Back to Key Issues
Soft Formations - Best Practices
Contacts
| Tony Settari | ASettari@TaurusRS.com | Taurus Reservoir Solutions |
| Dale Walters | DWalters@TaurusRS.com | Taurus Reservoir Solutions |
| Karim Zaki | karim@advantekinternational.com | Advantek International |
Summary
The analysis of field data from soft formations shows that a large amount of damage exists for PWRI injectors, and their performance can vary widely; depending on the completion method and how they are operated. Therefore, there is a considerable scope for improving the performance of these wells. Best Practices are an aid to such optimization, consisting of general observations and experience generated in the project. These general guidelines should be however supplemented by specific analysis of each project.
Key Issues
| Data Gathering | Data Analysis |
| Completion Selection | Drilling and Operations |
| Mechanisms/Models | Performance |
Recommendation 1.1: If possible, establish wellbore friction by direct measurements for accurate conversion of WHP to BHP.
This recommendation is not specific to soft formations, but it is especially important because soft formations tend to have high permeability and therefore low pressure gradients.
Ideally, continuous BHP measurements (permanent gages, optical fibers) are preferred. However, it is sufficient to use spot BHP measurements (injection/falloff, PLT) to verify wellbore friction calculations from WHP.
Conventional calculation of frictional pressure drop, based on pipe roughness, can be incorrect if friction reducers are used, Teflon coated tubing is run, etc. Note that reducing the value of roughness to zero does not decrease the friction beyond a certain limit, using conventional equations. Achieving the correct values may therefore require changes to the software.
An example is the friction measurement in Heidrun wells (Reference 1) and the consequences for BHP data interpretation (Reference 2).
Recommendation 1.2: If possible, measure the surface injection temperature.
Bottomhole temperature (BHT) is important for evaluation of stresses and fracturing pressure. Also, any reservoir or geomechanical simulation will require the BHT for the injection well. The experience gained in analyzing horizontal injectors has shown that one of the key factors for accurate matching of the observed data is accurate BHT for the injector. This is particularly important for onshore projects where injection temperature can vary.
Although this recommendation is also general, it is important for soft formations. While it is often argued that thermal effects are small because of a low modulus, it must be remembered that the pressure gradients are also small due to high permeability. Therefore, the relative importance of the thermal stresses is just as large as in lower permeability, "harder" formations.
Ideally continuous BHT measurements (optical fibers) are preferred, but one can use spot measurements (PLT) to verify BHT calculations from the WHT.
For surface measurements, the guidelines include:
At high rates, BHT can be assumed to be close to the surface temperature. At low pressure, it may be necessary to use thermal wellbore hydraulic software to translate the wellhead temperature to a bottomhole value.
Recommendation 1.3: Record both the rate and duration of the injection each day (rather than just the daily volumes).
If the injector is on for less than 24 hrs/day, use of the daily volume to calculate rate will result in a lower than actual rate. This will cause artificial features of the data in diagnostic plots (in particular in the pressure vs. rate plot) and lead to erroneous conclusions (e.g., about fracturing). Therefore, check the percentage of time that the injection is on or off or the hours on or off during the day (these figures are generally available) and correct the rates to true injection rates.
When the tubing head pressure data have been already converted to bottomhole values, they should be treated with caution. For example, Figure 1 shows a portion of the rate and bottomhoe data for an example well. It shows that the bottomhole pressure is increasing with decreasing rate, and results in a pressure vs. rate plot with a negative slope, as shown in Figure 2. The likely explanation is that the rate during injection is constant but that the injection daily time varies. The variation in the bottomhole pressure is then produced by a falsely-computed variation in the frictional pressure drop.
Figure 1. Example of already interpreted rate and bottomhole pressure.
Figure 2. Diagnostic plot based on accepting the bottomhole pressure interpretation in Figure 1.
If only daily volumes and surface pressures are available and it is known that the injection is significantly less than continuous, such data can be very difficult to analyse.
Recommendation 1.4: Record both the rate and pressure each hour during the first days or up to weeks of injection.
The duration of the period should be based on the estimated duration of transient effects due to compressibility, and also the time necessary for stabilization of the plugging effects. Hourly data give more information than the usual daily data, and provides better averages.
Recommendation 1.5: Determine the overall depletion or repressurization trend in the injection area (i.e., average reservoir pressure applicable to the injection well).
The variations in reservoir pressure can significantly affect the interpretation of the Hall plots, II and the damage. For example, this effect is significant in the Phillips T field data set (Reference 2, Section 4) and distorts all diagnostic plots.
One of the main cause of interpretation difficulties for pressure/rate plots and Hall plots is the lack of knowledge of reservoir pressure evolution in time, which must be evaluated as accurately as possible. This recommendation is also valid for other types of formations, but what is particular here as a cause of misinterpretation is that even in radial flow conditions the slope of the pressure-rate line is very small and a fracturing regime can be confused with radial flow with changing reservoir pressure.
