Casing-wear factors: How do they improve well integrity?

Casing-wear factors

Casing-wear factors form an integral part of casing-wear estimation techniques. In existing drilling practices, it has become customary to calculate the “field” casing-wear factors by back-modeling the measured casing-wear logs. It is expected that these reverse-calculated wear factors would more accurately represent the actual field scenarios.

However, a problem arises when these back-modeled wear factors are not able to accurately predict the downhole casing wear for further drilling activity in that well or for the drilling program of any other well, having a similar profile in the same field. This discrepancy has led to skepticism and reservations about the field applicability of the wear factors to actual drilling situations, and resulted in a loss of confidence in the existing modeling methods. As a result, most operators choose to over-design their casing strings, rather than accurately estimate casing wear to obtain a sound engineering design.

The authors believe that a lot of these suspicions are unwarranted and have emerged because of inconsistencies in accurately applying the casing-wear model. A large number of uncertainties involved in casing-wear prediction have led to significant misinterpretations of the existing casing-wear modeling techniques. The current approach to the problem is to “fudge” all of the unknowns on a single parameter—casing-wear factor. However, a detailed, comprehensive modeling method is required to account for all of the involved variables, to help achieve a more representative scenario and improve the accuracies of wear prediction. Different factors contribute toward under- or over-estimation of the wear, thus resulting in an overall, very complex analysis for any drilling program. Under-prediction of the wear can result in casing failure that compromises well integrity, while over-estimating it can cause over-designing of the casing string and unnecessary higher costs.

A thorough understanding of the underlying parameters involved in casing-wear estimation and a comprehensive step-by-step modeling approach, will help reduce unknowns, more accurately calculate the correct wear factors, and predict casing wear. This article provides a sound engineering basis for drilling engineers to effectively understand the challenges involved in casing-wear analyses. It provides detailed explanations to help understand the fundamental concepts and back-model appropriate wear factors for improving overall integrity in the well lifecycle.

KEY CASING-WEAR PARAMETERS

Some of the key parameters that influence casing wear modeling and wear estimation follow, along with further details on each.

1. Estimation of normal contact force. The normal contact force forms an integral part of the casing-wear equation. Inaccuracies in computing contact forces will lead to accumulating error in wear estimation. To calculate the normal contact force, the conventional soft-string model has been used in most applications. However, various torque and drag analyses presented in the industry prove that the more advanced stiff-string analysis, which includes the bending stiffness of the drillstring, should be performed for higher accuracy. In particular, the stiff-string model should be used in cases that have high doglegs, as the soft-string model has limitations in such scenarios.
   Another drawback of the conventional soft-string model is that it assumes contact between the tool joints and the casing at all points along the wellbore, leading to an over-estimation of the wear. The stiff-string model helps to depict a more realistic scenario and computes a value of normal contact force, only if any contact occurs and prevents over-estimation of the wear.
2. Detailed calculation steps and monitoring. Casing wear is estimated as a consequence of all of the operations that the well has undergone. Hence, it is important to capture and simulate all the operational steps in greater detail to improve the accuracy of wear estimation. During drilling of any wellbore section, the normal contact forces between the tool joints and the casing wall vary as the drill bit progresses farther down the hole to drill the formation. This variation in contact forces at all depths along the casing must be captured correctly and accounted for in wear calculations. For example, calculation of contact forces and the resulting wear as the bit penetrates each 30-ft section of the formation will represent more accurate wear calculation results, compared to a case in which these calculations are performed for every 500 ft of the drilled section.
   The length along the casing, over which the average casing wear is estimated and monitored, also should be sufficiently small to effectively capture the effect of all of the detailed calculation steps. Calculating and monitoring casing wear for every 10-ft section along the casing will more accurately represent a field scenario, compared to a case in which the average casing wear is reported for every 100 ft of the casing section. The objective of casing-wear analysis is to correctly predict the peaks of casing-wear logs, and it is important to independently track smaller casing lengths for all of the conducted operations to achieve this objective.
3. Well path and wellbore survey. All studies, conducted on casing-wear modeling, recommend that the accuracy of wear estimation depends tremendously on the accuracy of the well path and wellbore survey points. Wellbore surveys directly influence the calculation of contact forces, and coarse survey spacing will result in omissions of high dogleg points that cause maximum wear.
   In addition, wellbore surveys measured before running in the casing should be updated, using a stiff-string model after the casing is cemented in place to address any misalignments of the casing with the initially drilled openhole section. Gyro surveys should be run at a casing point, if there is excessive uncertainty in the actual wellpath or the cemented casing position inside the wellbore. The updated casing locations along the wellbore should help improve wear estimation and reduce suspicions about the back-modeled wear factors.
4. Non-linear behavior. Existing literature, based on laboratory experiments, explains that the wear groove formation on the inner casing wall does not increment linearly and displays a non-linear behavior, based on the groove depth. The wear factors estimated decrease with increasing wear depth for a given set of test conditions and approach an asymptotic value, as wear exceeds approximately 40%. The laboratory-reported wear factors are based on these asymptotic values of wear percent. However, for a given drilling situation, if the wear on the casing is below this asymptotic value, then an empirically derived correlation should be used to account for this non-linear behavior and correct the wear-factor values.
   Hence, the experimental wear factors would result in under-predicting the wear, if this non-linear characteristic is not accounted for in a case of small values of wear percent, while the back-modeled wear factors would erroneously be higher than the actual value. 
5. Multiple wear factors. Different drillstring and casing configurations are used to cater to different drilling programs, and assuming that casing wear along the entire wellbore can be simulated, using one single wear factor will be erroneous. For example, to drill multi-lateral wells, special aluminum casing pipes are used in the main borehole from where a lateral is to be initiated. Hence, in this case, separate wear factors should be used in the casing-wear analysis for aluminum and steel casings.
   Casing-friendly tool joints and pipe protectors are used in certain portions along the drillstring. Using a single-wear factor to represent the more aggressive and friendlier types of drillpipe tool joints in a single drillstring will lead to erroneous wear estimation. In addition, using multiple wear factors is especially useful while back-modeling wear factors from casing-wear logs, as different values of wear factors might be required for different casing sections to accurately match the field data.
6. Wear groove locations. One of the critical drawbacks of existing casing-wear modeling techniques is the failure to accurately estimate wear groove locations at any cross-section along the casing. Field experience has shown that more than one wear groove can be generated at the same casing depth, if the contact points between the casing and tool joints change as the drilling activity progresses. The soft-string model, used conventionally to perform wear analysis, is not able to identify variations in contact locations and hence, at any depth, higher values of wear are predicted, assuming that the contact is only at one position at a cross-section.
   
