Research On Inversion Method Of Twodimensional Nuclear Magnertic Resonance Spectrum In Low Field

The technique of time-domain nuclear magnetic resonance(NMR) in low field is well-known as a green, noninvasive, fast and accurate analytic tool. Coupled with the lower cost and easier miniaturization when using low-field magnet, this technique is having been involved into various fields, such as food, agriculture, energy, biomedical. But the sampling data is not understandable intuitionally, inversion is needed to produce a relaxation spectra easy to interpret.One-dimensional(1D) spectrum is currently the most widely used time-domain spectrum. But with the continuous deepening of the application of 1D spectrum, it is found that different peaks corresponding to different components are always overlapped together, which would be a fatal weakness in quantitative analysis. Two-dimensional spectroscopy emerged at this moment.To get a 2D map, an inverting of sampled data is needed. Compared with inversion in 1D case, 2D inversion should handle measured data in a much larger scale, and the problem itself is much more ill-posed. To address these challenges, researchers home and abroad have proposed many different inversion algorithms, but these algorithms have some limitations. In response to this situation, this thesis proposed a new 2D inversion algorithm based on noise fitting.This thesis introduces the application background and the research status of 2D inversion method firstly. Then the physical basis of NMR and the mechanisms of NMR relaxation of fluid in porous media is following. Also some most commonly used sequences and applications of time-domain NMR technique are addressed in this part. After a brief review, the existing algorithms are evaluated and the pros and cons of each method are summarized. This thesis proposed a new 2D inversion algorithm after a deep exploring of the advantages and disadvantages of the existing methods. Each step of the proposal algorithm is discussed in detail. Finally, this algorithm is used to process lots of synthetic and experimental data to verify its reliability.Experimental results show that this proposed noise fitting based 2D inversion algorithm for low field NMR is robust and accurate. And it can meet the analytical requirements in a variety of application areas and different sampling conditions. In conclusion, the proposed algorithm has broad application prospects.