In-line-inspection (ILI) inertial measurement unit (IMU) data have been used to identify sections of pipelines bent or deflected by an outside force (e.g., a geohazard) for over 25 years, and particularly since the recent widespread adoption of the use of the IMU tool to more accurately map pipeline centerlines. In principle, the concept is simple: the ILI vendor reviews the IMU data and excludes construction related bends (e.g., sag bends, over bends, side bends). Bends that extend across one or more girth welds are inferred not to have been deliberately constructed and are reported by the vendor as “bending strain features,” along with the magnitude of the bending strain and other attributes. Bending strain features exceeding a specified threshold magnitude are inferred to have formed post-construction (e.g., from an outside force such as a geohazard). As such, commissioning a bending strain report can be an attractive option to operators because the usage of an IMU has increasingly become a standard part of ILI data collections; thus, the additional cost of the bending strain report is low.

In practice, the use of IMU bending strain data is not nearly so simple. As multiple operators have experienced, the list of bending strain features provided by the ILI vendors often does not correlate to any apparent anomaly when excavated, nor any apparent geohazard. Absent context and supporting information, IMU bending strain data can appear to be essentially meaningless with little value added to real-World situations.

The goal of this paper is to provide context and examples for how IMU bending strain data have been used to successfully identify real-World impacts to pipelines from geohazards. The authors have access to a database containing over 4,600 IMU-identified bending features located in a variety of geographic settings, including areas with high densities of landslides. As described in the paper, the majority of reported bending strain features (over 90%) do not appear to correlate to a likely geohazard. Instead, most reported bending strain features are relatively low-magnitude (less than 0.2% strain) and are inferred to correlate with likely tie-in locations, such as at pipe replacement locations, road crossings, and river crossings, where low levels of bending may have been induced during construction. Although the number of bending strain features that correspond with a likely geohazard is a small percentage of the total, the authors have found statistical correlation between likely geohazard locations and relatively high magnitude bending strain features, as well as a correlation between geohazards and bending strain where the primary or significant component of strain is in the horizontal direction.

If IMU bending strain data can be put into context (e.g., overlain with known geohazard locations), the results can be used to aid in the prioritization of response, such as monitoring or mitigation of sites. Alternatively, through a close review of the IMU bending strain data and their locations, previously unidentified geohazards may be identified (e.g., the IMU data can be used to prioritize where to look for possible geohazards).