The Study Of NMR Logging Data Processing Method And Software Development

After studying the processing methods of various nuclear magnetic resonance logging tools which are produced by different companies, developed a set of uniform data processing flow. On the basis of data processing methods of various nuclear magnetic resonance logging instruments, consolidated the different processing flows of different instruments and designed a system of nuclear magnetic resonance data processing (CIFLog-NMR) in the CIFLog (a new generation of well logging data processing system which was developed by PetroChina). This paper focuses on the following aspects:First, the research of data processing method. Including raw data preprocessing, T2spectrum inversion, parameters calculation and oil-gas analysis method, the main research object of which was T2spectrum inversion method, and implemented some inversion methods, such as singular value decomposition (SVD), model smoothing algorithm, SIRT algorithm and fast scan operator algorithm. According to the shortcoming of conventional SVD inversion algorithm, the paper put forward to the best singular value points number and the modified SVD method, which improved the problem of conventional SVD algorithm that does not apply to low signal-to-noise ratio (SNR) data. In addition, the paper implemented raw data preprocessing algorithm, nuclear magnetic porosity and permeability calculation algorithm. The cut-off value method, time domain analysis (TDA) and diffusion analysis (DIFAN) were also implemented for analysis of oil and gas. The pore structure parameter calculation and pseudo capillary pressure curve calculation method were implemented for the analysis of pore structure.Second, the parallel processing improvement of data processing. Aiming at the problem of imaging logging data processing speed, the paper put forward to the idea of parallel processing, which is based on the personal computer with multi-core. According to various situations, different parallel processing models were designed to fit the NMR data processing methods, which make the speed of NMR data processing largely faster. This parallel model not only could be used in NMR data processing, it also would be used in improving other time-consuming logging data processing situation.Third, software development and integration. NMR logging data processing system (CIFLog-NMR) was designed. It was based on the new generation of logging data processing platform CIFLog1.0, which was developed by PetroChina. The system not only has simple operation and friendly user interface, the processing flow is also simplified. The operators only need to set a few parameters, and will achieve to a good treatment effect. It supports precise processing by detailed parameter settings. CIFLog-NMR has intelligent help system and online upgrade system. Intelligent help system can determine the user’s current position and status automatically, and provide the accurate help information. It reduces the difficulty of the NMR logging data processing and interpretation. Through the online upgrade system, users can easily achieve the latest version from server.