Adrian Buenrostro, Mohammed Al-Abdrabalnabi, Amro E. Mukhles, and Saad M. Al-Driwees

Reprinted with permission Journal of Technology Saudi Aramco

Fracturing is a complex technique utilized to stimulate hydro-carbon reservoirs to enhance their production. The theory, tools, and techniques utilized for fracturing are complex and very specific. With the demanding increase of hydrocarbon production, proppant and acid fracturing are widely implemented, and in some cases, is the unique solution to having commercial production, i.e., unconventional gas reservoirs.

The optimization of the fracturing process is important since this activity is complex, expensive, and in most of the cases, required for commercial production. Several parts of the process are known based on previous experiences. In other cases, complex studies still need to be done to better understand each case, to enable a more accurate propped fracturing execution. To be confident in the case of design, execution and evaluation of a proppant fracturing job, fracture geometry determination is very important, as this drives the adequate evaluation and job execution while also allowing for job optimization. One of the basic and most important values to estimate the fracture geometry is the fracture height; once this parameter is determined, fracture width and fracture length can be determined with less uncertainty, based on other parameters from log data, job execution, etc. The fracture geometry can be estimated more confidently if the fracture height value is set.

This article studied and compared multiple scenarios of sandstone reservoirs. The study was made for fracturing jobs focused on gas and hydrocarbon production. Tight conditions and unconsolidated sandstones were dominant scenarios of the fracture jobs. The parameters utilized for this study were focused on some log signatures, which were observed to drive the criteria of the fracture height growth. Based on those parameters, the project attempts to simplify the process of fracture height determination. The exercise was made for a group of wells in different areas of the reservoirs where some parameters prevail, and we are confident in the gathered information. The signatures observed proved helpful to confidently predict the expected fracture height. In the exercise of this project, all the examples were made with cases where the temperature log was made after reading the fracture data of the obtained frac-ture height by the temperature log criteria 1.

 

ANTECEDENTS

This project was made in wells that produced outstanding gas rates from its reservoirs over a period of decades. Stimulation by proppant fracturing at the sandstone reservoirs become a regular requirement to find commercial production after new areas were attempted for production. The process of proppant fracturing started with multiple uncertainties from the reservoir point of view, and also for the logistic part at the job execution as fracturing was not previously used. The regular fracturing startup process involves reservoir log-ging evaluation, injection test, completion hardware setup, soft-ware simulations, and lab tests for compatibility, and forecast studies for production results, among other challenges. Specific studies and tests were conducted in preparation to the first field applications. During the startup process, it was observed that a wide range of parameters need to be clarified prior to deploying the stimulation in the field.

Figure 1 shows the parameters used for a fracture geometry in a generic way, simplified to have the total length, height, and width, represent a volume. In real cases, a fracture is not a geo-metrically perfect square or rectangle; but to simplify equations and analysis, computational models assume simplified shapes of fracture geometry for processing calculations. In general, for a fracture geometry determination for the three main dimensions; the length (L in Fig. 1, or “d” in Table 1), width (W), and height (H), are based on:

  • The amount of energy placed (pressure while pumping and pumping rate).
  • The amount of mass used (fluids + proppant).
  • The amount of energy dissipated (fluid efficiency and pressure variations). Fluid efficiency is related to the fluid amount pumped, the amount lost in fracture geometry to the formation due to leak-off, and the fluid effectively creating a fracture geometry.
Fig. 2. Fracture geometry development from the wellbore related to the formation stress orientation, which can be either vertical or horizontal. Fig. 2. Fracture geometry development from the wellbore related to the formation stress orientation, which can be either vertical or horizontal.

Note that in Fig. 1, the length is only for one fracture wing, as the fracture is assumed to develop symmetrically — the same geometry on both sides from the wellbore to the formation. The fracture volume and geometry is standard, referring only to one wing or side of the fracture.Width, according to fracture geometry calculation, is based in different models of the fracture, and is directly related to2 : EE′=E1−ν2 (1)where H = frac height, L = frac length, R = radial growth, μ = viscosity of fluid, Q = pumping rate, E’ = plain strain modulus, E = Young’s modulus, and ν = Poison’s ratio.The fracture width can be calculated in general by:ww ∝ [QQQQ LLEE′]1/4(2)Following Fig. 1 and Table 1, and Eqns. 1 and 2, it is observed that the width is driven by Q, μ, L, and formation mechanical parameters. This concept is a simplified way to calculate the fracture “d”W (width)Perkins and KernH(μQL/E’)1/4RadialL(μQR/ E’)1/4Gertsma deKlerkR (radius)(μQL2/E’H)1/4 “d” W (width

