You must log in to edit PetroWiki. Help with editing

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information


Predicting CT fatigue

PetroWiki
Jump to navigation Jump to search

Over the years, attempts have been made to track the working history of coiled tubing (CT) strings in service to maximize the service utility of the tube while minimizing fatigue failures. As a result, three commonly used methodologies for predicting the fatigue condition of the CT were developed.

Running feet method

A relatively simplistic approach used to predict the working life of coil tubing is commonly described as the “running-feet” method, in which the footage of tubing deployed into a wellbore is recorded for each job performed. This deployed footage is then added to the existing record of footage deployed in service for any given string. Depending upon the service environment, type of commonly performed services, and local field history, the CT string is retired when the total number of running feet reaches a predetermined amount.

The running-feet method offers the service vendor relative simplicity of use, requiring only that the maximum depth of CT deployed into the wellbore be recorded. However, there are numerous limitations of this fatigue-tracking method as a reliable means of determining ultimate working life of a CT service string. Several limitations are described next.

The value of maximum footage to retirement for any CT string is based on the service vendor’s previous experience with the same type of tubing, performing wellsite operations with similar well depths and types of service. In this method, there is generally no consideration given to duration of corrosive services performed or effects of exposure to atmospheric corrosion.

  • The running-feet method typically focuses on the specified outside diameter (OD) of the CT string in service, with minimal consideration for tubing wall thickness, tube material type, and yield strength.
  • The running-feet method does not have a means of accounting for variations in tubing guide arch radius, service reel core radius, internal pressure loading, or identification of specific tube segments where additional bending cycles are applied.
  • The working life-derating method used in the running-feet approach cannot be extended to different tubing sizes or operating conditions. This method can be used only where working history for the specific tube material, geometry, and surface handling equipment has been gathered and analyzed to yield the prescribed maximum running-feet value.

Trip or empirical method

A natural extension of the running-feet fatigue derating approach can be found in what is commonly described as the “trip” method. In the trip method, numerous improvements have been incorporated to the running-feet approach, providing greater reliability in predicting working life of the CT string.

One major improvement entails evaluating the CT string as a series of partitioned segment lengths that can range from 100 to 500 ft long. This approach applies a greater sensitivity to the working life analysis by identifying sections of the CT that are subjected to more bending cycles than others during a specified service. The number of trips over the service reel and tubing guide arch for each discrete segment can then be tracked and recorded. When employing this method, a reduction in the length of the section increment increases the accuracy of the bend-cycle record. This type of analysis makes it possible to identify the CT string segments that have experienced the most bend-cycle fatigue damage.

Another major improvement with the trip method incorporates the effects of internal-pressure loading. For a given tubing guide arch and service reel core radius, CT bend-cycle fatigue life decreases significantly with increased internal pressure loading. The evolution of the trip method incorporated extensive CT bend-cycle fatigue testing using full-scale service equipment (injector, tubing guide arch, and service reel) and varying amounts of internal-pressure loading. In this scenario, numerous bend-cycle fatigue tests are performed for a given size of CT at specified amounts of internal pressure.

Data recorded in these tests were initially used to create a database for statistical projection of CT working life. From these types of tests, a segment of the CT string that had accumulated a considerable amount of bend-cycle fatigue damage can be identified, thereby providing the user with options for removing the heavily damaged segment of tubing from service.

As more full-scale fatigue cycle tests were performed, trends in CT fatigue were identified for various pipe sizes, tube geometry, and internal-pressure load conditions. Analysis of these trends provided the service vendor with the ability to “curve-fit” the data points and derive empirical coefficients that were incorporated into conventional multiaxial fatigue-life prediction approaches, yielding the early CT fatigue prediction models.

