Bolted junctions are ubiquitous in applications such as oil and gas pipelines, and making sure such fasteners remain adequately tightened is critical to ensuring a leak-proof seal at pipe flanges. Manual checking with traditional tools such as torque wrenches can be time-consuming and expensive, while conventional ultrasound methods essentially rely on a calibration having been performed for each bolt, which is often not practical for field use.
In this project Acentech was asked to use its expertise in signal processing, vibration, wave propagation and ultrasonics to help a manufacturer of bolting equipment develop a device to infer bolt tension in the field, without the need for a prior calibration step. Several possible methods were researched, including flange structural response across the joint, bolt bending response and various ultrasonic techniques before settling on a technique that monitors the roundtrip travel time ratio of axial to shear waves in the bolt. The travel times for these two wave types are affected not only by slight changes in the bolt length as it is tensioned, but also by a phenomenon known as the acousto-elastic effect, which changes the speed of the waves as the bolt tension changes. Since the travel times for the two wave types respond differently to changes in these two factors, their ratio (along with some geometric and material properties) can be exploited to directly determine bolt tension without the need for a prior calibration measurement, making this method a good candidate for practical use in the field. An ultrasonic transducer capable of launching and detecting both axial and shear waves was chosen, and an adaptor was designed so that the transducer could be coupled to a stud end in a quick and repeatable fashion. Signal processing techniques were developed to accurately detect the arrival times of waves reflected from the other end of the stud, from which the travel time ratio and the tension in the stud could be determined.
The main finding was that this method could be used to measure bolt tension to within an accuracy of about +/-5% (compared to the directly measured tension obtained from a load cell) as long as the stud ends were in good condition and roughly parallel to each other. A subsequent study was conducted in which the system was used on a large sample of bolts from the field, which led to the finding that a fairly high proportion of such bolts had end conditions that degraded the accuracy of the method.
The results and recommendations obtained from this project enabled the client to make informed decisions about further development of the system, and to refine the target market to those applications were bolt end conditions were more controlled (such as the nuclear industry).