Comparing the Results of a Full-Scale Buckling-Test Program to Actual Well Data: New Semiempirical Buckling Model and Methods of Reducing Buckling Effects

Buckling and its effects are topics of economic and technical interest as extended reach drilling (ERD) and horizontal wells become critical to maximizing recoverable reserves, particularly in the continental United States and Alaska. Previous work has resulted in important discoveries about drillstring buckling, but to date, little testing has been performed on actual drillpipe in a controlled manner, particularly in measuring drillstring whirl. As a result, there can be disparities between theoretical predictions of buckling effects vs. actual field results. These disparities can result in unrealistically high friction factors required to bring calculated values close to actual data, or in many cases, operational difficulties can result such as high torque, low rate of penetration (ROP), drillstring failures, inability to maintain directional control, or inability to reach the planned depth. To learn more about drillstring behavior in buckling conditions, a full-scale buckling test fixture was developed to evaluate the effects of buckling on 2 7/8-, 3 1/2-, and 4-in. drillpipe while sliding and rotating inside 7-in. casing. The test fixture incorporated a variety of sensors and cameras to characterize torque, drag, vibration, and drillstring deformation under buckling loads. As part of the test program, low-friction nonrotating protectors were also tested to measure performance under buckling conditions. The test results show that drillstring buckling occurs at far lower loads than predicted by current models, possibly caused by minor deformations inherent in real drillpipe. The results also show that for a given amount of torque or drag, protectors increased the available compressive load by 20 to 30% and substantially reduced vibration caused by drillstring whirl. The test results were used to develop a new semiempirical buckling model that predicts contact forces resulting from drilldrag modeling software in which it was compared against actual data from a large number and variety of wells. The results show an ability to more accurately predict torque, drag, and vibration caused by buckling and whirl.