管道地理坐标测量系统的误差校正方法研究

发布时间：2018-01-05 08:12

本文关键词：管道地理坐标测量系统的误差校正方法研究　出处：《沈阳工业大学》2015年博士论文　论文类型：学位论文

【摘要】：常态化长输油气管道内检测是保障管道安全运行的重要措施。管道地理坐标测量系统在常态化管道清洁和缺陷内检测的同时完成管道位置测量和管道地理轨迹的描绘以减少管道缺陷开挖维护成本,监测和评估管道应力、应变的变化,对提高管道维修效率、辅助管道安全预测、完善管道运维信息具有重要意义。管道地理坐标测量系统的核心是捷联惯性导航系统,该系统的最大特点是硬件结构简单,软件极为复杂多样。由于惯性敏感元件、初始值及安装等误差的存在,导致长航时复杂导航计算过程中位置等导航参数误差随时间不断累积甚至发散,这些故有的问题成为管道地理坐标测量系统工程应用的瓶颈。如何消除由于各种因素导致的导航参数误差(特别是位置误差)是管道地理坐标测量系统工程应用的关键。因此,论文基于管道地理坐标测量系统平台,围绕几种误差估计和补偿展开相关研究,主要工作有:(1)论文首先设计了一种捷联惯性导航系统(SINS—Strapdown Inertial Navigation System)导航解算程序。采用4阶龙格库塔法求解四元数完成姿态更新,提高了数值解算方法精度和速度。分析了SINS导航系统的噪声特性,针对管道焊缝、颠簸等运动引起的冲激噪声,采用均值法去除;针对MEMS敏感元件的随机噪声影响,建立了基于自回归模型AR(1)的随机噪声模型,采用基于AR(1)的kalman滤波(KF—Kalman Filter)方法去除。小车导航参数解算的仰俯角、横滚角及高度对比实验结果表明采用先对惯性器件采样信号进行均值法预处理后,再利用基于AR(1)的KF法对随机误差进行动态补偿的数据预处理效果优于单独采用均值法及平滑法的去噪效果。(2)SINS在管道内检测器上的安装误差是影响管道地理坐标测量精度的重要因素之一。论文研究了安装误差对导航解算影响的规律,定义了管道内检测器与SINS坐标轴间的偏角为安装误差,依据刚体转动原理,建立了安装误差动态传递模型;将俯仰和航向安装误差角作为状态变量,基于管道铺设起始段的平直特点,构建一种基于管道初始段方向约束和里程轮速度约束为观测量的SINS/里程仪/管道磁标点GPS组合导航系统,并采用非线性容积卡尔曼滤波(CKF—Cubature Kalman Filter)对俯仰角和航向角安装误差的估计和补偿。模拟管道内有滚转运动的三组拖车实验结果表明所提出方法能够有效地在线估计出安装误差,安装误差估计精度为0.2′~0.6′,并能够抑制和改善因俯仰角安装误差引起的高度解算误差及航向角安装误差引起的轨迹偏移,经安装误差修正后的水平地理坐标解算精度可达到1.07e-3°;高度解算精度可达1.59e-3m。(3)初始对准提供管道地理坐标测量系统初始姿态矩阵,其精度是影响管道地理坐标测量精度的重要因素。论文针对初始对准问题,提出了一种新的管道惯性测量系统动态初始对准算法。建立了以管道基准点为中心的100米区段的管道方向角、基准点位置及管道检测器运行速度为观测量的动态初始对准观测方程。滤波采用模型误差预测滤波(MEP—Model Error Predict)和CKF相结合算法,利用MEP算法预测系统模型误差以修正CKF状态一步预测值。该算法能够实时预测模型误差,降低了CKF系统状态维数,减少估计时间。小车实验结果表明:本文提出的动态初始对准算法适用于运行环境复杂,模型误差不确定,非线性强的系统,对准精度可达0.001°,能够满足管道地理坐标测量系统在管道内复杂运行环境下的动态初始对准。(4)封闭油气管道没有实时GPS信息用来修正导航参数误差,只有里程轮提供的管道地理坐标测量系统在管道内的运行速度和行程信息。常规SINS/GPS或SINS/GPS/里程轮组合导航方法不适用于管道地理坐标测量系统。根据管道地理坐标测量系统的结构及其在管内运行特征,设计了两级滤波误差校正方法。利用前级正向基于里程轮速度自约束与后级反向基准点位置平滑两级组合滤波算法实现管道位置的最优估计。因基准点位置信息在管道磁标点和出口处可获得,在由起点至磁标记点的运行过程中,导航误差由前级CKF和KF组合滤波校正,即强非线性部分采用CKF滤波,避免雅克比计算及线性化误差,线性部分采用KF,降低CKF滤波维数。该算法将前级正向滤波补偿后输出的位置信息为后级反向平滑滤波的参考量,将两级滤波有机地结合在一起。后级反向滤波以管道磁标记或末端基准点位置信息为观测初始值,进行后级平滑滤波,通过基准点校正进一步修正了前级位置参量误差。通过拖车实验轨迹的固定测量点值与GPS测量值比对证明该管道地理坐标解算算法的地理坐标测量精度可以达到10e-3°,能够满足长距离油气管道内检测与定位的需求。
[Abstract]:The normalization of long-distance oil and gas pipeline detection is an important measure to ensure the safe operation of the pipeline. The pipeline geographic coordinate measuring system in normal and clean pipeline defect detection simultaneously depicting geographic trajectory position measurement and pipeline pipeline to reduce the defects of the pipeline excavation and maintenance costs, monitoring and assessment of pipeline stress and strain changes, to improve the efficiency of pipeline maintenance, auxiliary pipeline safety prediction, it is important to improve the pipeline operation and maintenance information. The core pipeline geographic coordinate measuring system is a strapdown inertial navigation system, the biggest feature of the system is simple structure of hardware and software is very complicated. Because of the inertia sensitive element, initial value and installation errors, resulting in long endurance complex the navigation location computing process navigation error accumulation and divergence over time, so some of these problems become the pipeline geographic coordinate measurement Bottleneck of system engineering application. How to eliminate the error of navigation parameters caused by various factors (especially position error) is the key application of the measurement system of engineering pipeline geographic coordinate. Therefore, the pipeline geographic coordinate measuring system based on the platform, around several error estimation and compensation research, the main work is: (1) firstly, design a strapdown inertial navigation system (SINS - Strapdown Inertial Navigation System) navigation solution program. By using 4 order Runge Kutta method for solving four quaternion attitude updating, improves the numerical solution method of precision and speed are analyzed. The noise characteristics of SINS navigation system, aiming at the pipeline weld, bumpy motion caused by impulse excited noise removal using average method; according to the random noise of MEMS sensitive components, establish autoregressive model based on AR (1) random noise model, which is based on AR (1) Ka LMAN filter (KF Kalman Filter) method to remove. Car navigation parameters calculation of pitch angle, roll angle and height of comparative experimental results show that the first of the inertial device sampling signal pretreatment means, then based on the AR (1) KF method of dynamic compensation for random error data pretreatment effect better than the average value method and smoothing denoising effect. (2) SINS in the pipeline installation error detector is one of the important factors