Multi-sensor Data Fusion Study Of Internal Defect Location In Pipelines

With the rising focus on pipeline safety operation and production, the importance of internal pipeline defect detection is increasingly prominent. Internal pipeline defect positioning, which provides the exact location information of internal defects and excavation, is an important part of internal pipeline defect detection. Safety accident risk can be reduced by timely and effective maintenance, keeping the pipeline safe. In the process of internal pipeline defect positioning, odometer wheel and strapdown inertial navigation system both have their advantages, but there are disadvantages when they are worked alone. Due to the positioning characteristics and the uncertainty of inside pipeline environment, the positioning of odometer wheel tends to contain gross error, which is not easy for divergence, by taking a slippage. While strapdown inertial navigation system, which does not rely on the external environment, is an independent positioning. But its positioning error increase rapidly with growth time. Based on the complementarity, odometer wheel-assisted strapdown inertial navigation system is proposed in this paper, combining the multi-sensor data fusion method for internal pipeline defect positioning.At first, multi-odometer wheel is used for internal pipeline defect positioning. According to the positioning error characteristics of odometer wheel, robust estimation algorithm and consistency check algorithm is used for measurement data fusion. Through contrastive analysis, robust estimation algorithm, which relieves the positioning effect of gross error, is chosen in this paper. Then the posture model and position model of the movement trajectory are simulated. The mathematical models of accelerometer and gyro, which are the device of strapdown inertial navigation system, are researched. Combined with the coordinate transformation matrix, the output information of accelerometer and gyro in the machine coordinate system is gained. After that, according to the calculating model of strapdown inertial navigation system, the position information of internal pipeline defect positioning is gained. Finally, taking advantage of the complementarity, odometer wheel-assisted strapdown inertial navigation system is used for internal pipeline defect positioning. Due to the main function of robust estimation algorithm is to eliminate the influence of gross error on the fusion result, indirect kalman filter is using for data fusion of odometer wheel and strapdown inertial navigation system. On the basis of error models, the positioning information fusion of odometer wheel with robust estimation algorithm and strapdown inertial navigation system is completed. According to the speed of odometer wheel cannot be directly get, the displacement increment difference between the odometer wheel and strapdown inertial navigation system is chosen as observed value.MATLAB is used for the model programming of robust estimation, movement trajectory, strapdown inertial navigation device model, strapdown inertial navigation calculation model, and data fusion model of odometer wheel-assisted strapdown inertial navigation system. The experiment result shows that the maximum longitude error for the internal pipeline defect positioning with odometer wheel-assisted strapdown inertial navigation system is 4×10-8. The maximum latitude error is 8×10-8 and the relative positioning error can reach 0.01%. It greatly improves the positioning accuracy and provides reliable positioning information for pipeline safety maintenance.