Search for a command to run...
Abstract Dogleg severity (DLS), a crucial measure of the drilling process, refers to the curvature of the borehole. While typically a 100 ft span is used for the DLS calculation, field experience suggests that this span does not accurately reflect the load on drilling tools. In contrast, local DLS over shorter spans typically has a higher correlation with mechanical overloads and fatigue. However, reducing the span length with the intention of a finer resolution only increases DLS because of the mathematics behind the DLS calculation. Thus, the optimal span for calculating local DLS remains unquantified. This paper aims to determine the optimal span using statistical methods and operational field data. To identify a universally applicable span for the local DLS calculation, drilling operation data in the form of continuous survey data from various historical runs with different bottom-hole assembly (BHA) sizes and configurations are used. In a first step, the bending moment, as the main mechanical load imposed by the borehole curvature, is simulated at every position of the survey throughout the BHA using a finite element software. Utilizing the same continuous survey data, the local DLS is calculated for multiple spans for each run. To identify the span that best represents the bending moment on the BHA, a correlation study is conducted between the simulated bending moment and the calculated local DLS. In this process, the span is adjusted incrementally, and the correlation is calculated at different positions of the BHA. An overall trend for the correlation can be observed throughout the various historical runs, regardless of tool size or drilling conditions. The correlation between bending moment and DLS is typically relatively low for smaller spans (1 ft to 5 ft). With larger spans, the correlation increases, peaking around 30 ft with correlation factors between 0.8 and 0.95 at the position of the survey sensor. These high correlation factors indicate an almost linear relation between bending moment and DLS. Beyond 30 feet, the correlation decreases to lower values. This overall behavior is plausible, as local DLS calculated at very short spans tends to be high, because every small change in curvature is reflected. However, the BHA, due to its stiffness, will not bend into these small curvatures. Higher spans do not tend to correlate well because smaller curvatures in the borehole are neglected and averaged out, leading to an underestimation of the bending load. The local DLS calculated with the statistically identified optimal span of 30 ft provides relevant information about the mechanical state of the BHA. Using solely continuous survey data, a meaningful estimation of the bending load can be derived. This can also be utilized in real-time applications to mitigate high loads or identify root causes in post-run analyses.