Defect depth is an essential indicator in magnetic flux leakage (MFL) detection and estimation. The quantification errors for defect depth are closely related to length and width errors, and this feature has always been used to support the operator's judgment in defect identification. However, the existing defect quantification algorithms based on shallow and deep neural networks only employed simple general network structures inspired by the field of artificial intelligence; consequently, these network structures lack the support of physical concepts and result in large quantification errors regarding defect size, especially depth. In this article, to describe and integrate the above theory into a deep neural network, we propose a physics-informed doubly fed cross-residual network (DfedResNet) suitable for MFL defect detection based on deep learning. Physics-based MFL defect quantification theory is studied and integrated into loss functions during the neural network training. DfedResNet quantifies defects in MFL data and automatically extracts deep features of defects. The experimental results show that it effectively achieves high-precision quantification of defect length, width, and depth simultaneously, especially defect depth. Moreover, it considers data from all three dimensions during network training, and use the originally measured magnetic signal data in place of recognized images to avoid defect information loss and further improve the quantification accuracy. The deep DfedResNet model proposed in this article reduces defect length and width quantification errors to within 0.3 mm and defect depth quantification errors to within 0.4% t . In addition, compared with other network structures and traditional algorithms, DfedResNet improves defect quantification accuracy by 1–2 orders of magnitude and thus achieves a high quantification performance.

Magnetic flux leakage testing (MFL) is the most widely used nondestructive testing technology in the safety inspection of oil and gas pipelines. The analysis of MFL test data is essential for pipeline safety assessments. In recent years, deep-learning technologies have been applied gradually to the data analysis of pipeline MFL testing, and remarkable results have been achieved. To the best of our knowledge, this review is a pioneering effort on comprehensively summarizing deep learning for MFL detection and evaluation of oil and gas pipelines. The majority of the publications surveyed are from the last five years. In this work, the applications of deep learning for pipeline MFL inspection are reviewed in detail from three aspects: pipeline anomaly recognition, defect quantification, and MFL data augmentation. The traditional analysis method is compared with the deep-learning method. Moreover, several open research challenges and future directions are discussed. To better apply deep learning to MFL testing and data analysis of oil and gas pipelines, it is noted that suitable interpretable deep-learning models and data-augmentation methods are important directions for future research.

It is challenging to apply deep learning in professional fields that lack big data support, especially in industrial structure health assessments using ultrasonic guided wave nondestructive testing (NDT) method. To solve this problem, one feasible solution is to introduce the concept of NDT physics into a deep neural network to compensate for the network's poor predictive abilities when trained on small datasets. Therefore, we propose a physics-informed deep neural network, named GuwNet, based on a unidirectional oblique-focusing (UOF) high-frequency, high-order shear horizontal guided wave electromagnetic acoustic transducer (EMAT) to quantify microcrack defects more accurately. First, the designed focused-transmission omnidirectional-reception UOF-EMAT can produce pure high-frequency, high-order guided waves. Through a circumferential arrangement of multiple receiving transducers, the maximum amount of information can be obtained regarding the reflected waves of the defect. This method solves the inherent problems of an EMAT (i.e., low energy conversion efficiency) and achieves effective detection of microcrack defects. Second, we study the quantification principle of microcrack defects suitable for UOF-EMAT, and propose a deep neural network using physical knowledge regarding this theory. We rationally design the network structure based on the quantitative principles and logic obtained from this article. In addition, feedback and feedforward loss functions suitable for evaluating different forms of variables are proposed to integrate the physical concepts of ultrasonic guided wave testing into the neural network training. Finally, we verify the performance of the proposed GuwNet based on the UOF-EMAT. Compared with traditional nonphysics-informed methods, the length, depth, and direction of the quantification errors are reduced to 0.127 mm, 0.279% d_t , and 1.843°, respectively, and the average quantification error is reduced by more than 80%.

Cracks that may cause pipeline cracking and leakage become the main risk of in-service pipelines after conventional metal loss defects have been detected. Therefore, it is imperative to develop ultra-high-definition magnetic flux leakage (MFL) detection for cracks. Due to the larger spatial sampling interval of conventional MFL inspections than the crack defect width, sparse sampling will result in problems such as missed detection of crack defects or insufficient signal sampling of crack defect MFL signals. To address the problem, this paper presents a detection topology in which the magnetic field signal is integrated first and then sampled, and it can effectively sample the leakage magnetic field of cracks. Furthermore, a magnetic field spatial integration (MFSI) method is proposed based on the magnetic dipole model, which develops an analytical solution for the depth quantification of crack defects. Using the multi-MFL spatial integral values under different lift offs, iterative equations are constructed, and the numerical solution for the width quantification of crack defects is provided. The finite element simulation and physical experiment are constructed. And, the results show that the proposed MFSI method can realize the quantification of crack size. When the crack defect aspect ratio is greater than 4, the quantization error of the crack depth and width does not exceed 0.7 mm and 0.1 mm, respectively. The robustness of the MFSI method for crack quantification is also discussed and verified. Moreover, based on theoretical analysis and simulation data, this paper concludes that the distribution of cracks MFL signal is width-insensitive, which would cause that classic MFL opening profile recognition methods for metal loss defects are no longer applicable to crack defects. Therefore, the proposed MFSI method is useful and meaningful for crack defect quantification in ultra-high-definition MFL detection.

Ultrasonic guided wave testing technology is currently widely used in industrial nondestructive testing (NDT), including defect detection in the floors of large tanks and oil pipelines. However, in addition to noise, practical scenarios of signal detection also present mode aliasing problems, which make it difficult to accurately identify and locate defects. In this paper, we propose an improved shear horizontal guided wave mode identification method using a variational mode decomposition (VMD) algorithm and a time-of-flight (TOF) extraction method using the synchrosqueezed wavelet transform (SSWT) algorithm. Moreover, we use the wavelet denoising method to denoise the original signal before applying VMD. The results show that the TOF errors obtained by the VMD method are all less than 5% and that the wavelet denoising of the original guided wave data can further reduce the errors (to less than 2%). In addition, the SSWT can modify the time-frequency analysis results of intrinsic mode functions obtained by the VMD method and provides accurate TOF data for different modes. Therefore, the proposed ultrasonic guided wave signal method is helpful for improving the defect detection sensitivity and accuracy of SH waves.

Concrete structures have emerged as some of the most extensively utilized materials in the construction industry due to their inherent plasticity and high-strength characteristics. However, due to the temperature fluctuations, humidity, and damage caused by human activities, challenges such as crack propagation and structural failures pose threats to the safety of people’s lives and property. Meanwhile, conventional non-destructive testing methods are limited to defect detection and lack the capability to provide real-time monitoring and evaluating of concrete structural stability. Consequently, there is a growing emphasis on the development of effective techniques for monitoring the health of concrete structures, facilitating prompt repairs and mitigation of potential instabilities. This paper comprehensively presents traditional and novel methods for concrete structural properties and damage evolution monitoring, including emission techniques, electrical resistivity monitoring, electromagnetic radiation method, piezoelectric transducers, ultrasonic techniques, and the infrared thermography approach. Moreover, the fundamental principles, advantages, limitations, similarities and differences of each monitoring technique are extensively discussed, along with future research directions. Each method has its suitable monitoring scenarios, and in practical applications, several methods are often combined to achieve better monitoring results. The outcomes of this research provide valuable technical insights for future studies and advancements in the field of concrete structural health monitoring.