was 2 m with a permeability of 600 mD.The reservoir depth was set at 1000 m, the temperature was set at 50 °C, the pressure was 10 MPa, the reservoir porosity was 20%, the oil saturation rate was 40%, the reservoir crude oil viscosity was 4440 mPa•s, and the reservoir dip angle was set at 0°.The maximum development time for each simulation was set to 20 years.Figure 12 shows the relative permeability curves.Table 7 lists the key parameters of the numerical simulation model.We established a four-phase eight-component model using numerical simulation software.The eight components included water, heavy oil, light oil, oxygen, carbon dioxide, carbon monoxide, nitrogen, and coke.Table 8 lists the numerical simulation of the reaction kinetic parameters.Table 9 lists the parameters of each component in the numerical simulation.The reservoir-scale numerical simulation of in situ combustion revealed significant differences in the grid-scale simulation when using laboratory-scale combustion reaction kinetic parameters.This discrepancy led to the dispersion of the total heat released by the grid, making it challenging to observe temperature effects.We established a four-phase eight-component model using numerical simulation software.The eight components included water, heavy oil, light oil, oxygen, carbon dioxide, carbon monoxide, nitrogen, and coke.Table 8 lists the numerical simulation of the reaction kinetic parameters.Table 9 lists the parameters of each component in the numerical simulation.The reservoir-scale numerical simulation of in situ combustion revealed significant differences in the grid-scale simulation when using laboratory-scale combustion reaction kinetic parameters.This discrepancy led to the dispersion of the total heat released by the grid, making it challenging to observe temperature effects.
In this study, we employed reaction kinetic parameters from the Xinjiang H block crude oil combined with kinetic parameters obtained from the TGA experiments.These parameters were adjusted based on grid size, involving iterations and modifications of crucial factors such as the activation energy, pre-exponential factor, and gas-liquid equilibrium constants.This process resulted in refined kinetic parameters for the combustion dynamics of heavy oil in the key reaction equations.

The Impact of Activation Energies
Figure 13 shows the oil recovery results under different activation energies.As shown in Figure 13a-c, during the oxidation process, with an increase in the activation energy of light oil burning, the early daily oil production decreased and later gradually increased.As for heavy oil burning and heavy oil cracking, an increase in activation energy led to an initial increase in daily oil production, but it reached a point where no further increase was observed.The changes in the activation energy of these three reactions had a relatively small impact on the final oil recoveries.
According to the results in Figure 13d, the variation in the activation energy of coke oil burning significantly influenced ISC, and the oil recoveries differed.When the activation energy was greater than 59 kJ/mol, an increase in the activation energy of coke oil burning led to an initial increase in daily oil production and ultimately an increase in the recovery factor.However, when the activation energy increased to 179 kJ/mol, a downward trend was observed in the final oil recoveries.According to the results in Figure 13d, the variation in the activation energy of coke oil burning significantly influenced ISC, and the oil recoveries differed.When the activation energy was greater than 59 kJ/mol, an increase in the activation energy of coke oil burning led to an initial increase in daily oil production and ultimately an increase in the recovery factor.However, when the activation energy increased to 179 kJ/mol, a downward trend was observed in the final oil recoveries.This is because this reaction mainly occurred in the HTO stage.An excessively small activation energy clearly did not occur.However, when the activation energy was too high, the combustion zone temperature increased rapidly, accelerating the reaction.This resulted in an insufficient oxygen supply and an incomplete reaction, leading to a decrease in the final oil recoveries.Therefore, in the numerical simulations of heavy oil ISC, the activation energy for coke oil burning had to be in the range of 60 to 150 kJ/mol.The optimal activation energy value was around 130 kJ/mol, which was close to the activation energy for coke combustion, allowing for a better simulation of reactions in the high-temperature oxidation stage.

