Halliburton Energy Services, Inc.

Jie Yang

Rebecca Corina Jachmann

Songhua Chen

Arcady Reiderman

Lilong Li

Reena Agarwal Chanpura

Daniel Lee Miller

Cristian Lenin Sosareyes

Christopher Jon Conrad

You Wang

Pierre Aerens

A method for calibrating an NMR tool includes receiving N echo trains from an NMR tool; performing a linear fit of the echo trains to determine an initial parameter of an exponential decay curve; generating a plurality of test sets, each identifying a subset of the echo trains as test samples and identifying at least one echo train as a control sample. The selected test samples in each test set is independent of the selected test samples in other test sets. The method also includes performing, for each test set, a non-linear fit of the test samples based on the initial parameter to determine a test value for the parameter of the test set; determining an error value for each test set; selecting a test set having an error value less than an error threshold; and calibrating the NMR tool based on the test value of the selected test set.

Systems and methods for calibrating a nuclear magnetic resonance tool

Systems and methods for compressing nuclear magnetic resonance (NMR) data are disclosed. The methods include the steps of selecting a T1/T2 ratio value, inverting a first portion of the echo train data using the selected T1/T2 value, and reconstructing the inverted echo train using the selected T1/T2 value. A quality measure is calculated based on a comparison of the reconstructed echo train to the original echo train. When the quality measure is satisfactory, the NMR data is inverted using the selected T1/T2 value and transmitted to the surface.

Adaptive T1/T2 ratio selection for optimal compression

Described herein are methods for removing the vibration induced additional signal obtained during downhole NMR operations. The additional signal is removed by analyzing a number of instances of data sets neighbors, at either the raw echo, reconstructed echoes, or the spectrum which results from inversion. A number of neighboring data instances are analyzed together to find the minimal (lowest) common values in each. Thereafter, the minimal value replaces the previous value across the data instances, thereby removing the extra signal.

Correction of nuclear magnetic resonance data in high vibration environments

Described herein are systems and techniques for improving accuracies of determinations made using a nuclear magnetic resonance (NMR) sensing device when the NRM sensing device collects data in a wellbore. NMR sensing devices include a magnet that provides a magnetic field that aligns the spins of protons in substances near the NMR sensing device. The magnetic field strength provided by this magnet affects the sensitivity of the NMR sensing device and affects frequencies that the NMR sensing device effectively uses when the NMR sensing device operates. Furthermore, the field strength of magnets used in an NMR sensing device varies with temperature. Since temperatures within a wellbore vary significantly and since these temperatures affect how an NMR sensing device operates, systems and techniques of the present disclosure collect calibration data in a calibration chamber such that data sensed by the NMR sensing device can be interpreted more accurately when temperatures change.

Nuclear magnetic resonance sensing device calibration system and method

Some implementations transmit nuclear magnetic resonance (NMR) data between downhole tools and surface-based computers. A method may include generating a first vector in which each entry includes a spectrum coefficient associated with NMR echo signals. The method may include generating a cumulative sum vector in which each entry includes a sum including a spectrum coefficient of a corresponding entry in the first vector plus any spectrum coefficients of any entries that precede the first corresponding entry in the first vector. The method may include down sampling the cumulative sum vector to form a down-sampled vector including a subset of the cumulative sum vector. The method may include generating a compressed vector in which each entry includes a difference of a corresponding entry in the down sampled vector and any entry directly preceding the corresponding entry in the down sampled vector and transmitting the compressed vector over a communication channel.

Split inversion for nuclear magnetic resonance data

A NMR tool for use in a wellbore in a subterranean region includes a magnet assembly to produce a magnetic field in a volume in the subterranean region; an antenna to produce an excitation in the volume, and to receive a plurality of spin echo waveforms from the volume; and a computing system coupled to the antenna and configured to: apply a first acquisition window having a first duration to the spin echo waveforms to generate a corresponding first echo train including a first plurality of NMR echo signal amplitudes; apply a second acquisition window having a second duration different than the first duration, to at least some of the spin echo waveforms to generate a corresponding second echo train including a second plurality of NMR echo signal amplitudes; and determine a relaxation parameter based on a single inversion of the first and second echo trains.

Data inversion to reduce motion effects on nuclear magnetic resonance relaxation data

Systems and methods are disclosed for identifying one or more vibration mechanisms or other mechanism associated with tool failure. In one example, a machine learning data system comprises a data pre-processing module, a machine learning training module, and a predictive module. The pre-processing module analyzes dynamic sensor input to identify one or more vibration mechanisms associated with a downhole tool. The machine learning training module identifies a correlation between dynamic sensor data and failure data for a plurality of downhole tools. The predictive module correlates the one or more vibration mechanism with tool failure or performance.