The aforementioned improvements in fatigue damage tracking realized by the trip method offer enhanced accounting of operating conditions present when the bend-cycling events occurred, along with a greater sensitivity of identifying tubing-string segments subjected to bend cycling. The limitations with the trip method of empirical modeling include:

  • The derived empirical coefficients for fatigue-damage are generally different for each combination of:
  • CT material
  • OD
  • Wall thickness
  • Bending radius
  • Bend-cycle testing using full-scale equipment is required to obtain the fatigue coefficients experimentally (expensive and time consuming).
  • The trip method does not incorporate tube body damage incurred as a result of well-servicing operations. This type of damage includes:
  • Exterior tube body wear,
  • Interior and exterior corrosion (atmospheric and industrial)
  • Nicks, cuts, or scarring resulting from contact with surface handling equipment
  • The test data obtained from fatigue bend-cycling machines is usually at a constant internal pressure. In well-servicing operations in which fluid pumping is required, the amount of internal pressure present in the CT varies along the entire length of the string. Therefore, as the tubing is deployed and retrieved, each section of the string has a different internal pressure at the point where bend cycling occurs.
  • The varying internal-pressure loading at the point of bend cycling requires a complicated record and prediction procedure to provide a realistic working-life prediction. This requires investment in surface recording instrumentation and sophisticated data collection systems, such as portable computers, as well as complicated tubing management software systems for tracking and maintaining up-to-date records of the compiled tubing working life.

Theoretical method

This method is used for predicting bend-cycle fatigue. It incorporates the same approach developed in the “trip/empirical” method for estimating the bend-straighten-pressure history for each segment of tubing along a string. However, a theoretical model based on the fundamental principles of mechanics and fatigue is used to estimate the stress, strain, and fatigue behavior of each section in the string.

The theoretical modeling of fatigue typically involves use of “plasticity” algorithms and “damage” algorithms. The plasticity algorithm is used to estimate the stress and strain history of the CT material as it is bent or straightened over a particular bending radius at a particular internal pressure. The damage algorithm uses the concept of cumulative fatigue damage to quantify the reduction in working tube life caused by each bend or straighten event. The approach is very mechanistic, for instance, taking into account the fact that the pressure during each bending or straightening event can be different. The fatigue damage computed for each event is summed throughout life and is usually expressed as a percentage of the predicted working life. Since each section of tubing along the length of a string can endure differing bend-straighten-pressure histories, the damage profile can (and usually does) vary along the length of a typical working string.

The plasticity algorithm in the theoretical model requires input of the specific material properties. These properties come from two types of testing. First, the aforementioned low-cycle fatigue testing conducted on axial coupons, and second, full-scale data typically taken from a CT fatigue testing fixture, or from full-scale equipment.

The low-cycle fatigue data are used to compute both elastic material properties and the cyclic stress-strain curve for the particular CT alloy. Although these properties are generated for axial loading only, they serve as the “constitutive relations” (i.e., the relations between stress and strain) for a multiaxial plasticity algorithm, which is capable of estimating the history of all stress and strain components (axial, hoop and radial) in the tubing.

Since the plastic deformation caused by bend cycling is so severe, it was determined that conventional plasticity theory was inadequate to describe the behavior of CT accurately. Conventional theories tended to overpredict phenomena such as ballooning and wall thickness reduction. To overcome this, new theories were developed specifically for CT. These models are effectively “tuned” to specific alloys by collecting data from constant pressure bend-cycling tests conducted on laboratory testing fixtures (although data from full-scale equipment can also be utilized) to supplement the low-cycle fatigue data. The empirical parameters derived from these test results cause the algorithms to do an excellent job of estimating ballooning and wall thickness reduction, as well as fatigue under complex loading histories.

The use of empirically derived data in this approach assures that the model can be mapped back to realistic behavior exhibited by real CT sections. In reality, CT mechanical properties must be allowed to vary within a particular grade. For this reason, it is important to collect as many experimental data points as possible to characterize the scatter caused by typical material variation. The greater the number of experimental data points, the stronger the statistical validity of the model.

