【作者】 张伟；

【导师】 师奕兵；

【作者基本信息】 电子科技大学 ， 测试计量技术及仪器， 2010， 博士

【摘要】 阵列声波测井仪是目前石油工程测井作业中必不可少的仪器之一。我国阵列声波测井仪与国外的先进仪器相比,在信息采集的精度和可靠性,以及测井数据处理和资料解释方面都明显落后。研制具有自主知识产权的阵列声波测井仪及配套数据处理和解释软件是目前测井行业亟需解决的重要课题。本文结合攻读博士学位期间承担的多项阵列声波测井仪器研究任务,着眼于阵列声波测井仪器的研制和工程实现,围绕井下信号采集与处理系统设计、测井数据处理、测井资料的综合解释与应用三大主线,展开了深入研究,取得了如下成果和创新:(1)井下信号采集与处理系统的体系结构设计及工程实现。在分析阵列声波测井仪的工作原理及声波测井信号特点的基础上,将阵列声波测井仪井下信号采集与处理系统的体系结构分为中控模块、信号采集与处理模块以及信号调理模块三大部分,并充分考虑其抗干扰、可靠性、低功耗等技术设计要求,经过反复试验和改进,本文所研制的系统工作稳定可靠,能够满足井下高温高压恶劣工作环境的要求。(2)测井数据处理方法研究。测井数据处理是测井资料解释的一项前期重要工作,是保证测井解释结果精度的重要前提。在井下实时数据处理方面,针对阵列声波测井仪采集到的原始测井数据受井下高温高压、强震动的恶劣环境影响比较大的特点,本文研究了测井数据的井下实时数字滤波、自动增益控制等处理技术;同时分析了井下首波到时(首波初至时间)和井下时差等测井现场质量控制参数的求取方法。在地面现场数据处理方面,研究了一种提高首波到时提取精度的短窗-长窗能量比算法;研究了在互相关系数法基础上引入插值运算来提高声波时差提取精度的方法。在测井数据后处理方面,提出一种基于抗混叠Shannon小波包变换的测井曲线高分辨率处理方法,为解决油气薄层划分和厚层细分问题提供了新的解决途径。(3)测井数据井下实时压缩和传输方法研究。对于传统电缆测井系统,研究了适合阵列声波仪井下硬件实现的基于改进SPIHT算法的测井数据井下实时压缩算法;研究了基于ADSL技术的电缆测井高速传输方法。对于随钻测井系统,针对其传输率极低,大量数据只能存储于井下存储器的特点,研究了基于小波神经网络的数据压缩算法,该算法能大大提高数据压缩比,极大地节省了井下存储空间;同时在分析随钻系统泥浆信道传输机制的基础上,研究了从强背景噪声下提取有用信号的小波神经网络泥浆信号检测法,利用该方法能够准确检测泥浆信号,达到了随钻测井环境下数据遥传的目的。(4)测井储层特性智能解释方法研究。在应用李雅普诺夫理论分析得到单个粒子稳定收敛的参数取值条件基础上,提出一种粒子群改进算法,并利用该改进算法来训练小波神经网络权值,以此构建一种高效的粒子群小波神经网络分类器,并将该分类器用于测井储层特性智能解释,取得了良好的处理效果。更多还原

【Abstract】 Array acoustic logging tool is an indispensable logging tool during the production process of oil logging. The domestic developed array acoustic logging tools lag behind the foreign advanced ones in technology such as information acquisition accuracy, information acquisition reliability, logging data processing and logging interpretation. Developping modern array acoustic logging tool and the correspending data processing and interpretation software with our own intellectual property is an urgent and important research project. Combining with some array acoustic tool research task during doctoral study and focusing on the tool’s technical realization in this dissertation, the author research on the following three aspects: the design of downhole signal acquisition and processing system, the logging data processing, the logging information interpretation and engineering application. The main works and contributions of the dissertation are as follows:(1) The research on the architecture design and the implementation of the tool’s downhole signal acquisition and processing system.Based on analysis of the operation principle of array acoustic logging tool and the characteristic of acoustic logging signal, the downhole signal acquisition and processing system is designed as three modules: the controller module, the signal acquisition and processing module and the signal conditional module.At the same time, the tool’s performance of anti-interference, reliability and low-power dissipation are taken full account into the design.Through experiments and improvement in the design, the system can run reliably in the downhole working environment of high temperature and high pressure.(2) The research on Logging data processing method.Logging data processing is an important work of logging interpretation and it is the quality assurance of logging interpretation accuracy. In downhole real-time data processing, because of the raw logging data are noised by downhole high temperature, high pressure and high vibration working environment, the data must be preprocessed to remove the interference.So dowhole real-time digital filter and automatic gain controlling processing method are researched. At the same time, two quality control parameters, the downhole head wave arrival time and the downhole time difference, are calculated to enhance logging process monitor. In the surface field data processing, an algorithm based on energy ratio of short-window and long-window is presented to improve the detection accuracy of surface head wave arrival time. Then a new method which is based on cross-correlation coefficient with interpolation is presented to improve the accuracy of calculations for surface time difference. In the surface post-period data processing, a new method for high-resolution processing of well logs based on anti-aliasing Shannon wavelet packet transform algorithm is proposed to enhance the explanation of thin beds.(3) The research on the method of down hole real-time data compression and high-speed data transmission. For traditional cable logging system, a downhole real-time data compression method based on improved SPIHT algorithm which can be used in the downhole tool is presented for decreasing data transmission content. Nextly, a logging cable high-speed data transmission method based on ADSL technology is presented for raising data transmission rate. For LWD (Logging While Drilling) system, with the transmission rate is very low, and a lot of logging data must be stored in downhole tool, a data compression algorithm based on wavelet neural network is presented in the dissertation. Experimental results show that the algorithm can improve compression ratio and save data storage space greatly. Next, the characteristics of mud pulse transmission channel of LWD system are analyzed in detail. Then a mud signal detection method based on wavelet neural network is discussed. Experimental results show that the signal can be detected from strong noise and the goal of data telemetry in LWD system can be realized by the means of the method.(4) The research on the intelligent interpretation method of logging reservoir stratum.Firstly, Lyapunov stability theory was used to discuss the convergence conditions of single particle. Then, based on that, a new strategy was introduced to improve the performance of the PSO (Particle Swarm Optimization) algorithm. Secondly, the improved PSO algorithm is used to train the parameters of the WNN (Wavelet Neural Network) instead of the gradient algorithm. Then a high efficient classifier based on improve PSO-WNN is created. Finally, the new classifier is used to for intelligent interpretation of logging reservoir stratum.更多还原