基于遗传算法的核磁共振T_2分布反演

NMR T_2 Distribution Inversion Based on Genetic Algorithm

针对核磁共振回波的正演理论,提出了一种新的基于遗传算法的核磁共振(NMR)弛豫信号多指数反演方法,给出了新方法的具体实现过程,并对理想模型和实际测井数据进行了处理。GA算法与基于奇异值分解的MAPII算法和最小二乘法反演结果对比表明,GA算法与其他方法的结果具有较好的一致性,特别是在信噪比较低的情况下,该算法仍然能保证弛豫谱的真实性。该方法不仅能够满足各种核磁共振测井数据处理的需求,而且也能用于核磁共振岩心分析。

A new genetic algorithm （GA） based multi-exponential inversion algorithm for NMR T1 and T2 relaxation distributions of rock and free fluid is presented. The concrete implementation process of the GA is discussed in detail. Some GA inversions of modeling and practical logging data are calculated. Comparison between results of the GA algorithm and MAPII algorithm on SVD or LS algorithm verifies that the GA has rather good consistency with the others, especially, even in conditions of low signal-noise-ratio it can still give true relaxation distribution. The GA inversion algorithm is valuable for both NMR experimental core analysis and NMR log interpretation.