Real-Time Monitoring and Performance Optimization of Steam Injection in Heavy Oil Reservoirs Using Fiber Optic Sensing and Integrated Predictive Simulation Models

Introduction

After primary and secondary recovery, enhanced oil recovery (EOR) is usually the last stage of oil production. While secondary recovery uses water injection to sustain pressure when natural energy diminishes, primary recovery depends on the reservoir’s natural pressure or artificial lift to extract oil. Tertiary recovery, or EOR, is used when these techniques become economically inefficient, frequently as a result of increased water output. EOR increases oil output by extracting more oil using thermal, chemical, or other means. EOR can be utilized at any point when conventional techniques are inadequate, even though it is often utilized near the end of a field’s life [1]. The two main procedures in thermal EOR are steam stimulation (cyclic steam injection) and steam-flooding. Steam and hot-water flooding are the predominant approaches. These techniques improve flow to producing wells by decreasing the viscosity of heavy oils. The most efficient EOR technology is steam-flooding, which recovers around 410 million barrels of oil per day worldwide. In-situ combustion and steam stimulation recover 57 million barrels per day, while all other EOR techniques recover 220 million barrels per day [2]. Heavy oil production frequently uses thermal EOR, which includes steam injection methods including In-Situ Combustion (ISC), Cyclic Steam Stimulation (CSS), and Steam Assisted Gravity Drainage (SAGD). The geographic resolution of traditional monitoring systems is frequently restricted. Distributed, real-time temperature and acoustic data collecting along the whole wellbore has been made possible by the advent of Fiber Optic Sensing (FOS), namely Distributed Temperature Sensing (DTS) and Distributed Acoustic Sensing (DAS) [3, 4].

The fundamental idea of a FO system is based on a transmitted light signal in which the FO cable (measured light) acts as a carrier for light, sending out a light signal from the transmitter on the surface. Reflected light is produced by light pulses and has an extremely low attenuation index. In order to reach the receiver at the surface, the backscattered light waves must return via the optic medium. The data is decoded, or converted to temperature, pressure, acoustic, or strain data, and then recorded in a database based on the technology used. This process compares the backscattered light signal with the light pulse that was first fired from the transmitter (reference beam). FOS Techniques are widely applicable in thermal EOR. There are two technologies for FOS operations. DTS Technique is the initial technique. Raman scattering in optical fibers is used by the DTS technology to monitor temperature profiles [3]. In SAGD operations, DTS has been very helpful in tracking the growth of the steam chamber and optimizing the SOR [5]. Field research conducted in Alberta showed how DTS was able to record the steam chamber’s temporal development, allowing for modifications in steam injection to increase the effectiveness of oil recovery [5]. The second is DAS, which uses Rayleigh backscattering to create continuous acoustic sensors out of optical fibers [6]. By offering dynamic flow monitoring, DAS enhances DTS. DAS applications for detecting steam flow patterns during injection in thermal EOR processes were demonstrated by Mateeva, et al. [6]. A multi-physics sensing platform is provided by hybrid DTS/ DAS systems, which combine DTS and DAS [7]. According to recent advancements, integrated fiber systems can increase reservoir interpretation and control by providing temperature, acoustic, and strain data [7, 8].

FOS Applications in Thermal EOR Methods are several. DTS makes it possible to track the compliance of the steam chamber in SAGD wells in real time, which helps to prevent steam channeling and optimize production techniques [9].

Operators may dynamically modify steam injection profiles thanks to continuous feedback from DTS, which lowers heat losses and enhances SOR [9]. DTS has also been used in CSS projects to track the steam soak phase and evaluate production-related thermal behavior [10]. This type of monitoring optimizes following cycles and enables the identification of thermal inefficiencies [10]. Nonetheless, DTS detects sharp temperature gradients that correlate to combustion zones in ISC operations in order to follow thermal fronts [11]. According to field findings, DTS assists operators in tracking the combustion process and modifying injection plans as necessary [11].