affecting the accuracy of pipeline geographic coordinate measurement. The paper studies the installation error of the navigation solution of law, defines the angle detector and the SINS axis between the pipe for installation based on the principle of rigid body rotation error, installation error of the dynamic transfer model was established; the pitch and heading misalignment angles as state variables, based on the characteristics of the initial segment of the pipeline laying flat, which is based on the pipeline at the beginning of Gou Jianyi The initial segment direction constraint and mileage wheel speed constraint is SINS/ / odometer measurements of pipeline magnetic punctuation GPS integrated navigation system, and the nonlinear volume filter (CKF - Cubature Kalman Calman Filter) estimation and compensation of pitch angle and yaw angle error. Simulation installation pipe rolling motion of three groups of experimental results show that the trailer the proposed method can effectively estimate the installation error, installation error estimation accuracy is 0.2 '~0.6', and can restrain and improve due to pitch angle error caused by the installation height calculation error and the heading angle deviation caused by installation error, the installation level of geographic coordinates error calculation accuracy can reach 1.07e-3 high degrees; calculation accuracy of up to 1.59e-3m. (3) initial alignment provides pipeline geographic coordinate measuring system of initial attitude matrix, its precision is the pipeline geographic coordinate measurement accuracy effect Factors. The thesis focuses on the initial alignment problem, proposed a new pipeline dynamic initial alignment algorithm of inertial measurement system. Establish the pipeline pipeline with reference point as the center of the 100 meter direction angle, reference point position and speed of pipeline detector for dynamic measurement of the initial alignment error model filtering using the observation equation. Predictive filtering (MEP - Model Error Predict) combined with CKF algorithm, using MEP algorithm to predict the model errors by using the modified CKF one-step prediction value. The algorithm can reduce the error of real-time prediction model, CKF system state dimension, the estimation time is reduced. The experimental results show that the vehicle dynamic initial alignment algorithm is proposed in this paper in complex environment, model error uncertainty, nonlinear system, the alignment accuracy can reach 0.001 degrees, which can meet the pipeline geographic coordinate measuring system in the pipeline complex operation Dynamic initial alignment environment. (4) closed oil and gas pipeline no real-time GPS information is used to correct the error of navigation parameters, only provide the mileage wheel pipeline geographic coordinate measuring system running speed and travel information in the pipeline. The conventional SINS/GPS or SINS/GPS/ mileage wheel navigation method is not suitable for pipeline geographic coordinate measurement system according to the structure. Pipeline geographic coordinate measuring system and its operating characteristics in the tube, correction method for the design of two levels of filtering error. Using the positive mileage wheel speed based on self constraint and reverse the basic point of two smooth combination filtering algorithm to realize optimal pipeline location estimation. For location information available in the pipeline and magnetic punctuation exit, in the operation process by the starting point to the magnetic mark point, navigation error by former CKF and KF combination filter correction, strong nonlinear part by CKF filter, Avoid Jacobian calculation and linearization error, the linear part of the KF, CKF filter. This algorithm reduces the dimension of the location information before the forward filtering output after compensation for the reference level after reverse smoothing filter, two stage filter in combination. After reverse filtering to pipe magnetic marker or end point of reference the location information for the observation of the initial value, after the level of smoothing filter, through reference point correction before further amendments to the level of location parameter error. Comparison proved the geographical coordinates of the pipeline calculation precision of geographical coordinates measurement algorithm can reach 10e-3 degrees by GPS fixed value and measure the value of the trace the trailer, which can meet the detection and location of long distance the oil and gas pipeline needs.

【学位授予单位】：沈阳工业大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TE973.6

【相似文献】