Back to results
US 5,517,115 AGranted
Efficient processing of NMR echo trains
AssigneeNumar Corporation
InventorManfred G. Prammer
Life of this patent
Estimating the term…
Abstract
A method and apparatus is disclosed for efficient processing of nuclear magnetic resonance (NMR) echo trains in well logging. Apriori information about the nature of the expected signals is used to obtain an approximation of the signal using a set of pre-selected basis functions. A singular value decomposition (SVD) is applied to a matrix incorporating information about the basis functions, and is stored off-line in a memory. During the actual measurement, the apparatus estimates a parameter related to the SNR of the received NMR echo trains and uses it to determine a signal approximation model in conjunction with the SVD of the basis function matrix. This approximation is used to determine in real time attributes of the earth formation being investigated.
Claims29
A method for nuclear magnetic resonance (NMR) sensing of a material under investigation, comprising the steps of:
- (a) imparting a polarizing magnetic field signal to the material for a predetermined period of time;
- (b) receiving NMR signals from a population of particles in the material, following the polarization period;
- (c) constructing a mapping of the received signals from the signal space into a parameter space associated with a predetermined set of basis functions, wherein the mapping step comprises the steps of:
- 1. sampling and digitizing of the received signals to obtain a measurement vector b;
- 2. defining a matrix A incorporating information about the set of basis functions and forming a linear system of equations for the defined matrix A and measurement vector b;
- 3. solving the system of linear equations Ax=b to obtain a solution vector x to provide an approximation of the received signals;
- wherein the solution is obtained by applying an orthogonal decomposition to matrix A;
- and
- (d) determining from the provided approximation an indication of attributes of the material under investigation;
- wherein only basis functions for which the corresponding component of the solution vector is non-negative are used in obtaining the approximation parameters, and steps 2-3 are iterated until a non-negative solution vector is obtained by defining at each iteration a reduced size matrix A.sub.i ' for which the i-th basis function corresponding to the component of the solution vector x having the largest negative amplitude is removed.
7 more elements
A method for processing NMR signals in well logging comprising the steps of:
- (a) generating a static magnetic field substantially perpendicular to the axis of a bore hole in a region surrounding the bore hole which region includes material sought to be analyzed;
- (b) generating a radio frequency magnetic field in said region in a direction substantially perpendicular to both the axis of the bore hole and to the static magnetic field for exciting nuclei of the material sought to be analyzed;
- (c) receiving NMR signals from the excited nuclei;
- (d) constructing a mapping of the received signals from the signal space into a parameter space associated with a predetermined set of basis functions, wherein the mapping step comprises obtaining parameters to provide an approximation of the received signals, and further comprises the steps of:
- (1) defining a matrix incorporating information about the set of basis functions;
- (2) forming a linear system of equations for the defined matrix and a measurement vector constructed from the received NMR measurement signals;
- (3) solving the linear system of equations to obtain parameters providing an approximation of the received signals;
- wherein the solution of the linear system of equations is obtained by applying a singular value decomposition to the defined matrix and by weighting the singular values of the matrix in accordance with a parameter dependent on the signal-to-noise ratio of the received signal measurement;
- and
- (e) determining from the obtained parameters an indication of attributes of the material sought to be analyzed.
7 more elements
An apparatus for nuclear magnetic resonance (NMR) sensing comprising:
- (a) at least one magnet having a longitudinal axis to provide substantially uniform magnetization along an axis perpendicular to said longitudinal axis;
- (b) means for causing said at least one magnet to generate a static magnetic field of generally uniform amplitude in a region remote therefrom containing materials sought to be analyzed, said static magnetic field having a static field direction perpendicular to said longitudinal axis in said remote region;
- (c) means for generating a radio frequency magnetic field in said remote region for exciting nuclei of the material sought to be analyzed and having a radio frequency magnetic field direction substantially perpendicular to both said longitudinal axis and to said static field direction;
- (d) means for receiving NMR measurement signals from the excited nuclei;
- (e) means for constructing a mapping of the received signals from the signal space into a parameter space associated with a predetermined set of basis functions, wherein the mapping means receives input from the receiving means and comprises means for obtaining parameters to provide an approximation of the received signals;
- wherein said mapping means further comprises:
- 1) means for defining a matrix incorporating information about the set of basis functions;
- 2) means for forming a linear system of equations for the defined matrix and a measurement vector constructed from the received NMR measurement signals;
- and
- 3) means for solving the system of linear equations to obtain parameters to provide an approximation of the received signals;
- wherein said means for solving comprises means for applying a singular value decomposition to the defined matrix and means for weighting the singular values of the matrix;
- and
- (f) means for determining from the provided approximation an indication of attributes of the material sought to be analyzed.
10 more elements
Dependencies are inferred from the claim text.