The Impact of Pre-Exponential Factors
Figure 14 shows the oil recovery results under different pre-exponential factors.As shown in Figure 14a-c, during the oxidation process, with an increase in the pre-exponential factor for light oil burning, the early daily oil production increased but later gradually decreased.In the case of heavy oil burning and coke oil burning, an increase in the preexponential factor led to a decrease in early daily oil production, but this reduction stopped at a certain point.The changes in the pre-exponential factors of these three reactions had a relatively small impact on the final oil recoveries.
According to the results shown in Figure 14c, the variation in the pre-exponential factor for heavy oil cracking had a significant impact on ISC, with different effects on the recovery rates.When the activation energy was greater than 4.17 × 10 9 (kPa•s) −1 , an increase in the pre-exponential factor for heavy oil cracking led to an increase in early daily oil production in simulations.However, as the reaction progressed, a declining trend was observed in the final oil recoveries.This is because, with a constant activation energy, a smaller pre-exponential factor resulted in a lower frequency of molecular collisions during the reaction.Although the early-stage reaction was slower, it allowed for a more complete reaction, leading to a According to the results shown in Figure 14c, the variation in the pre-exponential factor for heavy oil cracking had a significant impact on ISC, with different effects on the recovery rates.When the activation energy was greater than 4.17 × 10 9 (kPa•s) −1 , an increase in the pre-exponential factor for heavy oil cracking led to an increase in early daily oil production in simulations.However, as the reaction progressed, a declining trend was observed in the final oil recoveries.This is because, with a constant activation energy, a smaller pre-exponential factor resulted in a lower frequency of molecular collisions during the reaction.Although the early-stage reaction was slower, it allowed for a more complete reaction, leading to a higher final oil recovery.Conversely, an excessively small pre-exponential factor could cause insufficient fuel supply in the HTO phase, resulting in slow oil recovery.On the other hand, when the pre-exponential factor was large, the numerical simulation process accelerated the conversion of reservoir crude oil into coke, burning a significant amount of crude oil and ultimately reducing the oil recoveries.Therefore, when performing numerical simulations for heavy oil ISC, the preexponential factor for heavy oil cracking should not be excessively large or small.The optimal pre-exponential factor value was around 4.17 × 10 9 (kPa•s) −1 .This value ensured a thorough reaction without burning too much crude oil, providing sufficient fuel for the HTO stage.

Conclusions
(1) The RTO and PDSC experiments indicated that, with an increase in pressure, the contact area between oxygen and crude oil expanded, intensifying and accelerating the oxidation process of crude oil.The oxidation process of the two heavy oil samples was divided into three stages, LTO, FD, and HTO, with descriptions of the reactions occurring in each stage.(2) The TGA experiments revealed that, with an increase in the heating rate, the overall oxidation rate of oil sands increased.The peak regions of both LTO and HTO reac-tions widened, and the corresponding temperatures shifted to the right.Using the Flynn-Wall-Ozawda method, the activation energy (E) and pre-exponential factor (A) for oil sample G were determined as E = 131.2kJ/mol and A = 9.46 × 10 4 (kPa•s) −1 , while for oil sample M, E = 127.9kJ/mol and A = 1.26 × 10 5 (kPa•s) −1 .(3) The numerical simulation revealed that changes in the activation energy and preexponential factor generally had opposite effects on the crude oil oxidation process.For light oil burning, an increase in activation energy led to a decrease in oil recoveries, while an increase in the pre-exponential factor resulted in higher recovery rates.In the case of heavy oil burning, heavy oil cracking, and coke oil burning, an increase in activation energy led to higher recovery rates, while an increase in the pre-exponential factor resulted in lower oil recoveries.(4) The impact of activation energy on ISC recovery rates was primarily influenced by changes in activation energy in coke oil burning, with a recommended value of 130 kJ/mol.The influence of the pre-exponential factor on ISC recovery rates was mainly attributed to changes in the pre-exponential factor in heavy oil cracking, with a recommended value of 4.17 × 10 9 (kPa•s) −1 .Altering the activation energy and pre-exponential factor for light oil burning and heavy oil burning had a relatively minor effect on the final oil recoveries.

Figure 1 .
Figure 1.The schematic diagram of the RTO system.Figure 1.The schematic diagram of the RTO system.

Figure 1 .
Figure 1.The schematic diagram of the RTO system.Figure 1.The schematic diagram of the RTO system.

Figure 1 .
Figure 1.The schematic diagram of the RTO system.

