Abstract
The friction factor plays a critical role in drag and torque calculations. In the drag and torque model, the friction factor is typically treated as an average value, or the wellbore is roughly divided into casedhole and openhole segments with different values assigned. Such simplified model parameters fail to account for variations in the friction factor across different well segments, impacting the accuracy of the computations. Precise drag and torque calculations are significant, and the prerequisite for accurate drag prediction is the inverse analysis of realistic friction factor parameters. This study proposes a friction factor inverse analysis method based on morphological characteristics of measured data. The Fisher optimal segmentation method is employed to divide the field-measured data into multiple reasonable subsegments based on curve morphology features, indirectly determining the segmentation of the friction factor. The 3D stiff-string drag and torque model is used to establish friction factors inverse model, with corresponding friction factors introduced separately for the segmented well sections. The sparrow search algorithm, enhanced by opposition-based learning strategy, is utilized to optimize the segmented friction factors, with the optimization evaluation criteria simultaneously assessing the deviation and similarity between simulated and actual curves. Case studies demonstrate that the friction factor inverse analysis method proposed in this study, leveraging the curve morphology features of real-time drilling data, can efficiently obtain high-precision segmented friction coefficients. This method significantly reduces the error between simulated and field-measured data, improving calculation accuracy by about 50%, thus providing robust data support for accurate drag and torque prediction. Furthermore, the division approach can be applied during a round trip to optimize drilling parameters, enhancing the efficiency of drilling parameter control.