Advanced Assessment of Drillpipe Fatigue and Application to Critical Well Engineering

Abstract

Abstract This paper addresses the nature and assessment of drillpipe fatigue. A detailed examination and statistical analysis of the three published sources of drillpipe S-N curve data supplemented by principles of machine design, provides a consistent basis to determine S-N curves for drillpipe of arbitrary size and grade. The machine design approach translates fatigue life characteristics of a small sample to that of drillpipe as a larger component. This translation accounts for size, manufacturing process and influences such as environment and vibrations. The resulting fatigue behavior agrees with previous data on grades D and E collected on full scale pipe specimens. Further, the current approach is demonstrated to be successful in predicting the fatigue behavior of higher drillpipe grades (G, X, S, V) for which scant data is currently available. Finally, the work is shown to be consistent with Hansford and Lubinski's 1964 analysis which informs the fatigue guidelines in API RP7G. Given the S-N curve and magnitude of variable loads, fatigue assessment requires tracking the number and magnitude of stress cycles. Since wellbore tortuosity and changes of drilling parameters with depth render the calculations cumbersome, a computer simulation is necessary. The simulation enables rapid and efficient evaluation of hole section, drillstring, trajectory and drilling parameter permutations. This is particularly useful in developments where a number of wells are drilled from a single structure. Application of the program is illustrated with a case study. The effect of a large shallow dogleg on accumulated fatigue in the drillpipe is discussed, enabling informed decisions on drilling subsequent hole sections. Introduction Fatigue damage is irreversible and cumulative, and it is caused by cyclic loads. Fatigue damage can occur at much lower stresses than the material yield strength used in designing the static limits of the drillstring. Fatigue accumulation propagates incipient microscopic cracks in drill stem components until they manifest as macroscopic cracks. If undetected by inspection techniques, these cracks can reach a critical dimension and cause catastrophic failure. Assessment and prediction of fatigue damage in drillpipe has been a persistent problem since the 1940s1. Designing drillpipe to withstand fatigue is significantly different from other aspects of drillpipe structural design. Fatigue is inherently scatter dependent, and the designer ideally requires significant amounts of data to assess fatigue damage or design a component to withstand fatigue. However, such data, especially for drillpipe and drill stem components, is limited. Due to the significant scatter that characterizes high cycle fatigue data, the design cannot be based on strict deterministic principles. The inherent randomness of fatigue data must be used with available fatigue models to estimate conservative limits within which there is a high probability of avoiding fatigue damage. The assessment of fatigue accumulation in drillpipe is a classical problem in mechanics of materials. A study of oilfield literature reveals two basic approaches to fatigue assessment: the classical cumulative fatigue damage (CFD) approach and the more recent linear elastic fracture mechanics (LEFM) approach. Three major sources of drillpipe fatigue data appear in oilfield literature: Bachman's data set published in 19512, Morgan and Roblin's data set published in 19693 and Grondin and Kulak's data set published in 19914,5. These data sets are the basis of most CFD based models and studies in the drilling industry. The current guidelines in API RP7G6 date back to the 1964 work of Hansford and Lubinski7,8 and are based on the Bachman data. Like the CFD approaches, studies based on the LEFM approach are characterized by limited data on crack propagation and fracture toughness, and field validation9. Besides these major data sets, a few references to fatigue tests conducted in the late 1950s for Project Mohole can be found10,11. This project, which was an attempt to drill in 12,000 ft. of water, required an understanding of dynamic stresses in the drillstring due to wave and current loads. Although it is evident from the project report that drillpipe fatigue tests were conducted, precise details are not available12.