  “d” W (width)
Perkins and Kern H (μQL/E’)1/4
Radial L (μQR/E’)1/4
Gertsma deKler R (radius) (μQL2/E’H)1/4

Table 1. Fracture Geometry Model’s to calculate the main dimensions

in correlation to a general fracture model, whereas other parameters, i.e., pumping pressure, net developed pressure, and leakoff, among others, may affect the fracture’s width. As previously mentioned, the three main dimensions, length, width, and height, are simplified for the analysis made to calculate them to obtain the fracture geometry. Since there are three main variables for one model, the definition of any of those three values simplifies the calculation of the other two dimensions. Therefore, when there is a chance to measure any of those parameters, the fracture geometry can be determined more confidently. It is possible for the fracture length and width to drastically vary, depending on the fracture height determination; whereas once one value is determined, fracture optimization is feasible. Additional technologies can be used to measure the width and fracture length. Those technologies are usually complex and present multiple concerns; as some can have limitations and logistic issues due to their particular requirements, which means that the implementation of them is then not always pos-sible. Fracture height in another way is easy to measure when the fracture plane is collinear to the wellbore, which commonly is the case of vertical wells, where most of the geomechanical conditions of the reservoirs place the wellbore collinear with the fracture plane, Fig. 2. The natural arrangement of the main stress orientations in the reservoirs make a vertical wellbore suitable for developing a vertical fracture plan, thereby making the fracture height collinear with the wellbore3 . The characteristic of a vertical well, being collinear with the minimum in situ stress of the rock, allows the fracture created to develop its height attached to the wellbore from the top to bottom height of the fracture.

An important temperature contrast is generated from the fluid injected in the formation at the depth where the fracture is placed at perforations made on the wellbore. The reservoir is usually dozens of degrees hotter than the injected fluids. The change in temperature is made by the fluids pumped into the formation, which starts cooling down the well from the surface, the wellbore completion, and finally reaching the reservoir. Equation 3 describes the parameters used to analyze the temperature change according to thermodynamics by three ways: conduction, convection, and radiation, Fig. 3. Conduction is the most important for the case of heat exchange in the formation between the rock, completion tubulars, and fluid injected during a fracture job.QQ/ tt∝kkkk(TTHHHHHH− TTCCHHCCCC); (3) where Q = energy (W), t = time(s), k = heat transfer coefficient [W/((m2) (°C))], A = heat transfer area (m2), THot = temperature of the hot mass (°C), and TCold = temperature of the cold mass (°C).

FIELD CASE APPLICATION

For the case of a fracture geometry investigation, fluid is pumped from the surface (bull heading) through the wellbore completion to the formation. Fluid is pumped at surface tem-perature, around 20 °C to 25 °C. The formation rock where the fracture is being made is several degrees hotter, and dissi-pation of the temperature will be driven by reservoir conditions depth, reservoir fluids, and rock type, among others. A reservoir at around 4,000 m deep will have approximate tempera-tures of 150 °Cto 180 °C. The difference from the reservoir temperature with the injected fluid generates an important cool-down at the area of injection where the fracture is generated. This temperature contrast dissipates with time, and depends on formation properties to accelerate or restrict the temperature to cool down or heat up. Usually, after several hours 6 to 12 hours a temperature contrast can be measured from the wellbore to the formation. During the process of a proppant fracture job for the study, injected cold fracturing fluid is usually no less than 50 m3 of base water fluid, injected into the rock, which is a porous media filled with gas, oil and/or water. The process is made in as high a rate as possible to fracture the formation under con-trolled parameters. The pumping rate used is from 3 m3/min to 7 m3/min, therefore, the fluid is injected in 10 to 15 minutes.

Fig. 3. General heat transfer paths by conduction, convection, and radiation.

Fig. 3. General heat transfer paths by conduction, convection, and radiation.

Quick injection allows for a higher temperature contrast  at the core of the injection point, which decreases with distance from that point, until a distance where original reservoir conditions (temperature) are not disturbed because no fluid reaches that point. Figure 4 shows the reservoir depth, when the fracture plane is collinear with the wellbore, it can be interpreted based on the temperature profile contrast observed on the readings of a log made with high-resolution readings, i.e., a high-resolution temperature (HRT) log. Figure 4 represents a logging job, where tools are placed into the wellbore by running in hole (RIH), and recording the temperature changes.

