磁/惯性组合的钻具重力信息自适应提取

针对旋转导向钻井过程中底部钻具振动引起的姿态解算参数严重失真问题，提出一种磁/惯性组合的钻具重力信息自适应提取方法。建立钻具重力信息的联邦融合提取框架，选取基于陀螺数据的重力递推模型作为主滤波器，基于加速度计、磁强计数据观测的重力信息模型分别作为子滤波器；对子滤波器状态和观测模型进行抗差估计融合，通过自适应因子和等价权矩阵自适应调节状态、观测模型融合权值，得到重力信息局部抗差估计；对抗差融合的子滤波器进行联邦融合，进一步得到钻具重力信息的全局估计。设计了模拟钻采及实钻井实验，实验结果表明，提出的磁惯性组合钻具重力提取方法能够有效提取钻具动态重力信息，重力提取后结解算的井斜角误差控制在2.5°以内。

Adaptive extraction of gravity information for drilling tool with magnetic and inertial integrated system

In order to solve the problem of the serious distortion of the attitude parameters about drilling tool caused by the strong vibration of the bottom drilling tool in the process of rotary steering drilling,a magnetic/inertial-based method is proposed to extract the gravity information of the drilling tool.Firstly,a federal extraction framework for the gravity information of drilling tool is designed with gyro data-based gravity recursive model selected as the main filter,and also the accelerometer data-based and magnetometer data-based gravity observation model are used as sub filter respectively.Secondly the robust estimation and fusion of the state and observation model about sub filters are carried out respectively,and the fusion weight between state and observation model is adaptively adjusted by the adaptive factor and the equivalent weight matrix,then the local robust estimation of gravity information is obtained.Thirdly the comprehensive estimation of gravity information about drilling tool is further improved by means of federal filter.Coal mine drilling experiments indicate that the proposed method could effectively extract the dynamic gravity information of drilling tools,with the corresponding average error of inclination within 2.5 degrees.