【作者】 刘爽；

【导师】 赵青；

【作者基本信息】 电子科技大学 ， 工程硕士（专业学位）， 2021， 硕士

【摘要】 井中雷达是一种高效的、无损的井中探测手段,能够工作在油田、气井等深层地层,因此得到了广泛的应用。井中雷达由于环境的复杂性以及探测条件的限制,在实际地质探测过程中井中雷达相较于常规的探地雷达,更难保证数据质量,因此如何从井中雷达数据中提取出有效信息,并且从中反映的地层信息,是目前国际竞相研究的重点。井中雷达系统尚未有统一的国际化标准,因此,井中雷达数据因其数据传输复杂、数据量传输速度有限、数据容易因为复杂的环境而受到污染等原因,如何从原始数据中提取出有效数据,并且对井中仪器进行实时控制及监控,是井中雷达系统数据处理所面临的首要问题。本文主要针对井中雷达信号处理方面展开,主要工作内容如下几个方面。首先,阐述了井中雷达信号的模型以及常用的井中雷达计算方法。从井中雷达信号着手,推导了井中雷达信号的信号模型以及井中雷达的数据类型,描述了井中雷达的信号分析内容,通过数学模型建立井中雷达数据处理的基础,编程实现了常规的数据处理流程。由麦克斯韦方程组出发,通过对计算流程的推导以及一维时域有限差分计算的推导,将正演计算井中雷达仿真描述完整,并将其计算方法拓展到高维,从而能够进行更加贴近真实模型的仿真计算。在此基础上,对于信号预处理方法,基于EMD分解的方法具有井中雷达信号处理适应性优良的表现,从EMD分解算法出发,对井中雷达信号预处理的时频域分解算法进行分析,通过将其拓展至高维,从而对井中雷达二维信号进行处理,能够有效保证信号质量。通过时频域的分解,使得信号处理工作有更好的时频域分析。本文还介绍了通过井中雷达能量的三维切片成像,能够直观可视化分析能量回波情况,极大提高了井中雷达分析地质资料的能力。最后,本文对井中雷达时域波形的反演进行了相关研究,对反演方程组进行了推导优化,进行了仿真编程验证,仿真结果表明,井中雷达基于最小二乘的反演计算方案能够达到较好的收敛结果,能够对柱状目标的电磁参数进行复现。通过将偏移成像及网格划分方法相结合,提出了对井中雷达目标的反演方案,对柱状目标的反演计算表明,该方法有效减小反演计算的工作量和复杂度,对模型的反演情况良好。更多还原

【Abstract】 As an efficient non-destructive detection method,borehole radar can work in deep formations such as oil fields and gas wells.Be constrained by the complexity of the environment and the limitation of the detection conditions,it is more difficult to guarantee the data quality than the conventional ground penetrating radar in practical engineering application.Therefore,how to extract the effective information from the borehole radar data and reflect the formation information is the focus of international research.There is no unified international standard for borehole radar system.Due to the complex data transmission,limited data transmission speed and data easily polluted by complex environment,how to extract effective data from the original data and carry out real-time control and monitoring of borehole instruments is the primary problem for borehole radar system data processing Questions.This paper mainly focuses on the signal processing of borehole radar,and the main contents are as follows.Firstly,the model of borehole radar signal and the common calculation method of borehole radar are described.Starting from the borehole radar signal,the signal model of borehole radar signal and the data type of borehole radar are deduced,the signal analysis content of borehole radar is described,the basis of borehole radar data processing is established through mathematical model,and the conventional data processing flow is realized by programming.Based on Maxwell equations,the forward calculation borehole radar simulation calculation is described completely through the derivation of calculation process and one-dimensional finite-difference time-domain calculation.Then the calculation method is extended to high-dimensional,so that the simulation calculation closer to the real model can be carried out.On this basis,in the methods of signal preprocessing,the method based on EMD decomposition has excellent performance of borehole radar signal processing adaptability.Starting from the EMD decomposition algorithm,the time-frequency domain decomposition algorithm of borehole radar signal preprocessing is analyzed.Expanding to high dimension,the two-dimensional signal of borehole radar is processed,which can effectively ensure the signal quality.Through the decomposition of time-frequency domain,the signal processing work has a better time-frequency analysis.This paper also introduces the 3D slice imaging of borehole radar energy,which can visually analyze the energy echo and greatly improve the ability of the borehole radar data to analyze geological data.Finally,this paper studies the inversion of borehole radar time-domain waveform,deduces and optimizes the inversion equations,and carries out simulation programming verification.The simulation results show that the inversion calculation scheme of borehole radar based on least square can achieve good convergence results,and can reproduce the electromagnetic parameters of cylindrical targets.Through the combination of migration imaging and grid generation method,a response scheme for the well radar target is proposed.The inversion calculation of cylindrical target shows that this method can effectively reduce the workload and complexity of inversion calculation,and the inversion of the model is good.更多还原