Challenges in FOS Deployment for Thermal EOR represent a key-element during applications. Despite its benefits, FOS has certain drawbacks, most notably fiber deterioration at temperatures above 300°C [12]. Furthermore, logistical and financial challenges arise due to the intricacy of installation in horizontal or deviated wells [13]. Moreover, sophisticated machine learning algorithms are needed to analyze the enormous volumes of data produced by DTS/DAS [14]. Using digital twins and machine learning in conjunction with FOS data to automate decision-making is one of the emerging trends [15]. In order to improve sensor survival for prolonged thermal EOR operations, research on high-temperature fiber coatings is also ongoing [16]. Furthermore, the development of 4D monitoring systems for subsurface operations appears to be promising when DTS, DAS, and DSS are applied together [17].

For better understanding the field monitoring and evaluation philosophy, it is advisable to know the best way to measure the changed parameters, how to measure them, and their effect on all field studies including wells and reservoir. By comprehending the observed variables and suggesting remedial action in real time, reservoir management seeks to maximize process efficiency in the reservoir. One crucial component of production facilities that might endanger the safe growth of a project is the detection of out-of-range operating situations. During injection and production cycles, continuous real-time temperature and pressure monitoring for heavy and viscous oil well production optimization aids in the optimization and improvement of CSS scenarios, provides information about heat transfer by convection and conduction into the reservoir, records any breakthroughs in steam or water, records any evidence of combustion gas breakthroughs, enhances comprehension of the processes occurring inside the borehole (downhole), and provide information for reservoir modeling and other optimization strategies.

Therefore, the main objective of this paper is to investigate temperature-pressure distribution using FOS in Suplac field wells during injecting a steam into a heavy crude oil reservoir. Two horizontal drilled wells are used in injecting steam into the reservoir. Two complete injection steam cycles are done for each, 130 and 150 days for well # 1 and 2 respectively. A comparison study between the two cycles is done for both well including pressure, temperature, steam quality, production parameters over the specified period. In addition, a model is developed using Prosper software to predict the fluid well, tubing and annuli temperature and pressure, and steam quality for steam injections with certain parameters. Thus, it is important to review the FOS applications, and EOR improvements. The Suplac field data are also described. Wells # 1 &2 data are described as well. Results and discussions are made.

A Chronological Review of Fiber Optic Sensing in Oil and Gas Fields

Thermal EOR monitoring capabilities have been significantly enhanced by FOS, especially DTS and DAS [3, 4]. FOS technologies are essential to creating intelligent, effective, and sustainable thermal EOR systems because of developments in fiber durability and AI-driven analytics [18]. In order to continue researching and apply FOS methods efficiently, a sequential review is made to study the historical studies, approaches, applications techniques of FOS especially in oil and gas field as shown in Table 1.

Study Methodology and Research Procedures

The goal of the project is to combine coupled wellbore– reservoir simulation with real-time distributed FOS in order to measure and optimize SAGD performance in the Suplacu heavy oil field. Early detection of out-of-range operating situations, optimization of the steam injection method, and enhancement of recovery efficiency while maintaining well integrity are the objectives.

Modelling & Analysis Workflow

• Record synchronous P-T-Q data from the wellhead. 3. Quality Control
• Perform automated detection of sensor drift and dropouts > 0.2 bar or > 3 °C; initiates field check within 24 hours.
• Prevent the previously mentioned data gaps by maintaining a backup power source.

Field Results and Discussions: Real Time Measurements, Simulations and Comparisons

Real-time surveillance that identifies departures from safe operating limits and permits timely remedial action is essential to effective reservoir management. This study uses distributed fiber-optic sensing (FOS) to evaluate the effectiveness of SAGD in Suplacu Field with a heavy oil reservoir of 8–13 °API, and 500–1000 cp viscosity. Wells #1 and #2, two horizontal injectors, were observed for two full steam-injection cycles. Table 3 shows the rock and fluid properties of Suplacu reservoir that were used during modeling with prosper. Figures 1 through 6 show the field map, well profile and data used in this study.

