【作者】 涂兵；

【导师】 李德胜；

【作者基本信息】 北京工业大学 ， 机械工程， 2013， 博士

【摘要】 当前石油钻井施工过程中，为了准确测量井眼参数、提高井眼轨迹控制力度、使钻头精确命中目标油层，钻井工艺中普遍采用了无线随钻测量(MWD，Measurement While Drilling)技术。无线随钻测量技术是在钻井过程中对井下信息进行实时测量和传输技术的简称：其特点是在不中断钻头正常钻进的情况下获得钻头附近的地质信息，并将这些信息以无线信号的方式传输到地面；地面系统再对这些信号进行分析与处理，按照井下仪器匹配的编码方式进行译码，获取施工所需要的定向数据、地层特性和钻井参数等各种信息。本文以提高无线随钻测量中泥浆脉冲信号的传输速率和解码的正确率为目的。因为泥浆脉冲信号在钻柱内的传输会受到现场各种条件的影响，所以安装在井口处的压力传感器检测到的泥浆脉冲信号的压力波中含有大量的噪声信号，这些噪声信号主要是由于井下各种机械的转动或者震动引起的，它们淹没了原始有用的脉冲信号；即使是没有经过噪声干扰的标准的泥浆脉冲波形，在经过长距离的传输或者由于泥浆本身的质量原因也会导致有用信号的大幅衰减。因此，如何从频率不固定的强噪声背景中检测出微弱的泥浆脉冲信号和提高泥浆脉冲信号的信噪比，就成为泥浆脉冲传输技术中的主要问题。本论文主要章节和内容安排如下：第一章绪论部分概述了无线随钻测量地面系统在石油钻井中的重要性，分析比较了各种无线传输方式的工作原理及其性能；介绍了关于采用泥浆脉冲信号作为传输介质的无线随钻测量最新国内外研究的进展。针对目前泥浆脉冲信号的准确提取和识别准确率高的问题，提出了高分辨率泥浆脉冲信号的提取和识别研究方法。第二章分析了泥浆脉冲信号传输特性机理，分析了泥浆脉冲信号粘频速度及钻井液在理想状态下传输速度的影响因素，分析了泥浆脉冲信号幅值沿着钻柱传播的衰减程度及其影响因素，并对其衰减模型的影响因素进行了详细的数学仿真，为下一步地面信号硬件电路的设计提供了依据。第三章分析了井下数据曼彻斯特编码和PLM(Pulse Location Manager)编码格式，对两种不同编码的传输格式以及传输格式中特殊信号的规定进行了分析。为无线随钻测量数据的编码方法和解码次序提供了一个理论基础。第四章对泥浆脉冲信号噪声的处理进行了研究，由于泥浆脉冲信号受到各种噪声的干扰，首先对原始泥浆脉冲信号进行了频谱特性的研究，分析出噪声和有用信号不同的频谱特性，分析出噪声信号的类型，建立比较完善的去噪算法和去噪流程，能对泥浆脉冲信号进行有效提取，为下一步的准确识别打下基础。第五章建立了泥浆脉冲信号识别模型，提出了泥浆脉冲信号识别算法的理论基础。针对曼彻斯特编码的泥浆脉冲信号提出了以下识别算法，单比特周期内曼彻斯特编码的识别算法有：局部特征识别算法、相位识别算法、相关识别算法；多比特周期内曼彻斯特编码的识别算法有：聚类识别算法、模糊聚类识别算法；对PLM编码规则的泥浆脉冲信号识别算法上提出了上升沿特征识别算法和波形相似特征识别算法。第六章对地面信号采集箱和地面解码软件进行了设计。在了解了泥浆脉冲信号传输特性的基础上，设计一套地面信号预处理电路和硬件滤波电路，对信号进行初步的预处理和滤除非同频信号对有用信号的干扰。地面解码软件采用USB接口对数据进行采集，采用VC++6.0开发了一套地面解码软件，软件的开发设计过程中：根据不同的需求实时选择不同的数字滤波器和相关的滤波参数；在泥浆脉冲信号的识别过程中加入了各种识别算法来实时对数据进行译码；实现了数据的多线程动态存储功能。软件设计界面友好，可操作性好。最后现场实验结果表明：地面系统达到了预期的设计功能，基本完成了设计目标。最后为本文的结论部分以及未来研究展望。更多还原

