Abstract:An accurate data processing algorithm is at the heart of successful Nuclear Magnetic Resonance (NMR) log interpretation. The first step in the traditional inversion algorithm is inverting for the T2 distribution from the magnetization data at a single depth. A recent innovation involves finding volume fractions of the different components (e.g., capillary bound water, clay bound water, free water, oil, etc.) using data from multiple depths/measurements and adopting Blind Source Separation (BSS) techniques. NMR data inversion and BSS are both ill-posed problems. These types of algorithms are strongly influenced by noise and have significant error bars, especially for low values of T2. In this work we develop a new algorithm that utilizes a joint inversion and blind source separation., "Kernel Incorporated Non-Negative Matrix Factorization" (KINMF). The outputs are T2 distributions and the volume fractions of different components from the magnetization data. This is accomplished by incorporating multiple measurements with varying volume fraction of the individual components. This single-step, hybrid approach has a de-noising effect which generates accurate results without regularization. It leads to one-to-two orders of magnitude improvement in processing speeds (e.g., compute time for the conventional method in a clastic system is higher than 120 sec; for the KINMF it is less than 15 sec). Improved interpretation of NMR physics. This algorithm also eliminates the “smearing effects” due to the standard inversion which implements regularization using an L2 norm. This norm is typically used because of its computational efficiency and ease of implementation. The impact of other regularization techniques including L1 normalization, Entropy and Elastic–Net are also investigated. These results justify the use of conventional inversion techniques and demonstrates the improvements in interpretation due to the use of the KINMF technique. The results of the KINMF algorithm were validated using forward modeling and comparison with several experimental datasets. These included varying mixtures of brine, oil, and sand. The technique was also applied to a suite of carbonates with varying volumes of vuggy and matrix porosity. The T2 distributions in the carbonates were validated using image analysis. In all cases the KINMF technique improves the results of the inversion. The major impact of applying KINMF is for T2 relaxation times less than 100 ms and it also significantly improved computational times (enhanced real-time data processing). This should lead to broader applicability and improved physical interpretation of NMR data.

Application of Multi-exponential Inversion of NMR Relaxation Signal

Inversion Method for NMR Weak Signals with Short Relaxation Time

A statistical learning perspective on the inversion of NMR relaxation data

The Inversion of 2D NMR Relaxometry Data Using L1 Regularization.

Multi-Exponential Inversion of T-2 Spectrum in Nmr Based on Differential Evolution Algorithm

Joint Inversion Of T-1 -T-2 Spectrum Combining The Iterative Truncated Singular Value Decomposition And The Parallel Particle Swarm Optimization Algorithms

A Novel Efficient Inversion Method for Three-Dimensional NMR and the Optimization of Activation Sequences and Acquisition Parameters

Low-field NMR Inversion Based on Low-Rank and Sparsity Restraint of Relaxation Spectra

A Novel Inversion Method of 2D TD-NMR Signals Based on Realizing Unconstrained Maximization of Objective Function.

Fast Multidimensional NMR Inversion Based on Randomized Singular Value Decomposition

A Novel Method to Enhance the Inversion Speed and Precision of the NMR T2 Spectrum by the TSVD Based Linearized Bregman Iteration

An Improved TSVD Based Linearized Bregman Iterative Algorithm for the Inversion of NMR T2 Spectrum

An Inversion for LF-NMR Signals Processing with BFGS Algorithm

A Generally Regularized Inversion for NMR Applications and Beyond

A Method of Choosing the Optimal Number of Singular Values in the Inverse Laplace Transform for the Two-Dimensional Nmr Distribution Function

A Fast Inversion Method of 2D Nuclear Magnetic Resonance Based on the Novel Hybrid Algorithm

An Improved Pulse Sequence and Inversion Algorithm of T2 Spectrum

Joint Inversion and Unsupervised Learning Applied to NMR Data Processing That Eliminates the Need for Regularization

MRS signal T2 inversion by multi-exponential decomposition with inequality constraint

A Novel Efficient Inversion Method for Three-Dimensional NMR

The Research on 1D SNMR Inversion by Tikhonov Regularization Method