The temperature distribution in a well over time – almost one year– and depth is depicted in the 3D graph shown in Figure 7, most likely during a steam injection or other thermal recovery procedure. The z-axis displays temperature oC, colour-coded in bands (from 90–100°C up to 190–200°C), the y-axis displays depth at different sensor sites (varying from ~215 m to ~610 m), and the x-axis depicts time (from mid-November to mid-September). The temperature zones’ contours and layers indicate changes in heat penetration at various depths and periods as well as the movement of the thermal front. Temperature profile spikes and plateaus signify active steam injection times, well shut-in times, or injection technique adjustments. For tracking thermal efficiency and seeing heat breakout zones or steam channeling over time, this visualization is helpful. FO-based temperature and pressure monitoring in well #1 throughout a year of steam-assisted gravity drainage (SAGD) operations is shown in Figure 8. Sharp rises in wellhead injection pressure (up to about 13.5 bar) and accompanying temperature spikes reaching 220°C, with ΔT values of 56°C and 30°C, emphasize two main steam injection cycles. Effective steam propagation is shown by steady baseline conditions and distinct heat and pressure responses during injection, as shown by FO data. Additionally, the graphic shows a short loss of data due to an RMS (Remote Monitoring System) breakdown that occurred over a period of 1.5 years. All things considered, the figure illustrates how FOS makes it possible for ongoing downhole monitoring for real-time diagnostics, well integrity evaluation, and thermal EOR optimization.

Click to enlarge

Click to enlarge

Figures 9 through 13 show how temperature and pressure changed throughout the course of 130 operating days along a slotted liner, signifying two production cycles. The five graphs, which are divided into different stages— after 15, 30, 60, 91, and 130 days—show the temperature and pressure behavior along a slotted liner in a wellbore during the period of 130 days of production. Temperature (°C) and pressure (bar) are shown against the slotted liner’s measured depth (MD) in each graph. Interestingly, two manufacturing cycles are recorded; the first is often indicated by red text, while the second is shown by purple. Water cut (WC) and liquid production rate (Qliq) are monitored for each cycle as production time goes on, and cumulative liquid production is shown to measure reservoir performance. Along with the injection tubing termination sites for each cycle, the graphs also indicate important completions and formations like “SAND” and “CARBONATE”. The temperature profile gradually decreases over the course of the five time periods, particularly in the vicinity of the well’s toe (in the direction of greater MD), suggesting thermal front movement brought on by production drawdown and cooling by produced fluids. Higher Qliq and WC values are first seen, particularly in the second cycle. But with time, WC improves as Qliq declines, indicating better oil cut as water output declines. By day 130, WC is at 55% and Qliq for both cycles is between 14 and 17 m³/d. The pressure seems to stabilize at 5 bar or less, and the temperature becomes more consistent. As is common in thermal recovery operations, the data show front stabilization, production drop, and reservoir cooling. A mature thermal sweep and decreased thermal losses are implied by the improvement in WC and comparatively stable temperatures at later stages, which suggest a manufacturing regime that is steadily stabilizing.

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

The next graphs (Figures 14&15) show the temperature and pressure profiles along a wellbore during steam injection as observed by FO and modeled (Prosper). The y-axes display pressure (right) and temperature/steam quality (left), while the x-axis shows the well length (in measured depth). In both pictures, temperature, steam quality, and pressure are shown versus measured depth, comparing calculated Prosper outputs with measured FO data for Well #1 during steam injection. The graphs provide a number of temperature profiles, such as FO, Prosper fluid, annulus, and casing temperatures, in addition to Prosper-calculated fluid pressure and steam quality. Important well components, such as the liner hanger, casing shoe, and injection tubing string end, are indicated by comments. One noteworthy finding in both figures is that, as the FO data suggests, Prosper model overestimates temperatures above the reservoir. The results also demonstrate the steam entering the reservoir at about 220 m MD and indicate constant injection circumstances (90 tons of steam per day, WHP = 12.84 pressure, THP = 220°C). In model 1, it is possible that the reservoir temperature is warmer than initially because of gases from combustion + previous steam cycles. In this model, the pressure is reasonably matched, but the temperature is off. Also, it was not possible to derive a model with both matched (Figure 14). In the second model (Figure 15), the temperature is matched, but the pressure is not matched. Also, Prosper will predict saturation conditions, which was not what was recorded. In both unmatched models, the calculated steam quality is ~80%.

