Claims
- 1. A method for generating an improved estimate of at least one of horizontal conductivity, dip angle, azimuth and anisotropy parameter of an earth formation penetrated by a wellbore, from dual-frequency transverse electromagnetic induction measurements of said formations, comprising:
- generating an initial estimate of said at least one of horizontal conductivity, dip angle, azimuth and anisotropy parameter from said dual-frequency transverse induction measurements made at each one of a plurality of base frequencies;
- imparting said initial estimates from each of said plurality of base frequencies into a trained neural network; and
- calculating said improved estimate using said trained neural network.
- 2. The method as defined in claim 1 wherein said neural network comprises a radial basis function model.
- 3. The method as defined in claim 1 wherein training of said neural network comprises:
- generating models of earth formations each said model having a known value of said at least one of horizontal conductivity, anisotropy parameter, dip angle and azimuth;
- synthesizing voltages which would be measured by said transverse electromagnetic induction logging instrument in response to each of said models;
- calculating initial estimates from said synthesized voltages; and
- inputting said initial estimates and said known values from each of said models to said neural network to cause said network to learn a relationship between said initial estimates from said synthesized voltages and said known values.
- 4. The method as defined in claim 1 wherein generating said initial estimate comprises:
- measuring electromagnetic induction signals induced by induction transmitters oriented along three mutually orthogonal axes, one of said axes substantially parallel to an instrument axis, said signals measured using first receivers each having a magnetic moment parallel to one of said orthogonal axes and second receivers each having a magnetic moment perpendicular to a one of said orthogonal axes which is perpendicular to said instrument axis, the method comprising;
- calculating a relative angle of rotation of said one of said orthogonal axes which is perpendicular to said instrument axis, with respect to a direction of said horizontal conductivity, from said received signals measured perpendicular to said instrument axis;
- calculating an intermediate measurement tensor by rotating magnitudes of said receiver signals through a negative of said angle of rotation;
- calculating a relative angle of inclination of one of said orthogonal axes parallel to said axis of said instrument, with respect to said direction of said vertical conductivity, from said rotated magnitudes;
- rotating said rotated magnitudes through a negative of said angle of inclination;
- calculating said horizontal conductivity from said magnitudes of said receiver signals after said second step of rotation; and
- calculating an anisotropy parameter from said magnitudes after said second step of rotation and calculating said vertical conductivity from said horizontal conductivity and said anisotropy parameter.
- 5. A method for generating an improved estimate of at least one of horizontal conductivity, dip angle, azimuth and anisotropy parameter of an earth formation penetrated by a wellbore, comprising:
- inserting a transverse electromagnetic induction logging instrument comprising mutually orthogonal transmitters and receivers and a dual-frequency alternating current source selectively connectable to each of said transmitters to induce alternating magnetic fields in said formations along an axis parallel thereto, said dual frequency source comprising a plurality of base frequencies;
- detecting voltages at each said receiver corresponding to electromagnetic induction phenomena in said earth formation, said voltages including one component corresponding to reception along the same axis as each of said induced magnetic fields, and reception orthogonal to said induced magnetic fields;
- generating an initial estimate of said at least one of horizontal conductivity, apparent dip angle, apparent azimuth and anisotropy parameter from said detected voltages at each one of said plurality of base frequencies;
- imparting said initial estimates from each of said plurality of base frequencies into a trained neural network; and
- calculating said improved estimate using said trained neural network.
- 6. The method as defined in claim 5 wherein said neural network comprises a radial basis function model.
- 7. The method as defined in claim 5 wherein training of said neural network comprises:
- generating models of earth formations each said model having a known value of said at least one of horizontal conductivity, anisotropy parameter, dip angle and azimuth;
- synthesizing voltages which would be measured by said transverse electromagnetic induction logging instrument in response to each of said models;
- calculating initial estimates from said synthesized voltages; and
- inputting said initial estimates and said known values from each of said models to said neural network to cause said network to learn a relationship between said initial estimates from said synthesized voltages and said known values.
- 8. The method as defined in claim 5 wherein generating said initial estimate comprises:
- measuring electromagnetic induction signals induced by induction transmitters oriented along three mutually orthogonal axes, one of said axes substantially parallel to an instrument axis, said signals measured using first receivers each having a magnetic moment parallel to one of said orthogonal axes and second receivers each having a magnetic moment perpendicular to a one of said orthogonal axes which is perpendicular to said instrument axis, the method comprising;
- calculating a relative angle of rotation of said one of said orthogonal axes which is perpendicular to said instrument axis, with respect to a direction of said horizontal conductivity, from said received signals measured perpendicular to said instrument axis;
- calculating an intermediate measurement tensor by rotating magnitudes of said receiver signals through a negative of said angle of rotation;
- calculating a relative angle of inclination of one of said orthogonal axes parallel to said axis of said instrument, with respect to said direction of said vertical conductivity, from said rotated magnitudes;
- rotating said rotated magnitudes through a negative of said angle of inclination;
- calculating said horizontal conductivity from said magnitudes of said receiver signals after said second step of rotation; and
- calculating an anisotropy parameter from said magnitudes after said second step of rotation and calculating said vertical conductivity from said horizontal conductivity and said anisotropy parameter.
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is a continuation-in-part of application Ser. No. 09/042,982 filed on Mar. 17, 1997, entitled, "Conductivity Anisotropy Estimation Method for Inversion Processing of Measurements Made by a Transverse Electromagnetic Induction Logging Instrument", and assigned to the assignee of this invention.
US Referenced Citations (4)
Number |
Name |
Date |
Kind |
5251286 |
Wiener et al. |
Oct 1993 |
|
5300770 |
Allen et al. |
Apr 1994 |
|
5444619 |
Hoskins et al. |
Aug 1995 |
|
5828981 |
Callender et al. |
Oct 1998 |
|
Continuation in Parts (1)
|
Number |
Date |
Country |
Parent |
042982 |
Mar 1998 |
|