NMR Tool Calibration Using Nonlinear Echo Train Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional NMR tool calibration methods are inaccurate due to the assumption of multiplicative noise, which affects the determination of exponential decay parameters, and are limited by the short window of time for data collection, leading to inefficiencies and increased costs.
Innovation Solution
A non-linear fit of echo trains is performed using statistical techniques to reduce noise influence, with outlier removal and generation of independent test sets to determine more accurate parameters (A0 and T2) within a limited data collection window.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional NMR tool calibration methods are used, then the calibration process is simple, but the calibration accuracy is poor due to multiplicative noise assumption
Solution Approach 1:
The patent changes the mathematical model parameters from assuming multiplicative noise to using non-linear least squares with outlier removal. This involves transforming the calibration approach to use robust statistical methods that can handle noise more effectively, thereby improving calibration accuracy without requiring overly complex equipment.
Solution Approach 2:
The patent replaces the conventional mechanical calibration approach with a statistical and computational method. Instead of relying on traditional calibration procedures that assume multiplicative noise, the invention uses non-linear fitting algorithms and outlier removal techniques to determine exponential decay parameters, improving accuracy through computational rather than mechanical means.
2Quantity of substance
If data collection window is extended, then more data can be collected for calibration, but the time available for calibration is limited
Solution Approach 1:
The patent applies partial action by using a subset of available data points strategically. Instead of requiring the entire data collection window, the method identifies and uses the most informative data points through non-linear least squares fitting and outlier removal, achieving accurate calibration with a portion of the available data thereby reducing the time required.
Solution Approach 2:
The patent performs preliminary actions by pre-processing the data to identify and remove outliers before performing the calibration fit. This preliminary cleaning of the data allows for more efficient calibration using fewer data points, reducing the overall time required while maintaining accuracy.
3Measurement precision
If more calibration data is collected, then calibration accuracy improves, but the cost and time increase
Solution Approach 1:
The patent uses feedback mechanisms through iterative non-linear least squares fitting. The algorithm continuously adjusts the calibration parameters based on the fit quality and can identify when sufficient accuracy has been achieved, stopping the process early rather than requiring all possible data to be collected, thus reducing time and cost while maintaining accuracy.
Solution Approach 2:
The patent replaces time-consuming mechanical data collection with computational methods. By using non-linear fitting algorithms and statistical outlier removal, the system achieves high calibration accuracy through computational analysis of fewer data points, substituting time with computational efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances calibration accuracy and efficiency by improving the signal-to-noise ratio and reducing the time and expense associated with NMR tool calibration.
Implementation Method 1
a coil assembly that generates radio frequency (RF) control signals and detects magnetic resonance phenomena in the subsurface material
Implementation Method 2
a magnet assembly that produces a static magnetic field
Data Source
AI summary
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.