As observed in the left log plotted lines, there are two continuous lines, thick and thin each one close to a dashed line the dashed line in each case is straight, and used as a reference to the perfect incremental profile of the temperature for this log case. The continuous lines are the real temperature readings by the log in the well. They are not straight different layers on the reservoir commonly have small variations in temperature. Most of the time, temperature logs are repeated over periods of time; for particular objectives where changes on the temperature profile is expected to provide information to make decisions on well conditions. To have a case by case reference about temperature profiles, the performance for each well was studied at initial temperature conditions. For this, an initial log was made when the well was undisturbed, at the startup of the well life; the initial log was then the base reference for comparison, so further temperature changes could be identified.

 

Fig. 4. Logging a well for temperature log.

Fig. 4. Logging a well for temperature log.

Due to the disturbance of wellbore fluids, and the thermal  hysteresis of the temperature log for the bottom-hole assembly, temperature readings differ if made RIH or pulling out of hole.For a proppant fracture job, the temperature  log when the well was undisturbed, is compared with the temperature log made after injection of the fluids, either for fracture evaluation or after the main fracture job, to document the final fracture height. This information was obtained by selecting the points where the temperature profile differs from the normal temperature base line. The temperature increases as the depth increases, resulting in a deviation of temperature profile, where a cool-down is observed close to the injection point at perforations which suggests that the pumped fluid has invaded the reservoir at the perforation zone. The readings must be done above and below the disturbed zone of injection, where the temperature deviates from a normal trend above and below the perforations, suggesting fluid invasion related to height

Fig. 5. Example of temperature drop at the injection zone; the temperature falls from expected undisturbed profile due to fluid recently injected.

Fig. 5. Example of temperature drop at the injection zone; the temperature falls from expected undisturbed profile due to fluid recently injected.

Above the perforations, the temperature has a similar slope to the original temperature log, but below the perforations where the fluid was not disturbed, only the  area of fluid invasion shows a cool-down. Deeper points show undisturbed temperatures from the initial conditions.
To verify the validation on readings as well as dissipation of a cool-downat the injection section, subsequent logs in specific time intervalsare made. For this study, the standard time is 2 hours between readings. The first reading should be done as soon as possible afterinjection. Due to logistics
and the surface operation sequence of activities, usually the first temperature log is made from 4 to 6 hours after injection, Fig. 6.

Therefore, the energy of the fluid, which can influence a change in the temperature of the rock, is less than when it is exposed to a smaller rock section. This criteria is utilized for the multiple (usually a total of three) temperature logs made to determine if with time, therock tends to dissipate the fluid in the fracture created under a non-homogeneous profile all lines on the temperature log becoming parallel to each other at any single depth. This case can be interpreted as the presence of a “dominant” fracture zone, which shall be considered for the fracture design and execution as the risk of fracture propagation could not be constant during the execution, increasing the chances of issues to complete the job as desired. Note that trend variations among subsequent logs could indicate differential closure conditions at various reservoir layers related to the fracture. The interpretation of such cases is important, but is a particular topic to be covered by specific study, which is not the case for this article.

 

After several years and dozens of temperature logs made,the temperature logs showed mostly average behaviors on fracture geometry “height” description. So far, no more than 85m (~275 ft) of fracture height has being identified, while also never finding a developed fracture of less than 21 m (70 ft) in height. The average fracture height in general is found to be around 33 m (120 ft), and it was observed that certain parameters on the log are usually related to the fracture height limits, suggesting those values are affecting or directly impacting the fracture height development. 

Based on those observations, a tracking data study over the recorded logs was made; it revealed a trend on the well logs compared to fracture geometry. The designed and post-job pressure matched fracture jobs were used and correlated to observe the trend of the fracture height, which was related to the most influential values on the logs.

Fig. 6. Example of a temperature log after injection with time intervals for subsequent logs; base line temperature log as a reference.

Fig. 6. Example of a temperature log after injection with time intervals for subsequent logs; base line temperature log as a reference.

These values can be utilized to confidently predict the fracture height without needing a temperature log. Among more than 20 values observed on the logs, e.g.,gamma ray, porosity, Poison’s ratio, Young’s modulus, silts, saturations, rock quality by reservoir point of view, etc., the study was made  to compare which parameters had a larger influence on the fracture height limitation.

