Neural Network Defect Geometry Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the geometry of defects in objects, such as pipelines, are inaccurate due to conservative geometric approximations, leading to underestimated burst pressures and operating pressures, and are subjective and costly, especially when evaluating complex geometries with multiple defects.
Innovation Solution
A method using multiple non-destructive measurement data sets, such as MFL and EMAT, to reconstruct defect geometry through neural networks, providing a more accurate representation of defect geometry and load limits by combining data from different measurement methods and iterative adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conservative geometric approximations (box model) are used to evaluate defects, then the evaluation process is simplified and can be performed by trained individuals, but the defect geometry accuracy deteriorates and burst pressure is underestimated
Solution Approach 1:
The patent replaces manual interpretation methods with automated neural network-based image processing. The neural network automatically identifies defect boundaries and reconstructs three-dimensional geometry from measurement data, eliminating the need for manual box-model approximation while achieving both high accuracy and automation.
Solution Approach 2:
The patent transforms the evaluation approach from simple dimensional parameters (length, width, depth) to complex three-dimensional geometric reconstruction. By changing the parameter representation from basic box dimensions to detailed surface geometry captured through image processing, the system achieves accurate burst pressure calculation while maintaining ease of use through automation.
2Measurement precision
If multiple non-destructive measurement methods are combined to improve defect detection accuracy, then the measurement precision improves, but the device complexity and evaluation complexity increase
Solution Approach 1:
The patent merges multiple measurement data sets (MFL, EMAT, UT, EC) into a unified evaluation process. The neural network simultaneously processes data from different measurement methods, integrating their complementary strengths to achieve comprehensive defect characterization while presenting a single automated evaluation output.
Solution Approach 2:
The patent creates a universal evaluation system that can handle multiple types of measurement data through a single neural network framework. The system is designed to process various non-destructive measurement methods using the same image processing algorithms, making the complex multi-method approach as easy to use as single-method evaluation.
3Reliability
If in-person site inspections are performed to verify defect predictions, then the reliability of defect assessment improves, but the inspection costs and time consumption increase significantly
Solution Approach 1:
The patent enables the measurement system to self-verify through cross-validation of multiple measurement methods. The neural network analyzes consistency among different measurement data sets (MFL, EMAT, UT, EC) to automatically assess prediction reliability, eliminating the need for external verification through costly in-person inspections while maintaining high confidence in results.
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 increases the accuracy of burst pressure calculations by 10%-20%, allowing pipelines to operate at higher pressures and reducing the need for frequent in-person inspections, while minimizing the risk of singular solutions.
Implementation Method 1
scans of magnetic flux leakage data (MFL data) based on magnetic flux leakage measurements (MFL measurements)
Implementation Method 2
electromagnetic-acoustic methods (EMAT methods), in which sound waves, especially in the form of guided waves, are generated in the pipe wall of the object to be examined due to eddy current-induced magnetic fields
Implementation Method 3
electromagnetic-acoustic methods (EMAT methods), in which sound waves, especially in the form of guided waves, are generated in the pipe wall of the object to be examined due to eddy current-induced magnetic fields
Implementation Method 4
methods directly introducing ultrasound into the object wall, hereinafter referred to as ultrasonic methods
Data Source
AI summary
Method for determining the geometry of one or more real, examined defects of a metallic, in particular magnetizable object, in particular a pipe or a tank, by means of at least two reference data sets of the object generated on the basis of different, non-destructive measurement methods,wherein the object is at least partially represented on or by an at least two-dimensional, preferably three-dimensional, object grid, in an EDP unit,wherein an output defect geometry, in particular on the object grid or an at least two-dimensional defect grid, is generated by inversion of at least parts of the reference data sets, in particular by at least one neural network (NN) trained for this object, a respective prediction data set for the non-destructive measurement methods used in the generation of the reference data sets is calculated on the basis of the output defect geometry by a simulation routine, a comparison of at least parts of the prediction data sets with at least parts of the reference data sets is carried out and, depending on at least one accuracy measure, the method for determining the geometry of the defect is terminated or an iterative adjustment of the output defect geometry to the geometry of the real defect(s) is carried out, as well as methods for determining a load limit (FIG. 1).