【Abstract】 Presently, in the process of drilling the oil-wells, wireless MWD (measurementwhile drilling) technology has been widely used with purposes of measuring theborehole parameters accurately, improving the control to well trajectory, and guidingthe drill head to the target oil layer precisely. Wireless MWD is an abbreviation forthe measurement and transmission of the real-time underground information duringthe drilling process, with such characteristics: to obtain geological information nearthe drill head without interrupting its normal drilling process, and turn the informationinto wireless signals to transmit them up to the ground; next, these signals will beanalyzed and treated within ground system, by which the signals can be decoded in anencoding way that matches the underground apparatus, to get the orientation data,stratum features, drilling parameters and other information that are required in drillingprocess.The objective of this thesis is to improve the transmission speed and decodingaccuracy of MWD mud pulse signals. Signal transmission in drill string will beaffected by various site conditions. Consequently, pressure wave of mud pulse signal,detected by pressure sensor which is placed at the mouth of the well, contains a greatnumber of noise signals, which are generated due largely to the rotations or vibrationsof the underground machines. Such noise signals would dominate the original pulsesignals. Even that the mud pulse signals have not been disturbed by noises, theywould be attenuated sharply after a long way of transmission, or, due to the quality ofmud itself. Therefore, for the technology of transmitting the mud pulse signals, it is aprimary problem to detect the weak mud pulse signals out of those strong noiseswhich have unfixed frequencies. This paper was arranged with following mainchapters and content, as:Chapter I, introduction. This part summarized the importance of wireless MWDground system for oil-well drilling, analyzed and compared the working principlesand performances of many types of wireless transmission technologies, introducedlatest developments of China’s as well as foreign literatures which made researcheson the wireless MWD that takes the mud pulse signals as its transmission media, andpresented methods of extracting and identifying mud pulse signals with highresolution, for the accurate extraction and identification of mud pulse signals. In chapter II, characteristics of mud pulse signal transmission were analyzed;factors, which have impacts on the sticky speed of mud pulse signals as well as on thetransmission speed of drilling liquid under ideal conditions, were studied; attenuationand its influencing factors of the mud pulse signals when they run through the drillstring were researched, and a mathematical simulation model to attenuation’sinfluencing factors was built, to provide basis for the design of ground signals’hardware circuit.In chapter III, the Manchester Encoding and PLM Encoding for undergrounddata were described, Transmittal formats of these two encoding methods and theprovisions to special signals in transmission were taken into consideration, to offer atheoretical basis for wireless MWD data’s encoding mode and decoding order.In chapter IV, treatment to noises that mixed in mud pulse signals was studied.The mud pulse signals are disturbed by various noises, so, this part firstly researchedthe spectral characteristics of original mud pulse signals, figured out differentcharacteristics between noises and useful signals, analyzed the types of noise signals,and set up a relatively complete denoising algorithm and process which caneffectively extract the mud pulse signals, to lay a foundation for next accurateidentification.In chapter V, a model for identifying mud pulse signals was established topropose the theoretical basis of recognition algorithm for mud pulse signals. For thoseencoded by Manchester, the recognition algorithm ban be conducted within single-bitperiod, such as local feature recognition algorithm, phase recognition algorithm,correlation recognition algorithm, and within multi-bits period, such as clusteringrecognition algorithm, fuzzy clustering recognition algorithm; for those encoded byPLM, the rising edge feature recognition algorithm and waveform similarity featurerecognition algorithm were presented.In chapter VI, ground signal collection box and decoding software were designed.Based on knowing the transmission characteristics of mud pulse signals, this partdesigned for ground signals a set of preprocessing circuit and hardware filter circuit,by which the signals can be initially preprocessed and the signals of differentfrequencies can be filtered to eliminate the disturbance on useful signals. The grounddecoding software which was developed by using VC++6.0, includes a USB interfacewas employed to collect data. In the process of designing the software, differentdigital filters and relevant parameters were chosen in real time according to different needs. In the identification of mud pulse signals, the real time data were decoded bythose identification algorithms, realizing the multi-thread dynamic data storage. Thesoftware has friendly interface with good maneuverability. Finally, the experimentalresults showed that the ground system can reach the expected functions and thedesigned objectives had been achieved basically.The last part of this thesis is about the conclusions and prospects in futureresearches.更多还原