Based on the observations made, the logs compared all inputs with fracture geometry predicted by design, fracture geometry adjusted by pressure match of the job execution (with real measured values), and with well log interpretation (additional criteria over the basic logs, gamma ray, sonic travel time, etc.). The jobs were visually arranged, with the observation and correlation of all data in images used to match the depth of the perforations, and  to compare the log values with the fracture height obtained  by design, and then compared with the real temperature log avail-able after injection, Fig. 7.

The study was made over more than 50 wells, which at that moment had available all the information required. Several more wells were also studied, but some of the information was not available or was incomplete. To have the most complete examples of better discretization on the parameters for this study, only the cases with all information available were used. Table 2 shows the details of the first base line study made, which corresponds to 40% of the wells analyzed as those cases for which observations were made with complete confidence. Note that some data was considered important for some experts, but was not mapped on the study, and is not shown in Table 2.

Table 2. First base line study to set the criteria of fracture height prediction based on logs

Table 2. First base line study to set the criteria of fracture height prediction based on logs

This decision was made as those parameters were constant for all cases, and did not influence the fracture height interpretation, e.g., perforation gun type4, base line frac fluid, and injection test fluid type, among others. The normal sequence of every job was:

1.Do a small injection test with potassium chloridebased brine; in all cases the same fluid is used.
2. Perform a step rate test and/or step down test.

Fig. 7. Example of the exercise made on every case of the study to analyze logs vs. fracture geometry suggested by computational analysis. Fracture geometry generated by computer is the yellow oval shape over the Young’s modulus (YM) and Poisson’s ratio (Pr) logs. The red dash section represents the fracture height related to all other parameters related by depth.

Fig. 7. Example of the exercise made on every case of the study to analyze logs vs. fracture geometry suggested by computational analysis. Fracture geometry generated by computer is the yellow oval shape over the Young’s modulus (YM) and Poisson’s ratio (Pr) logs. The red dash section represents the fracture height related to all other parameters related by depth.

3. Calibrate the injection with cross-linked fluid asper the planned one for the main job.
4. Run a temperature log (HRT), and study the datafrom step 3.
5. Execute the main job proppant fracture execution(based on step 3).
6. Perform a pressure match of the job, to be adjusted as per the HRT.
This study allowed us to upgrade the operation by swapping the sequence as per the following steps:
1. Interpretation of the logs.
2. Comparison of correlated wells to match the criteria on possible fracture height by dominant parameters.
3. Compare with fracture simulation by fracture design computational software.
4. Execution of the job.
5. Pressure match to evaluate differences on geometry results by software analysis.

The main advantage of the upgraded steps is that no temperature log is required, DataFRAC is avoided, thereby saving no less than 24 hours while optimizing the fracture design based on previous experience.

OBSERVATIONS AND CONCLUSIONS

The study’s conclusion was reviewed and utilized to set the forecasted or predicted fracture height in wells where the HRT was still required, to prove with confidence that the criteria found was reliable. So far, over 25 cases were observed to match the criteria of the fracture height prediction with less than a 15% error, which for the purpose of these main fracture job designs, the adjustments for time savings and cost optimization, and the impact on operations could be obtained.

The study set this criteria as reliable for application in field operations to confidently avoid the usage of a temperature log. The study continues to be fine-tuned as areas outside of the actual studied wells are still under development. It was found to have similar trends on criteria for those areas. Table 2 prevailed to be valid in terms of holding the most important criteria to confidently predict the fracture height based on log interpretation. The trends for this criteria narrows more when the type of reservoir and average job profile is similar to the ones pumped in the studied cases. Two job types, in terms of volume and rate, were used. When discretization is made at the same job approach, the fracture height can be predicted with minimal error.

OUTLOOK

1. The project proved to be valid based on previous data and a new test was made with the HRT to validate the prediction. More wells are being added to the study, and the addition of new cases is important when the well is drilled in areas where correlations are expected to be similar to studied scenarios. This will help to finetune the criteria to predict the fracture height with more confidence.

2. The study could be extended to other areas of application, i.e., prediction of fracture volume, fracture total geometry (length in particular, width as well), perforation enhancements5 by fracture observations, the fluid success ratio on fracture execution, and conditions to evaluate potential screen outs, etc.

 
 
 
ACKNOWLEDGMENTS

The authors would like to thank the management of Saudi Aramco for their support and permission to publish this article. The authors would also like to thank Nahr Abulhamayel and Yousef Noaman, for this project execution and the continuation of it.

The article was presented at the SPE Kingdom of Saudi Ara-bia Annual Technical Symposium and Exhibition, Dammam, Saudi Arabia, April 23-26, 2018.

