No 12 (2023)

In the current NDT 4.0 revolution, machine learning and artificial intelligence have emerged as the major enablers for non-destructive testing and evaluation (NDT&E) of industrial components. However, recent developments in active thermal NDT (TNDT) support its use as a practical method for checking a range of industrial components. Additionally, recent post-processing research in TNDT has developed several machine learning models to replace human interaction and offer automatic defect detection. However, the smaller area of the flaws and their related few thermal profiles than the wide sound area, leading to imbalanced datasets, make it difficult to train a supervised deep neural. Recently added to TNDT are anomaly detection models and one-class classifiers, both of which are commonly applied machine learning models to real-world issues. The accuracy and other important metrics in autonomous defect detection are influenced by the hyper-parameters of these models, such as contamination factor, volume of training data, and initialization parameter of the relevant model. The current paper investigates how initialization parameters affect these models’ TNDT capabilities for automated flaw detection. Using quadratic frequency modulated thermal wave imaging (QFMTWI), a carbon fiber-reinforced polymer specimen with variously sized artificially produced back-holes at different depths is examined. A good hyper-parameter for automatic flaw identification is chosen after qualitatively comparing testing accuracy, precision, recall, F-score, and probability.

Quality control of welding is usually carried out during the visual inspection process and is highly dependent on an operator experience. In the paper, it is proposed an approach to automatic detection and classification of a defective region, where segmentation of the analyzed photographic image of a weld (i.e., its division into defective and defect-free regions) is performed using the region growing procedure. The starting points for this procedure are selected by the authors' robust method of interval fusion with preference aggregation (IF&PA) on the base of image histogram analysis. Testing of the proposed approach for real life photographic images showed its ability to detect different types of weld defects with higher accuracy compared to traditional methods such as Otsu method and k-means.

A technique for approximate solution of the inverse geometric problem of magnetostatics for a plate made of a soft magnetic ferromagnet in a magnetic field is presented. The technique is presented both for the case of the location of magnetic transducers directly above the surface defect of loss of continuity of the metal, and for the case in which the magnetic transducers are located above the defect-free surface of the plate. It is assumed that the plate is accessible from one side only. The sizes of defects in which the proposed technique works reliably are determined. It is shown that the proposed technique can be used in mobile devices to carry out flaw detection of drill pipes using the MFL (Magnetic flux leakage) method directly at drilling sites.

The issues of classification and determination of parameters of surface operational defects according to the results of ultrasonic, eddy current and visual and measuring methods of nondestructive testing are considered. At the same time, the visual and measuring method was realized with the use of a television inspection camera equipped with a computer vision function and a laser triangulation sensor. The paper presents a dataset containing 5760 images of pipelines with and without pitting corrosion. A convolutional neural network (CNN) is presented, which has been applied to classify the images obtained from a TV inspection camera into images without corrosion and images with pitting corrosion. The paper presents a dataset containing 269 measurements of planar and volumetric surface defects. A model for surface defect sizing based on gradient boosting is presented. The paper develops an algorithm for classification and sizing of surface defects in complex diagnostics, in which the obtained models are applied, and determines the accuracy of this algorithm by the RMSE metric, which was calculated within the studying test data set and amounted to 0.011 mm.