Improvement of Drill-Collar Fatigue Property by Last-Engaged -Thread Height Reduction of Pin
- Yasushi Tsukano (Nippon Steel Corp.) | Shunji Nishi (Nippon Steel Corp.) | Shin-Ichi Nishida (Nippon Steel Corp.) | Masakatsu Ueno (Nippon Steel Corp.)
- Document ID
- Society of Petroleum Engineers
- SPE Drilling Engineering
- Publication Date
- December 1990
- Document Type
- Journal Paper
- 325 - 330
- 1990. Society of Petroleum Engineers
- 1.6 Drilling Operations, 1.10.1 Drill string components and drilling tools (tubulars, jars, subs, stabilisers, reamers, etc), 1.10 Drilling Equipment
- 0 in the last 30 days
- 199 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
Improvement of Drill-Collar Fatigue Property by Last-Engaged -Thread Property by Last-Engaged -Thread Height Reduction of Pin
Summary. This paper describes a method to increase the fatigue life of a drill collar by reducing the height of a pins's last-engaged thread (LET). The effectiveness of the method was verified with finite-element analysis (FEA), and the results were evaluated by full-siz fatigue tests. Also, the practically of the redesigned connection was confirmed through overmakeup, make and break and full-size tensile testing.
Drill-collar fatigue failure is a costly problem for the oil industry This type of failure often occurs at a pin's LET because of the concentration of thread loading in this area. Thus, the best way to increase fatigue resistance is to reduce load concentration.
Although some papers I have dealt with redesigning threads to reduce load concentration, most redesigned connections restrict in terchangeability with API standard connections. A redesigned con nection should be interchangeable with any standard connection.
LET-height reduction is a simple method that reduces the load concentration while preserving interchangeability with any standard connection. This paper describes the effectiveness of the method against fatigue.
Maduschka developed a numerical method to obtain the thread-load distribution in a made-up, threaded connection. With this method, Nishida obtained thread distributions for several as of connections. According to Nishida's results, about 30% of the total thread load was carried by the pin's LET, which is why drill-collar fatigue failures often occur there.
Figs. 1a and 1b show the thread contact in the standard and the redesigned connections, respectively. In each connection, the box thread is assumed to be a cantilever beam and the contact load a point load. If the same deflection is assumed at both free ends, the point load. If the same deflection is assumed at both free ends, the beam of Fig lb can be deflected with a lower point load because its load is applied farther from the fixed end than that of Fig. la.
Although whether the deflections obtained at both free ends are the same under identical makeup conditions is unknown, LET reduction may still effectively reduce the contact load.
Finite-Element Model. To prove the effectiveness of thread-height reduction we conducted an FEA with the MARC computer program. program. Although the previous section discussed the reduction of LET height only, the same load-reduction technique was applied to subsequent threads to study the effect of gradual reduction in thread height across several threads.
We modeled a 5 14-in.-OD X 2 13/16-in. (13.3 cm OD X7.1 cm ID) drill collar with an API NC40 connection (see Fig. 2). Model A shows the API standard connection, and Model B shows the LET-height-reduced connection- (The LET height was reduced to be 70% lower than the standard height.) Model C shows a connection on which the LET and the second-thread heights were reduced. Model C was prepared to investigate the effect of gradual reduction. AH three models used four-node axisymmetric quadrilateral elements with bending effects. The pin side of each model was fixed against both axial and radial displacement.
The contact conditions between the pin and the box were given by the interface element, which can be used with the bending element. The interface element requires the coordinates of two contacting nodes, the contact on, and the interference. These dam determine the accuracy of the analysis. Through several examinations, we selected the data for the interface elements (see Fig. 3). These data indicate the contact conditions of the connection nude up at API recommended torque (12,100 lbf-ft [16 405 N-m]).
Our analysis neglected friction between threads because of the difficulty of convergence.
An elastoplastic element method was used because the stresses at the thread roots could have exceeded the yield situ. Table 1 lists the mechanical properties used in this analysis.
The analysis was composed of three loading steps. First, a makeup load determined with data from Fig. 3 was applied to the models. Second, a bending moment of 23,600 lbf-ft [32 000 N - m] was uniformly applied to the made-up connections. Third, the bendingmoment direction was reversed. This load step was required to investigate the mean stress and the stress amplitude at the @ roots. Although the stresses in the compression side can be obtained in the second step (without the reversed moment), the loading path is not equal to that in the third step. From a viewpoint of pathdependent plasticity, the reversed moment was needed. plasticity, the reversed moment was needed. Model Accuracy. To confirm the accuracy of the models, we compared stress-analysis results with strain-gauge-test results. Fig. 4 shows the comparisons for the API recommended makeup torque (the first load step) of the standard connection. The agreement between analytical and experimental results is fairly good.
Solutions. Fig. 5 shows the comparisons of thread-load distributions among Models A, B, and C under the makeup load (first step). The values of these loads do not include the radial components of thread loading because the components are most often imed with drill-collar fatigue failures.
As Fig. 5 shows, the thread load in Model A is concentrated on the LET. The load carried by the LET is 58,780 lbf [261 464 N]. The load concentrations shown for Models B and C are about 22 % lower than that of Model A.
The effects of gradual reduction in thread height are not clearly shown in Fig. 5. Little difference in thread load on the LET exists between Models B and C, even though the load on the second thread of Model C was reduced by height reduction.
Figs. 6a and 6b show the axial stress contours in the LET area of Models A and B, respectively, under the makeup load (first step). Although the LET roots are plastic-deformed on both models, the maximum stress of Model B is lower than that of Model A.
FW. 7a and 7b show the threat-root along the pin lengths of Models A and B, respectively. The stresses for the three loading steps are given so that the mean stress and the stress amplitude at the thread roots can be compared. The mean stress at the LET of Model A is 232,760 psi [1598 MPa], while the mean stress of
Model B is 212,760 psi [ 1467 MPaj. The reduction is about 8 %. The stress amplitude for Model A is 57,580 psi [397 MPa], while amplitude for Model B is 55,112 psi [380 MPa], a 4 % reduction.
|File Size||923 KB||Number of Pages||6|