REFERENCES
  1. Burra, A., Esterle, J.S. and Golding, S.D.: “Use of Temperature Logs in Coal Seam Gas Reservoirs:  Application to the Sydney Basin,” International Journal of Coal Geology, Vol. 143, April 2015, pp. 68-77.
  2. Buenrostro, A., Abdulkareem, H.S., Noaman, Y.M., Al-Driweesh, S.M., et al.: “Controlled Breakdown Technique Enables Proppant Fracture Placement by Enhancing Fracture Initiation; Fracture Pressure is Reduced when Applied,” SPE paper 187981, presented at the SPE Kingdom of Saudi Arabia Annual Technical Symposium and Exhibition, Dammam, Saudi Arabia, April 24-27, 2017.
  3. Hamid, O., Buenrostro, A., Suzart, W., Zoghbi, B., et al.: “New Methodology to Calibrate Rock Mechanics Parameters via Tuned Geomechanics
    Technique after Falloff Matched Curve,” paper presented at the 10th Edition of the Scientific and Technical Days, the 2nd Edition of the Scientific Exhibition, Oran, Algeria, October 5-8, 2015.
  4. Behrmann, L.A. and Elbel, J.L.: “Effect of Perforations on Fracture Initiation,” Journal of Petroleum Technology, Vol. 43, Issue 5, May 1991,
    pp. 608-615.
  5. Jackson, R.A. and Craig, D.P.: “Depth-ofInvestigation and Volume-of-Investigation Calculated from Fracture-Injection/ Falloff Test
    DFIT Data,” SPE paper 185029, presented at the SPE Unconventional Resources Conference, Calgary, Alberta, Canada, February 15-16, 2017.
ABOUT THE AUTHORS

Adrian Buenrostro is a Senior Petroleum Engineer in the Gas Production Engineering Division of the Southern Area Production Engineering Department of Saudi Aramco. He has 16 years of experience in the oil and gas  industry, which includes production and completion engineering, well management, training and mentoring, sales and marketing, particularly developed in the stimulation area in Latin-America, Europe, and Saudi Arabia. Adrian has been a Society of Petroleum Engineers (SPE) certified Engineer since 2015. He has written multiple technical articles for international publications and coauthored several more. Adrian developed the concept of channel fracturing with resin coated natural sand, which is patented.He received both his B.S. and M.S. degrees in Mechanical Engineering from the National Autonomous University of Mexico, Mexico City, Mexico. Adrian received the Mexican national Engineering Medal for his engineering work in 2001.

 

Mohammed Al-Abdrabalnabi is a Petroleum Engineering Scientist working with the Production Technology Division of Saudi Aramco’s E x p l o r a t i o n and Petroleum Engineering Center – Advanced Research Center (EXPEC ARC). The team targets well productivity enhancements such as laser oriented fracturing, optimized fracture conductivity, and water shut-off by chemical means. In 2015, he received his B.S. degree with honors in Chemical Engineering from King Fahd University of Petroleum and Minerals (KFUPM), Dhahran, Saudi Arabia.

 

 

Amro E. Mukhles is a Petroleum Engineer supervisor in Saudi Aramco’s Southern Ghawar Production Engineering Department. He has 12 years of experience in the oil and gas industry in areas like production optimization , well completion, stimulation, well intervention  operations as well as scale and corrosion mitigation. Amro received his B.S. degree in Petroleum Engineering from West Virginia University, Morgan town, WV, and his M.S. degree in Petroleum Engineering from the University of Texas at Austin, Austin, TX. 

 

 

 

Saad M. Al-Driweesh is a General Supervisor in the Southern Area Production Engineering Department, where he is involved in gas production engineering, well completion, and fracturing and stimulation activities. Saad is an active member of the Society of Petroleum Engineers (SPE), where he has chaired several technical sessions at local, regional and international conferences. He is also the 2013 recipient of the SPE Production and Operations Award for the Middle East, North Africa and India region. In addition, Saad chaired the first Unconventional Gas Technical Event and Exhibition in Saudi Arabia.He has published several technical articles addressing innovations in science and technology. Saad’s main interest is in the field of production engineering, including production optimization, fracturing and stimulation, and new well completion applications. He has 26 years of experience in areas related to gas and oil production engineering. In 1988, he received his B.S. degree in Petroleum Engineering from King Fahd University of Petroleum and Minerals (KFUPM), Dhahran, Saudi Arabia.