Click to enlarge

Click to enlarge

During injection through well #1 (Figures 8 through 14), the temperature is uniform along the slotted liner. It appears that the well shoe remains warmer than the rest of the well section, even when the well is shut down. This corresponds also to the Panonien sand section, after the long section of carbonate. Furthermore, the steam is not injected uniformly in the reservoir, but rather to zones with least resistivity (good rock properties & good fluid mobility: either warm zone, or zone with gas), or/and closer to the end of the injection string. With increase of gas production, the temperature profile does not change and the gas could also come from this most productive area. With high water cut, the temperatures in the middle well section are smaller than with lower water cut. It is not possible to see where the water corresponding to the increasing water cut is coming from – most likely also from the most productive zone, i.e. the well toe. During steam injection, the measured temperature and pressure do not correspond to the saturation conditions – big concerns that only hot water is injected. To get to the saturations conditions, the pressure should be ~2bar lower, could it be measurement error (~22% error)? Could the pressure gauge be disrupted by high temperatures? Prosper model could not reproduce those measurements. For the given injection conditions it would predict a ~80% steam quality. During steam injection in well #1, the temperatures are uniform in the slotted liner. It seems that more steam goes into reservoir sections with least resistivity and close to the end of the injection tubing string. The following points are extracted based on the results:

• The location of the end of the injection tubing string has an effect
• Preliminary Prosper models would predict: steam quality ~80% - 60% to tubing shoe for injection rate 90- 50 t/d
• Differences between measured wellhead temperature in time of injection phase and FO sensor temperature are around 40 -50°C
• Differences between measured wellhead pressure and FO sensor pressure are around 3-4 bar in CSS time
• Measurements in well #1 during steam injection did not correspond to saturation conditions, which should indicate hot water injection only
• This is the reason why Prosper model needs improvements regarding the pressure and temperature measurement error margin or further investigations over reservoir parameters
• It’s hard to qualitatively assess what is happening toward the well heel. The measured temperature may reflect more the well flow from the other well section, than the inflow from the reservoir; 1. Other pressure sensors are required apart from toe (i.e. to heel ); 2. Need of the permanent electrical power source to avoid wrong readings after electrical breaks Another steam injected well called well # 2 is also studied as shown from Figures 16 through 23. The temperature profile after 16 days of production for the 2 cycles is different (Figure 16). This could indicate that the location of the injection tubing string had an effect. It seems that injecting where mud losses where encountered helped cleaning this part of the well, as the temperatures in this zone are higher than in the previous cycle. After 29 days of production (Figure 17), both cycles showed a hot producing zone at ~460m MD. For the 2nd cycle the cold zone at 480-490m MD can be seen, but then the well is relatively warmer toward the toe, which could confirm its contribution – except for the last sensor that records smaller temperature. However, the difference in the temperature profile of the 2 cycles toward the well toe is confirmed over the next five periods: 39, 60, 90, 120, and 150 days (Figures 18 through 22). Moreover, another model was built using Prosper software to predict well temperatures and pressures along the wellbore profile, then compare with the FO measurements for well # 2. Figure 23 show a slightly match between Prosper model and FO data measured. However, there are still lots of things to sort in Prosper, but for this rate, Petroleum Expert 5 correlation gives a reasonable result. The calculated steam quality is ~68%. The problem with Prosper is to match the range of rates with the same correlation. This could not be achieved. During injection in well # 2, the temperatures are uniform along the well during steam injection. However having the steam in the slotted liner during injection, does not guaranty that the steam is injected uniformly along the well section. Changing the end of the injection tubing string location in the well had an effect in the temperature profile. It seems that for the 2nd cycle, the zone that had severe mud losses during drilling has been cleaned and has a better contribution. There is a zone at ~460m MD that remains warmer than the rest of the well – it corresponds to the end of the tubing string of the Aug 14 steam injection – but it can also be observed for the 2nd cycle. Either this zone is hot because it received directly a lot of steam in Aug14, i.e. this is due to the injection tubing string location Also, it has particularly good properties as shown in the well log, so it is easier for the steam to get into this zone. With time a colder zone can be observed from 480 to 490m MD. This is most likely aquifer water coning. Another one could be briefly observed at 430m MD, but seemed later masked either by fluid coming from the other well sections or because its contribution is small. The 1st well section [380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 422, 423, 424, 425, 426, 427, 428, 429, 430] mMD where logs interpretation attributes good porosity does not seem to contribute much to production. A Prosper model would predict a 68% downhole steam quality during the February 15 steam injection.