Advantages of theoretical models

The advantages to the use of theoretical models include greater accuracy of bend-cycle fatigue life prediction with the capability to predict fatigue life for variable loading conditions. The use of such an algorithm in the field is dependent upon the use of a reliable string-management routine that keeps track of the depth and pressure history of the string throughout its use and is capable of computing the bend-straighten-pressure history for each section of tubing, based on that depth-pressure log. Fortunately, software is available commercially to implement the approach either in real time or following the job

Another advantage of this model is its ability to make quantitative predictions that are based on statistically significant quantities of empirical data. Fig. 1 shows estimated trips to failure vs. internal pressure for an 80-ksi tubing material with three different diameters and a 0.134-in. wall thickness, run over a 96-in. reel core diameter and 72-in. tubing guide arch. (The life prediction algorithm used to make these predictions comes from the Flexor TU4 Life Prediction Model).

In the field, the model is capable of monitoring the fatigue life profile along the length of the tubing. But, more importantly, it can predict the effect of impending jobs on that profile. This allows the engineer to modify the use of each string to effectively maximize its life, avoiding potentially dangerous situations and reducing costs.

Limitations of theoretical models

The limitations of theoretical models include:

  • Two sets of input data are required to characterize a particular material: (a) low-cycle fatigue data taken from axial coupons and (b) constant pressure bend-cycle data taken from CT fatigue fixtures or full-scale equipment. The greater the number of data points from the latter set, the stronger the statistical validity of the model.
  • Current theoretical models do not incorporate tube body mechanical damage incurred as a result of well intervention or drilling operations. However, research is under way to develop models that can estimate the influence of surface defects. Such a model exists and is currently being refined. This type of damage includes:
  • Exterior tube body wear
  • Interior and exterior corrosion (atmospheric and industrial)
  • Nicks, cuts, or scarring resulting from contact with surface handling equipment
  • The implementation of a theoretical approach must be in concert with a routine capable of estimating the varying internal-pressure loading at the point of bend cycling. This requires a reliable record of not only the depth-pressure history of a string throughout its use, but also a record of any string modification (section splicing and/or removal). This requires:
  • Investment in surface recording instrumentation
  • Sophisticated data-collection systems, such as portable computers
  • Complicated tubing management software systems for tracking and maintaining up-to-date records of the compiled tubing working life.

CT management

The service vendor must maintain a history of the various services for which each CT string has been employed to ensure prudent management of CT strings. This tubing-string record should include the following information (as a minimum):

  • Pressure/bending fatigue-cycle history and locations of repeated bending cycling. The data for these records should be obtained from daily service activity reports or through electronic record-keeping devices.
  • If pressure bend cycling is not recorded, then the service vendor should provide the “total running feet” (records must reflect footage into and out of the well).
  • Maximum pumping pressures through the CT string when stationary.
  • Exposure of tubing string to acid service. This record should list the number of:
  • Acid jobs performed
  • Type and volume of the acid system pumped
  • Duration of the acid-pumping program
  • Vendor-recommended derating factors for the string
  • Locations of:
  • Welds
  • Identification of type of weld
  • Observations of deformity, ovality, or surface damage
  • Locations in the CT string where tensile loads exceeding 80% minimum yield were placed upon the pipe
  • A detailed record of any splicing or section removal that takes place along the length of the string.

References

Noteworthy papers in OnePetro

Use this section to list papers in OnePetro that a reader who wants to learn more should definitely read

External links

Tipton, S, M. and Nelson, D. V., Advances in Multiaxial Fatigue Life Prediction for Components with Stress Concentrations, International Journal of Fatigue, Vol. 19, No. 6, pp. 503-515 (1997)

Newman, K. N., Brown, P. A., Development of a Standard Coiled-Tubing Fatigue Test http://dx.doi.org/10.2118/26539-MS

Tipton, S. M, Surface Characteristics of Coiled Tubing and Effects on Fatigue Behavior, http://dx.doi.org/10.2118/38411-MS


See also

Coiled tubing fatigue

PEH:Coiled-Tubing_Well_Intervention_and_Drilling_Operations