Measurement of Stresses in Pipelines Using Hole Drilling Rosettes

Abstract

On-shore buried pipelines are loaded by internal pressure, by longitudinal displacement restrictions caused by support or soil interaction, and by temperature differential. If a referential initial state of stress can be established by some means, conventional strain gage techniques can be used to determine the existing loads relative to the reference state. If a reference cannot be established, residual stress measurement techniques are strong candidates for use in the measurements because they measure the absolute state of stress present at the assessed point of the pipeline. This paper aims to provide laboratory test results to show how uncertain measurements using conventional strain gage or residual stress techniques can be when used in determining existing loads. The tests utilized a special device designed and built for applying combined states of stress. The device had a U-shaped format. The horizontal part of the U-shaped device consisted of the specimen to be tested: a segment of an API 5L X60 steel pipe with length, diameter and thickness equal to 910mm, 323.3mm and 9.7mm, respectively. The pipe segment was capped at both ends by welding two reinforced plates that formed the two vertical legs or arms of the U-shaped device. A pressurized water inlet was introduced into one of the caps to provide for the internal pressure loading of the test specimen. Besides capping the pipe, the vertical leg-plates were connected by two threaded spindles. The spindles and the long plate arms made it possible to apply axial forces and intrinsically coupled bending moments to the pipe segment. The spindles were instrumented with four electrical resistance strain gages in a full bridge arrangement to measure the applied forces. The resulting applied stress states in the pipe walls caused by normal force, bending moment and internal pressure were measured using several rosettes of electrical resistance strain gages that were appropriate for applying the blind-holedrilling technique. Three sets of data points were collected during the tests. The first set consisted of measurements of strains caused by applying simple or complex load combinations. The second set consisted of strain measurements made after the blind holes were drilled in order to determine the residual stresses caused in the fabrication of the steel pipe. The third set of data consisted of measurements of strain variations caused by unloading the device. All the data of the experiment were analyzed and compared with strains calculated by means of analytic (Strength of Materials) and numeric (Finite Elements) methods. These comparisons helped to reach the conclusion that the use of a residual stress measurement technique such as the blind-hole drilling method to determine pressure induced and soil movement loading in operating pipeline will furnish inaccurate results even if a reasonable number of measurement points are used to describe the stress states of points along the analyzed cross sections. The reason for this is the possibility of a large variation in the residual fabrication stresses not only along the outer perimeter of one cross section of the pipe, but also the large variation in the distributions that occur along neighboring sections of a short pipe segment. For the case of predicting strains generated by the complex loading applied to the pipe test specimen, the comparison of experimentally and analytically determined strains allowed for the evaluation of an overall strain prediction uncertainty of 50µe. Furthermore, an uncertainty value of 26µe for the hole drilling method was also determined by comparing the initially imposed combined strain states with those measured after drilling the blind holes and unloading the test specimen.