旋转导向系统有色噪声的改进无迹卡尔曼滤波方法

在旋转导向钻井系统的姿态测量过程中,三轴加速度计测量数据中包含大量有色噪声,严重影响井底组合钻具姿态测量的精度。基于旋转坐标变换的四元数理论,结合所建立的三轴加速度噪声模型,提出一种改进无迹卡尔曼(UKF)迭代滤波算法。此方法利用陀螺测量原理构造观测方程和时变状态方程,将实时解算出的钻具姿态以四元数的形式更新时变状态方程中三轴角速度,通过更新观测方程加速度噪声模型实现加速度计传感器数据中有色噪声的UKF迭代滤波。实测数据滤波的结果表明,此方法可有效滤除加速度传感器数据中的有色噪声,保证旋转导向钻具姿态测量的精度。

Improved unscented Kalman filtering method for colored noises of rotary steerable system

In general, the measurements of three-axial accelerometers contain a great amount of colored noises, which can significantly affect the accuracy of the attitude measurement by the rotary steerable drilling system. Combined with the noise model of the three-axial accelerometer, an improved unscented Kalman filtering(UKF)algorithm based on quaternion was proposed. Based on the theory of gyroscope, the observation equation and time-varying state equation were constructed in real-time, and then the three-axis angular velocities of time-varying state equation were updated by the attitude of drilling assembly in the way of quaternion, and the UKF iterative filtering was applied. The filtering results measured from oil fields show that this method can filter the colored noises of accelerometer effectively and ensure the accuracy in the attitude measurement of rotary steerable assembly.