Ultrasonic NDT Autoencoder for Surface Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasonic Non-Destructive Test (NDT) methods struggle to accurately analyze defect signals when they interfere with initial pulses, leading to low accuracy and reliance on empirical signal processing methods.
Innovation Solution
An ultrasonic NDT method using deep learning and an autoencoder-based prediction model is employed to accurately remove initial pulses from echo signals, allowing for the extraction and analysis of defect signals. The method involves training a prediction model using normal signals and retraining it with pseudo-normal signals to enhance defect analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a defect is located around the surface of a test object, then the defect signal is reflected back to the ultrasonic transducer, but the defect signal interferes with the initial pulse making it difficult to detect
Solution Approach 1:
The patent extracts the initial pulse signal from the composite ultrasonic signal using signal processing techniques. By separating the initial pulse from the echo signal, the system can identify and analyze defect signals that would otherwise be masked by the initial pulse interference, thereby resolving the contradiction between detecting surface defects and eliminating signal interference
Solution Approach 2:
The patent introduces reference signals from defect-free regions as an intermediary to facilitate defect detection. By comparing the measured signal with reference signals, the system can identify defect signals even when they interfere with the initial pulse, effectively using the reference signal as a mediator to overcome the interference problem
2Measurement precision
If residual-based analysis method is used to compare signals, then defect analysis can be performed, but signal distortion or phase modulation errors occur and accuracy is low
Solution Approach 1:
The patent replaces the traditional mechanical signal comparison method with a wavelet transform-based signal processing approach. Instead of simply comparing residual signals which causes distortion and phase modulation errors, the system uses wavelet decomposition to analyze signals in the time-frequency domain, substituting the flawed mechanical comparison with a more sophisticated mathematical transformation that preserves signal integrity and improves accuracy
Solution Approach 2:
The patent changes the analysis parameters by transforming the signal from the time domain to the time-frequency domain using wavelet transform. This parameter change allows for more reliable defect analysis by examining both temporal and frequency characteristics simultaneously, avoiding the distortion and phase modulation errors that occur in simple time-domain signal comparison
3Measurement precision
If signal processing methods such as differentiation, filtering, deconvolution are used, then signal analysis can be performed, but considerable efforts are required to acquire optimal results and availability is reduced
Solution Approach 1:
The patent implements a universal wavelet transform-based signal processing framework that can handle multiple types of ultrasonic signals and defect conditions with a single methodology. Instead of requiring different processing techniques for different scenarios, the wavelet transform provides a unified approach that works across various signal types and defect configurations, reducing complexity while maintaining analysis capability
Solution Approach 2:
The patent applies dynamic signal processing by using wavelet transform with variable scaling and translation parameters. This allows the system to adaptively analyze signals at different time scales and frequency resolutions, providing flexible and efficient signal analysis without requiring complex fixed-parameter processing methods, thereby reducing overall system complexity while maintaining high measurement precision
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 enables accurate defect analysis on the surface of test objects despite interference with initial signals, improving the reliability and accuracy of ultrasonic NDT methods.
Implementation Method 1
an ultrasonic transducer is used to emit an ultrasonic wave onto a test object. At this time, the ultrasonic signal emitted from the ultrasonic transducer is reflected from a defect such as void or crack on the rear surface of the test object or inside the test object, and returned to the ultrasonic transducer
Data Source
AI summary
An ultrasonic NDT method and system, which can extract and analyze a defect signal even when a signal reflected from a defect interferes with a unique initial pulse of an ultrasonic transducer or a signal reflected from the surface of a test object, and an autoencoder-based prediction model training method used therefor. The method may include acquiring a measured signal by transmitting an ultrasonic wave to a test object and receiving an ultrasonic wave reflected from the test object; inputting the measured signal to an autoencoder-based prediction model and predicting a reference signal which is to be expected to be measured from a test object with no defect; calculating a residual signal as the absolute value of a difference between the measured signal and the reference signal; and analyzing information on a defect contained in the test object by analyzing the residual signal.