振动下惯性随钻测量的方位漂移闭环补偿

Closed-Loop Compensation for Azimuth Drift of Inertial Measurement While Drilling Under Vibration

为解决振动下惯性随钻测量(Measurement While Drilling,MWD)方位角失真的问题,提出一种方位漂移在线闭环补偿方法。考虑陀螺仪随时间漂移而加速度计长时测量稳定,利用多传感器分布式信息融合,获得微惯性测量单元(Micro Inertial Measurement Unit,MIMU)测量加速度和重力加速度在钻具坐标系下的向量误差,建立存在振动的方位陀螺漂移PI控制模型,设计一种基于函数逼近理论的新型神经网络结构,通过优化权值直接确定法(Optimized Weights Direct Determination,OWDD)快速确定网络权值,将神经网络辨识信息反馈给PI控制器,实现对方位陀螺漂移的自适应控制,实时跟踪误差趋向并补偿。最后设计实验,振动台实验中方位角误差由10.15°/h减小到0.21°/h,实际钻井实验中方位角误差由10.53°/h减小到0.3°/h。结果表明振动下方位角测量精度明显提高。

In order to solve the problem of azimuth distortion in Measurement While Drilling(MWD)under vibration,an online closed-loop compensation method for azimuth drift is proposed.Considering that the gyroscope drifters with time and the accelerometer is stable in long term Measurement,the vector errors of the Measurement acceleration and gravity acceleration in the drilling tool coordinate system of Micro Inertial Measurement Unit(MIMU)are obtained by using multi-sensor distributed information fusion.The PI control model of azimuth gyro drift with vibration is established,and a new neural network structure based on function approximation theory is designed.The weight of network is quickly determined by the Optimized Weights Direct Determination(OWDD).The neural network identification information is fed back to the PI controller to realize the adaptive control of the opposite position gyro drift,and the error trend is tracked and compensated in real time.Finally,the design experiment shows that the azimuth error in the shaking table experiment decreases from 10.15°/h to 0.21°/h,and the azimuth error in the actual drilling experiment decreases from 10.53°/h to 0.3°/h.The results show that the measurement accuracy of the azimuth Angle under vibration is significantly improved.