Severe damage can be hidden by interpreting a low differential pressure at a given reservoir pressure while in fact the reservoir pressure is lower or decreasing with time. Therefore, an accurate knowledge of reservoir pressure evolution is essential for good interpretation.
Recommendation 1.6: If possible, measure the skin caused by the completion equipment such as screens and excluders.
In order to determine the fracture pressure and damage in the formation and/or fracture, it is necessary to subtract the completion skin from the data. Opinions of operators vary as to the skin of some completions, notably Excluder screens. Some operators believe that large pressure drops may exist, but Marathon offered data showing tests with a 230 micron Excluder screen that showed low pressure drops (24 psia for pumping at 18 bpm into a 16 ft long screen, Reference 3).
Recommendation 1.7: Determine if injection equipment pressure limitations or other operational constraints are reflected in the data.
Operational constraints may be important for understanding the character of the diagnostic plots. Pressure limitations can be confused with fracturing pressure, and if their values are close, it may be impossible to separate them (for example, refer to the ELF3 data set in Reference 2).
Recommendation 1.8: Obtain an accurate estimate of the virgin reservoir permeability and the true injection height (falloff analysis, PLT).
The magnitude of the estimated damage is highly dependent on the virgin properties of the formation, which are often poorly defined. For example, a permeability estimate of between 1 and 2 darcies is often considered to be sufficiently accurate. However, detailed analysis could show that, in an extreme case, this could reflect a real virgin permeability of 2 darcies and a mean overall damaged permeability of 1 darcy (50% loss of injectivity). Therefore, to determine the amount of damage, it is essential to have an accurate knowledge of the virgin permeability.
Generally, soft sands exhibit very a high kh and sufficient injectivities for clean water with only few bars of differential pressure. Even with a severe damage and loss of injectivity, the differential pressure can be acceptable. However, this differential pressure can be very high when dealing with soft sands with high permeability but very low h or small entry height into the formation. Then, an accurate knowledge of the real injection height or entry height is essential.
Recommendation 1.9: Water Quality Measurements.
Measure water quality at least once a week at the beginning of PW injection and every time abnormal changes in injectivity are observed. These measurements should include OIW and especially TSS. The potential to form scale should be eliminated. This is a basis for further efforts to correlate the relationship between damage and water quality. The lack of water quality measurements in the present soft formations database does not allow this exercise.
2. Methodologies for Data Analysis
Recommendation 2.1: Follow the general methodology outlined in Section 1 of Reference 2 for the analysis of field pressure and rate data.
The main steps in the analysis are:
Recommendation 2.2: Check the quality of the conversion of THP to BHP before doing any further analysis.
The important checks are:
A negative slope may be caused by an incorrect caluclation of the friction pressure, or by incorrect interpretation of the rates (see Recommendation 1.3).
Recommendation 2.3: Check the consistency of the interpretation by independent calculations and comparison with other data.
The more data that are considered, the more reliable the interpretation becomes. The most common criteria to check are already included in Recommendation 2.1. For example:
In addition, well test data, such as initial injectivity or drawdown/buildup tests, can provide independent estimates of the initial completion skin, reservoir pressure and the reservoir kh.
Recommendation 2.4: Do each analysis in-depth since any given well may have unique features not covered in the general guidelines of Recommendation 2.1.
There is always danger in applying a standardized methodology to a given set of data, in particular if the guidelines for data collection in Section 1 have not been followed. Usually problems are indicated by inconsistencies as per Recommendation 2.2 above.
Applying the Data Analysis Methodology to Soft Formations
The general monitoring and analysis methods apply for soft sands. Among them, interpretation methods and tools were used to analyze well performance for soft formations. In this section, the points specific to soft formation will be emphasized.
The simplest and most powerful methods to diagnose the type of injection regime are the pressure/rate and the Hall plot analyses. Although there are situations where radial flow or fracturing regimes can be easily identified, there are many cases where it is difficult to discriminate between the two regimes. The impact of plugging on these plots is relatively well known and one of the main causes of interpretation difficulties is poor knowledge of reservoir pressure evolution - which must be evaluated as accurately as possible. This is valid for every formation but what is particular to many soft formations, as a secondary cause of misinterpretation, is that even in radial flow conditions, the slope of the pressure-rate line is very small and fracturing regime could be confused with radial flow under conditions of increasing reservoir pressure.
Other simple methods of well performance evaluation are the different analytical tools included in the toolbox. One difficulty here is to discriminate between the different competitive skins (completion, fracturing and plugging). This difficulty is less due to the tool capabilities than to inadequate basic data (see data acquisition).
For plugging in radial flow conditions, use the simple one-dimensional numerical model that is in the Toolbox ("PWRAD"). This tool is not specific to soft formations, and uses a simple two-parameter damage law. It has been applied to both hard and soft formations and successfully matched the field data in all cases. It is hypothesized that the parameters of the damage law can be determined in the laboratory and that the formulation is suitable for incorporation in commercial numerical simulators. In principle, this code could be used in a predictive mode, but there is a need for further work to evaluate the consistency between parameters derived from laboratory experiments and from history matching.
