An Adaptive Sampling Method for In-Line Inspection of Pipelines Based on Multi-Sensor Fusion and SSI-MPC

In long-distance subsea pipelines, pipeline inspection gauge (PIG) speed fluctuations under complex operating conditions can cause a mismatch between the sampling frequency and the operating state. This mismatch may compromise the integrity of defect detection signals. To address this problem, this study proposes an adaptive sampling method for in-line inspection of pipeline data based on multi-sensor fusion and a spatial sampling interval model predictive controller (SSI-MPC). The impact of PIG velocity fluctuations on defect detection signal integrity was analyzed, and a multi-sensor velocity fusion estimation model based on Kalman filtering was established to fuse inertial measurement unit (IMU) acceleration data with cumulative displacement information from the odometer. The sampling frequency adjustment was formulated as a spatial sampling interval tracking problem, and an SSI-MPC controller was developed. The effectiveness of the proposed method was verified through comparative simulations with fixed-time sampling and odometer-based equidistant sampling. The results indicate that SSI-MPC adaptive sampling more effectively restores the key quantitative features of defect detection signals under velocity-jump and high-speed conditions. The RSE and RAE of the defect detection signals are controlled within 3.44% and 2.10%, respectively. Compared with fixed-time sampling, the proposed method reduces the detection signal data volume by 39.37%.