Research Of MRS Data Processing Based On Genetic Algorithm And Nonlinear Programming

Currently, Magnetic Resonance Sounding (MRS) is the only direct detectiongroundwater of the ground geophysical methods. Compared with the traditionalmethods,this technique has the advantage for the unique explanatory of inversion andhigh efficiency. JLMRS-1MRS instrument is researched and developed by JilinUniversity independent.In the specific project, the data of measurement needs forprocessing. Hydrogeological data is based on the measured signal has been treated. Inthis paper, groundwater detection MRS signal data has been processed, research andexperimental comparison. According to the JLMRS-1underground detection system,based on the collection MRS signal, the data processing methods aiming at the variousideal models and data included the different SNR are put forward, for example, datafitting method and inverse interpretation method. In addition, the correspondingsoftware is also designed including Polynomial fitting and Fourier fitting. Through thecalculation by the former two methods, the response result can be received, and thenthe ideal data fitting method can be identified. The inverse calculation for MRSresponse result includes the depth and thickness of aquifer and the calculation about the size of moisture content. Aiming at the problems existing in the process ofTikhonov regularization inverse calculation, the corresponding parameter can bereceived by using the nonlinear optimization theory. At last, according to the availableresults, the Tikhonov regularization parameter corresponding to the actual situationcould be set using the Tikhonov regularization inverse method. The optimal inversemethod for MRS response signal data adopts the genetic algorithm combined withnonlinear programming to calculate the underground water content. Applying the fileddata, the test compared the drilling results can confirm the validity and veracity ofthese data processing methods.For nuclear magnetic resonance (MRS) groundwaterdetectors to further improve the overall performance to lay the foundation.According to the JLMRS-1underground detection system, based on the collectionMRS signal, the data processing methods aiming at the various ideal models and dataincluded the different SNR are put forward, for example, data fitting method andinverse interpretation method. In addition, the corresponding software is also designedincluding Polynomial fitting and Fourier fitting. Through the calculation by the formertwo methods, the response result can be received, and then the ideal data fittingmethod can be identified. The inverse calculation for MRS response result includes thedepth and thickness of aquifer and the calculation about the size of moisture content.Aiming at the problems existing in the process of Tikhonov regularization inversecalculation, the corresponding parameter can be received by using the nonlinearoptimization theory. At last, according to the available results, the Tikhonovregularization parameter corresponding to the actual situation could be set using theTikhonov regularization inverse method. The optimal inverse method for MRSresponse signal data adopts the genetic algorithm combined with nonlinearprogramming to calculate the underground water content. Applying the filed data, thetest compared the drilling results can confirm the validity and veracity of these dataprocessing methods.The first chapter is the introduction; it mainly elaborates the research meaning andthe status of research at home and abroad, puts forward the necessity improving the processing method of MRS underground detection data, aims to explain that the newdata processing method and unconventional numerical algorithm are worth studying.The second chapter introduces the basic theory of the MRS underground detectionsystem. At first, it elaborates the MRS theory and the component of detection system;secondly, it analyses the inverse method of MRS underground detection signal and theoptimal design of corresponding Tikhonov regularization inverse calculate.The third chapter explicitly states the preprocessing of inverse data of MRSunderground detection signal, it includes Polynomial fitting and Fourier fitting.Through the applying example, the data fitting strategy is synthetically analyzed andsummarized.The forth chapter designs the data inversion of MRS signal basing the geneticalgorithm combined with nonlinear programming, at the same time, this chapter alsoreceives the corresponding results.The fifth chapter mainly verifies the correctness of the most optimal algorithm,which is Fourier fitting algorithm combined with the nonlinear programming basingthe heredity.The sixth chapter is the conclusion and the next study proposal.Through verifying by the experiment, this article studies the data processingmethod of MRS underground detection instrument, conquers the problems existing inthe data processing of MRS underground detection instrument, and elevates theaccuracy of data fitting and inverse calculation, sets up the foundation to furtherimprove the overall performance of MRS underground detection instrument.