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Click to enlarge

Conclusions

In conclusion, the Suplacu field is served as a prototype for other fields of a similar kind in Romania and is a significant testing ground for cutting-edge heavy oil exploitation technology. The extraction of heavy crude oil at advanced maturity circumstances, when thermal processes and specialized pumps are necessary to sustain economic production, is shown by this reservoir. In heavy- oil fields like Suplacu, this study has demonstrated that real- time distributed fiber-optic sensing (FOS), which includes distributed temperature sensing (DTS) and distributed acoustic sensing (DAS), is a useful and effective method for improving SAGD performance monitoring. In order to accurately visualize the evolution of the steam front and identify injection irregularities, the field pilot showed that FOS systems can consistently record high-resolution thermal and acoustic data over the whole length of horizontal injector wells. Because it provides geographically and temporally rich information, this continuous monitoring technique performs noticeably better than conventional point-based sensors. More precise history matching and a better comprehension of reservoir behavior, including the identification of heterogeneities and steam channeling, were made possible by the integration of FOS data with a coupled wellbore- reservoir simulation model. The efficacy of the existing data infrastructure, which was shown to be reliable enough for prolonged monitoring periods, was also confirmed by the pilot. Despite early problems with sensor power failures and data synchronization, they were resolved later in the trial. The technical viability of applying FOS in established, onshore heavy-oil operations was generally validated by the field study. The system was useful not just for real-time monitoring but also for optimizing well performance during the SAGD cycle and guiding steam injection schemes.

Recommendations

Based on the results and discussions, the recommendations are divided into the main six items as following:

• Wider Deployment of FOS Systems: To provide complete temperature and flow profiling, assist cross-well diagnostics, and enhance compliance control, DTS and DAS systems should be installed across injector and producer wells.
• Improved Instrumentation: To further enhance coupled model calibration and improve pressure diagnostics, install pressure gauges at the wells’ toe and heel portions.
• Workflows for Automated Optimization: Use automated simulation-feedback loops with real-time FOS data to continually modify steam injection rates and identify operational irregularities early.
• Advanced Data Analytics: Use machine learning techniques to analyze and analyze vast amounts of auditory and thermal data, allowing for automatic anomaly identification (such as coning, steam breakthrough, or crossflow).
• Scalability Assessment: To evaluate the wider applicability and scalability of this integrated monitoring and modeling process, conduct further pilots in other Suplacu field regions and in heavy-oil reservoirs with comparable geology.
• Increase Power and Data Redundancy: Reduce the possibility of data gaps during crucial injection or production stages by increasing system power redundancy and communication dependability. This will guarantee continuous data collecting. Operators may considerably enhance reservoir management, lessen their impact on the environment by using steam more efficiently, and increase the productive life of heavy-oil assets by integrating high-resolution sensor technologies with dynamic simulation and optimization.