Recommendation 3.1: Use the Completion Selection Tool from the PWRI Toolbox to "highgrade" the most appropriate completion techniques.
The Completion Selection Tool is a spreadsheet that represents the collective experience of the operators in rating completion techniques in terms of likely success for long term operation of the well. As such, it includes factors such as possibility of sand production, ease of workovers, etc.
The tool only provides general ratings! Llocal constraints may override the recommendations. Also, it is also likely to come up with several choices with no clear preference. Therefore, it does not completely replace an individual approach to completion design.
Use the "Completions Selection Tool" for initial screening, to select the best types of completions and to eliminate completions which are not appropriate. This screening exercise does not replace (or eliminate the need for) further use of more sophisticated tools, that are capable of including more variables and allowing one to make final choices.
Recommendation 3.2: Compare competing alternatives in terms of completion skin.
The Completion Selection Tool may yield several alternative completion methods with similar ratings. In the absence of other operational or logistical preferences, the expected skin from these alternatives can be quite different and can be used as a secondary selection criterion. Methods and data for determining completion-related skin are found in Reference 5.
Use formulations and charts from the "Completion Skin Tool" in the PWRI Toolbox to estimate the skin for various completion types. In particular, a detailed study of perforated completions, assuming that the perforations can be plugged or can collapse, showed that very large skins could result. In addition, in some instances, turbulence can cause a large additional skin. These two factors potentially contributed to the high total skins observed in the analyzed data sets.
Recommendation 4.1: Minimize the drilling damage, particularly in high permeability zones.
Differences in the drilling and completion program can account for large differences in the performance of the injectors. Also large drilling damage makes it difficult to design an efficient perforating strategy. An example of the differences in injectivity, which can be at least partially attributed to differences in drilling damage, is the comparison of KMG field G with the analog BP field D (Ref. 2, Sec. 5.3).
Recommendation 4.2: Consider using non-penetrating charges for cased and perforated completions.
Marathon recommends KISS charges for injection, in preference to deep penetrators to reduce collapse damage in the perforation tunnels. Whatever type of perforator is selected, remember that collapsed perforations can potentially cause very large skin as shown by detailed modeling (Reference 6). KISS charges are designed to create a big hole and only penetrate casing and cement with very minimal formation contact.
However, this recommendation is contingent on not creating deep drilling and cementing damage. In the case of deep drilling mud invasion, long perforations may be required to overcome the damage, with the associated risk of collapse.
Recommendation 4.3: Shut in the injectors for sufficient time to prevent backflowed material from being injected back into the formation(s). Minimize backflow!
The risk of plugging with suspended backflowed material due to crossflow or early resumption of injection has been well demonstrated in the Heidrun field (Reference 1). The operational measures recommended, based on this experience, include:
Recommendation 5.1: Be careful when using conventional methods or tools for soft formations.
The physical mechanisms (interactions between reservoir flow, fracture flow, formation plugging, completion skin and fracture propagation) are more complicated than for hard formations. This is because of the particular geomechanical behavior of soft formations (plasticity, dilation, stress- dependent porosity and permeability). As a result, one has to be careful when using conventional methods or tools for soft formations.
General guidelines on how to identify geomechanical effects from field data, drawn on experience and the literature, are included in Reference 7.
Recommendation 5.2: What existing models can't do.
None of the existing simulation models are particularly satisfactory for soft formations. The industry still needs simple, practical tools. If a model were to be developed, a reasonable approach is to first develop or implement the additional needed physics into existing fully integrated coupled models and then attempt to built simpler (analytical) tools by performing parametric studies using the full models, in order to define what can be simplified, under which conditions, and the loss of accuracy that the simplifications cause.
Recommendation 5.3: If possible, identify the important geomechanical mechanisms which may be present.
It is important to recognize and identify geomechanical effects, such as stress-dependent permeability, dilation, joint or fracture opening, based on available data. A checklist of some of the mechanisms is found in Table 1 of Reference 7.
Recommendation 5.4: Use the simplest possible model that is compatible with the physics of the problem.
If there are no indications of strong geomechanical effects, a conventional reservoir simulation may be adequate. In some instances, stress-dependent properties can be successfully approximated by pressure-dependent properties. In complex situations, coupled geomechanical modeling may be necessary.
As it was concluded in Reference 8, all current models have some shortcomings and their results must be critically evaluated in that light.
Recommendation 5.5: Do not use jointed (fractured) media models to represent fracturing in soft formations which do not have fractures or joints.
Experience with modeling PWRI fractures in horizontal injectors has shown that models using a continuum representation of sets of joints may overestimate the injectivity and may not properly model fracture pressure. Modeling with discrete fractures is preferred in most soft formations except in fractured chalk.
More Detail?