Figure 4 .
Figure 4.The schematic diagram of zero-point TGA.

Figure 4 .
Figure 4.The schematic diagram of zero-point TGA.

Figure 6 .
Figure 6.Oxidation stages division of air composition variation at 3.0 MPa with different oil samples: (a) oil sample G; (b) oil sample M.

Figure 6 .
Figure 6.Oxidation stages division of air composition variation at 3.0 MPa with different oil samples: (a) oil sample G; (b) oil sample M.

Figure 7
Figure7shows the PDSC results of the heat-flow experiment.As shown in Figure7, before 200 • C, there was no significant exothermic change in either oil sample.With the increase in temperature, the crude oil sample underwent an evaporation phase change.When the reaction temperature was below 200 • C, the evaporation phase change exceeded lowtemperature oxidation, resulting in a constant zero heat-flow curve without any changes.As the temperature increased between 200 • C and 300 • C, the samples began to show changes in heat flow, marked by the appearance of the first heat-flow peak.In this temperature range, the exothermic heat released during the low-temperature oxidation of crude oil exceeded the endothermic heat associated with the evaporation process.When the temperature reached between 400 • C and 500 • C, a second heat-flow peak appeared.It was lower compared to the first peak, with a more pronounced decrease, especially in the M oil sample.

Figure 6 .
Figure 6.Oxidation stages division of air composition variation at 3.0 MPa with different oil samples: (a) oil sample G; (b) oil sample M.

Figure 7 .
Figure 7.The PDSC results of the heat-flow experiment with different pressures: (a) oil sample G; (b) oil sample M.

Figure 7 .
Figure 7.The PDSC results of the heat-flow experiment with different pressures: (a) oil sample G; (b) oil sample M.

Figure 8 .
Figure 8.The weight loss rates with different heating rates: (a) oil sample G; (b) oil sample M.

Figure 8 .Figure 9 .
Figure 8.The weight loss rates with different heating rates: (a) oil sample G; (b) oil sample M. Appl.Sci.2024, 14, x FOR PEER REVIEW 13 of 20

Figure 9 .
Figure 9.The conversion rates with different heating rates: (a) oil sample G; (b) oil sample M.

Figure 11 . 11 .
Figure 11.The numerical model grid diagram: (a) 3D grid diagram; (b) well patterns for IS Figure 11.The numerical model grid diagram: (a) 3D grid diagram; (b) well patterns for ISC.

Figure 12 .
Figure 12.The relative permeability curves: (a) relative permeability of oil-water phase; (b) relative permeability of gas-liquid phase.

Figure 12 .
Figure 12.The relative permeability curves: (a) relative permeability of oil-water phase; (b) relative permeability of gas-liquid phase.

Figure 13 .
Figure 13.Oil recovery results under different activation energies: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.Figure 13.Oil recovery results under different activation energies: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.

Figure 13 .
Figure 13.Oil recovery results under different activation energies: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.Figure 13.Oil recovery results under different activation energies: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.

Figure 14 .
Figure 14.Oil recovery results under different pre-exponential factors: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.

Figure 14 .
Figure 14.Oil recovery results under different pre-exponential factors: (a) light oil burning; (b) heavy oil burning; (c) heavy oil cracking; (d) coke oil burning.

Table 1 .
The reservoir conditions.

Table 2 .
The SARA information.

Table 3 .
The RTO experiment conditions.

Table 3 .
The RTO experiment conditions.

Table 4 .
The PDSC experiment conditions.Start the DSC power supply.After the instrument is fully started, open the pressure chamber cover, place the sample crucible in the reaction chamber, and cover it successively.

Table 5 .
The TGA experiment conditions.

Table 6 .
Comparison of oxidation kinetic parameters with equal conversion.

Table 6 .
Comparison of oxidation kinetic parameters with equal conversion.

Table 7
lists the key parameters of the numerical simulation model.

Table 7 .
Key parameters of the numerical simulation model.

Table 8 .
The numerical simulation of reaction kinetic parameters.

Table 7 .
Key parameters of the numerical simulation model.

Table 8 .
The numerical simulation of reaction kinetic parameters.

Table 9 .
The parameters of each component in numerical simulation.