   

   Fig. 1. Variation in contact positions of the drill pipe with inner casing wall, estimated by applying the stiff-string model.
   Advanced stiff-string models are able to accurately estimate the points of contact between the casing and tool joints as drilling proceeds, and help perform a more realistic analysis. Figure 1 estimates this variation in contact positions of the drillpipe with the casing wall along the wellbore depths by applying the stiff-string model. Accurately monitoring the casing cross-sections at all depths, and estimating wear, only if there is any tool joint contact at a location, improves wear predictions.
7. Rig site operations. The total wear on the casing is a consequence of all the operations conducted through it. It is crucial to vigilantly track each operation performed that could contribute to wear. For example, a conventional drilling operation can comprise a short reaming of the hole after drilling every stand. It is important to account for any additional wear resulting from this small reaming operation, as over the course of the entire drilling program, this additional wear would accumulate. Field observations suggest that casing also wears out, even in situations where the drillstring is not rotating, but is just reciprocating inside the casing shoe, resulting from wave heave in an offshore environment. Hence, it is important to capture all downhole scenarios that could contribute to wear.
   Additionally, during rig site operations, the values of the drilling parameters, such as weight-on-bit and rotational speed, can vary. Accounting for variations in these parameters, to estimate normal contact force and the resulting wear, should help improve the accuracies of the back-modeled wear factors.
8. Casing-wear logs. Obtaining accurate casing wear measurement logs is crucial when trying to estimate casing wear, using back-modeled wear-factors. Multi-finger caliper logs are generally considered to be more accurate compared to ultrasonic logs, as the estimation of casing wall inner diameter is a direct physical measurement for caliper logs, whereas ultrasonic logs determine the casing wall thickness as an indirect estimation derived from other parameters. All proper precautions should be taken while running these logs, to be as accurate as possible when associating a particular casing depth with a certain value of measured casing wear groove depth. Inaccurate casing-wear logs will certainly result in erroneous back-modeled wear factors and incorrect wear predictions.

All of these parameters and uncertainties, listed for casing-wear estimation, affect significantly the accuracy of the wear-prediction models. An improper analysis, performed by discounting any of these parameters, will result in incorrect back-modeled wear factors and cause an inaccurate casing wear estimation for the next set of operations, when these wear factors are used.

IMPACT ON WELL INTEGRITY 

Wearing out of the inner casing wall results in degradation of the casing strength and causes a reduction in the burst and collapse pressure ratings, thus compromising casing and well integrity. Casing-wear factors affect the determination of the remaining wall thickness after casing wear, and thus have a great influence on estimating the overall well integrity. Inaccurate back-modeled wear factors can result in an under-estimation of the groove depth. Casing-wear analysis, performed by discounting some of the underlying uncertainties, would lead to estimation of inaccurate burst and collapse pressure ratings of the worn-out casing. 

Fig. 2. Collapse and burst pressure rating estimations for worn-out casing.