Use of vibration indexes as classifiers for tool performance assessment and failure detection

Methods to operate an NMR tool, methods to simulate a numerically-controlled oscillator of an NMR tool in real time, and downhole NMR tools are presented. A method to operate an NMR tool includes determining a phase shift of a sinusoidal wave, determining a number of look-up tables and a number of terms of Taylor Expansions performed to obtain a value corresponding to a phase angle of the phase shift, and separating the phase angle into a first component and a second component. The method also includes obtaining a first value corresponding to the first component from the number of look-up tables, performing the number of terms of Taylor Expansions on the second component to obtain a second value corresponding to the second component, combining the first value and the second value to obtain the value of the phase angle, and generating the sinusoidal wave having the phase shift.

Methods to operate a nuclear magnetic resonance tool, methods to simulate a numerically-controlled oscillator of a nuclear magnetic resonance tool in real time, and downhole nuclear magnetic resonance tools

Noise in drilling operation measurements can be eliminated or reduced using machine learning. For example, a system described herein can receive one or more measured signals in a logging-while-drilling process for drilling a wellbore. The system can determine a coupling factor for noise in the one or more measured signals. The system can generate a corrected signal by removing the noise multiplied by the coupling factor from the one or more measured signals. The system can output the corrected signal for use in drilling operations in the wellbore.

Noise elimination or reduction in drilling operation measurements using machine learning

A method and system for transmitting an electromagnetic (EM) signal in a wellbore to improve an open loop system. The method comprises measuring a parameter indicative of a property of a transmitter using a sensor in the wellbore and generating an input signal for the transmitter based at least in part on the measured parameter to correct distortions exhibited by the output signal of the transmitter. The method comprises also comprises applying the input signal to the transmitter to transmit the EM signal corrected for distortions. The system comprises an antenna, a transmitter, a sensor, and a controller. The controller is configured and operable to generate an input signal for the transmitter based at least in part on the measured parameter to correct distortions exhibited by an output signal of the transmitter; and apply the input signal to the transmitter to transmit the EM signal corrected for the distortions.

Method and system to improve open loop systems

Data is received downhole from a downhole logging tool to produce received data. The received data is processed to produce processed data. A binary formatting option is selected from a group of binary formatting options, wherein applying the selected binary formatting option to the processed data produces fewer bits than applying any other binary formatting option of the group of binary formatting options to the processed data. The selected binary formatting option is applied to the processed data to produce binary data. The binary data is transmitted uphole. The binary data is received uphole. The received binary data is decompressed to produce uncompressed data. The uncompressed data is processed uphole.

Downhole adaptive data compression and formatting

Echo signals are acquired from operation of a nuclear magnetic resonance logging (NMR) tool in a borehole. A multi-inversion process is performed on the echo signals. The multi-inversion process includes a downhole processor performing a downhole inversion. The multi-inversion process includes reconstructing echoes from coefficients in the downhole inversion. Reconstructing echoes is performed by one of the downhole processor or an uphole processor. The multi-inversion process includes transmitting one of: (a) the coefficients in the downhole inversion or (b) the reconstructed echoes to the uphole processor. The multi-inversion process includes the uphole processor performing an uphole inversion on the reconstructed echoes to produce final coefficients used in a spectrum.

Reducing data bandwidth requirements in downhole nuclear magnetic resonance processing

Echo signals are acquired from operation of a nuclear magnetic resonance logging (NMR) tool in a borehole. The acquired echo signals are processed with respect to an inversion of each matrix in a plurality of matrices, wherein each matrix in the plurality of matrices accounts for motion of the NMR tool at a respective rate of penetration (ROP) of the NMR tool in the borehole, to produce respective spectrum coefficients for each ROP. Each spectrum is processed with respect to the respective matrices in the plurality of matrices to produce a test echo signal for each ROP. The acquired echo signals and the test echo signal for each ROP are processed to produce an error for each ROP. The ROP for the minimum of the errors is identified as the ROP of the NMR tool.

Speed of tool assessment via speed kernels

A nuclear magnetic resonance (NMR) logging tool configured to be positioned in a borehole created in a subsurface formation. The NMR logging tool includes a magnet to induce a magnetic field in the subsurface formation, a transmission line, and a first shield positioned between the magnet and the transmission line. The NMR logging tool includes a transmitter and an antenna coupled to the transmitter by the transmission line. The antenna is to transmit a radio frequency (RF) magnetic field into the subsurface formation in response to the transmitter supplying power by the transmission line.

Growing community of inventors

Paoli, PA, United States of America

Jie Yang