Several case studies were performed to understand the importance of the different influencing parameters, while attempting to estimate the back-modeled wear factors and the subsequent casing wear. Three variables—detailed calculation steps, accounting for wellbore tortuosity, and multiple wear factors along the drillstring—were analyzed together to understand their combined influence. As a result, higher resulting wear on the casing was predicted. Figure 2 shows the difference between the burst and the collapse pressure ratings for the worn-out casing when all these three factors are considered as compared to a wear analysis performed by discounting them. Here, Analysis 1 represents the situation when these three parameters are ignored, while Analysis 2 applies them as described above. For more details on the parameters of this case study, please refer to the paper, SPE/IADC 173053.

The reduction in casing collapse and burst-pressure ratings, estimated in Analysis 2, emphasizes the importance of accounting for all the underlying uncertainties while performing a casing-wear analysis. A comprehensive wear-estimation method would result in a more accurate prediction of the remaining casing wall thickness and hence, a more reliable collapse and burst pressure rating for the worn-out casing. For the simple case presented in Fig. 2, consisting of a single drilling operation, reductions in pressure ratings of up to 20% for collapse and 15% for burst were estimated, which may have a significant influence during casing selection and design. Standard API equations were used for collapse and burst-pressure estimation. A similar analysis for casing integrity also can be performed for the maximum allowable wear limit, based on all the designed load cases for the wellbore and comparing it against the predicted casing wear for a given well.

Fig. 3. vonMises equivalent-stress estimation and design limits for worn-out casing.

Figure 3 represents the vonMises equivalent stress plot and the design limits for the worn-out casing for both Analysis 1 and Analysis 2, performed here, and compares it with that of the new casing. Casing wear leads to a reduction in wall thickness of the casing, hence the design limits and the vonMises equivalent stress ellipse reduce significantly for carrying out all operations safely. The more comprehensive Analysis 2, performed here, results in even a further reduction in the design limits for safe operations.

The safety factors applied in Fig. 3 for the new casing design are 1.2 for burst and collapse, 1.3 for tension and compression, and 1.25 for the triaxial stress limit. Two load cases for pressure test and casing evacuation have been analyzed to understand the influence of casing wear on the underlying casing design limits. The safety factor for casing evacuation load case reduced from 1.18 for collapse for new casing to 0.66 after casing wear in Analysis 2, while the reduction in safety factor for triaxial stresses was from 1.7 to 1.35. This significant reduction in collapse safety factor needs to be accounted for appropriately during casing design, to mitigate any casing failures.

COMPREHENSIVE APPROACH 

The significance of performing a comprehensive casing-wear analysis, and its corresponding effect on wear factors and wear groove estimation, has been analyzed in this study. Correct back-modeled wear factors should be used to perform an accurate casing-wear analysis, and for precise estimation of remaining wall thickness after wear. The authors believe that all of the skepticism concerning casing-wear factors and casing-wear estimation methods has resulted from ignoring some of the fundamental underlying parameters while performing casing wear analyses. A detailed, comprehensive approach, addressing all of the involved uncertainties, should help simulate a more realistic scenario and improve the accuracies of wear estimation. The objective of performing wear analysis is to accurately predict the peaks of casing wear logs and the minimum remaining wall thickness. The authors believe that this can be achieved most certainly by following the detailed steps, as outlined in this study, and disregarding any skepticism concerning casing-wear factors.  

ACKNOWLEDGEMENT

This article is based on SPE/IADC paper 173053, presented at the 2015 SPE/IADC Drilling Conference and Exhibition in London, UK, in March 2015.

REFERENCES

1. Calhoun, B., S. Langdon, J. Wu, J. et al., “Casing wear prediction and management in deepwater wells,” SPE paper 137223 presented at the SPE Deepwater Drilling and Completions Conference, Galveston, Texas, Oct. 5–6, 2010.
2. Hall, R.W. Jr., A. Garkasi and G. Deskins, G. et al, “Recent advances in casing wear technology,” SPE paper 27532 presented at the IADC/SPE Drilling Conference, Dallas, Texas, Feb. 15-18, 1994.
3. Hall, R.W. Jr. and K. P. Malloy, “Contact pressure threshold: An important new aspect of casing wear,” presented at the SPE Production and Operations Symposium, Oklahoma City, Okla., April 17-19, 2015. 
4. Kumar, A., J. Nwachukwu, J. and R. Samuel, “Analytical model to estimate the downhole casing wear, using the total wellbore energy,” Journal of Energy Resources Technology, No. 135, 2013, pp 1–8.
5. Kumar, A. and R. Samuel, “Modeling method to estimate the casing wear caused by vibrational impacts of the drillstring,” SPE paper 167999, presented at the IADC/SPE Drilling Conference and Exhibition, Fort Worth, Texas, March 4-6, 2014.
6. Mitchell, S. and Y. Xiang, “Improving casing wear prediction and mitigation, using a statistically based model,” SPE paper 151448 presented at the IADC/SPE Drilling Conference and Exhibition, San Diego, Calif., March 6-8, 2012.
7. Samuel, R. and D. Gao, “Horizontal drilling engineering: Theory, methods and applications,” Sigmaquadrant Engineering, Houston, Texas, 2014.