Automatic Fault Research Articles

EXplainable artificial intelligence for automatic defect detection in additively manufactured parts using CT scan analysis

AbstractAdditive Manufacturing (AM) and in particular has gained significant attention due to its capability to produce complex geometries using various materials, resulting in cost and mass reduction per part. However, metal AM parts often contain internal defects inherent to the manufacturing process. Non-Destructive Testing (NDT), particularly Computed Tomography (CT), is commonly employed for defect analysis. Today adopted standard inspection techniques are costly and time-consuming, therefore an automatic approach is needed. This paper presents a novel eXplainable Artificial Intelligence (XAI) methodology for defect detection and characterization. To classify pixel data from CT images as pores or inclusions, the proposed method utilizes Support Vector Machine (SVM), a supervised machine learning algorithm, trained with an Area Under the Curve (AUC) of 0.94. Density-Based Spatial Clustering with the Application of Noise (DBSCAN) is subsequently applied to cluster the identified pixels into separate defects, and finally, a convex hull is employed to characterize the identified clusters based on their size and shape. The effectiveness of the methodology is evaluated on Ti6Al4V specimens, comparing the results obtained from manual inspection and the ML-based approach with the guidance of a domain expert. This work establishes a foundation for automated defect detection, highlighting the crucial role of XAI in ensuring trust in NDT, thereby offering new possibilities for the evaluation of AM components.

Parametric study of anomaly detection models for defect detection in infrared thermography

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.

Research progress on bearing fault diagnosis with signal processing methods for rolling element bearings

As a prerequisite for rotating machinery to operate effectively, rolling element bearings play an essential role. The focus of condition monitoring has initially been on defect identification, then on its measurement, and eventually on automatic defect prediction. The improvement in signal processing has made this breakthrough possible. The quality of characteristics taken from the bearing signals strongly impacts how effective these techniques are. Aiming to provide the researchers with the option to choose and implement the optimum signal analysis method, the authors have described numerous signal processing techniques used to diagnose faults in rolling element bearings. The research study examines several important studies and explains their relevance to locating rolling bearing defects. It analyzed recent research, ones from the past, and developments in the field of diagnosing bearing defects. The main goal of the research is to investigate different vibration signal processing and analysis methods for locating and evaluating bearing faults. After that, each of these subjects is rigorously analyzed in order to draw conclusions, spot new trends, and pinpoint areas that still need more research. This article is meant to serve as a guide for those who operate in the condition monitoring domain.

Deep Learning-Based Automatic Defect Detection of Additive Manufactured Stainless Steel

Accumulating interest from academia and industry, the part of quality assurance in metal additive manufacturing (AM) is achieving incremental recognition owing to its distinct advantages over conventional manufacturing methods. In this paper, we introduced a convolutional neural network, YOLOv8 approach toward robust metallographic image quality inspection. Metallographic images accommodate key information relating to metal properties, such as structural strength, ductility, toughness, and defects, which are employed to select suitable materials for multiple engineering execution. Therefore, by comprehending the microstructures, one can understand insights into the behavior of a metal component and make predictive assessments of failure under specific conditions. Deep learning-based image segmentation is a robust technique for the detection of microstructural defects like cracks, inclusion, and gas porosity. Therefore, we improvise the YOLOv8 with dilated convolution mechanisms to acquire automatic micro-structure defect characterization. More specifically, for the first time, the YOLOv8 algorithm was proposed in the metallography dataset from additive manufacturing of steels (Metal DAM) to identify defects like cracks and porosity as a novel approach. A total of 414 images from ArcelorMittal engineers were used as an open-access database. The experimental results demonstrated that the YOLOv8 model successfully detected and identified cracks and porosity in the metal AM dataset, achieving an improved defect detection accuracy of up to 96% within just 0.5 h compared to previous automatic defect recognition processes.

Skin Imaging: A Digital Twin for Geometric Deviations on Manufactured Surfaces

Closed-loop manufacturing is crucial in Industry 4.0, since it provides an online detection–correction cycle to optimize the production line by using the live data provided from the product being manufactured. By integrating the inspection system and manufacturing processes, the production line achieves a new level of accuracy and savings on costs. This is far more crucial than only inspecting the finished product as an accepted or rejected part. Modeling the actual surface of the workpiece in production, including the manufacturing errors, enables the potential to process the provided live data and give feedback to production planning. Recently introduced “skin imaging” methodology can generate 2D images as a comprehensive digital twin for geometric deviations on any scanned 3D surface including analytical geometries and sculptured surfaces. Skin-Image has been addressed as a novel methodology for continuous representation of unorganized discrete 3D points, by which the geometric deviation on the surface is shown using image intensity. Skin-Image can be readily used in online surface inspection for automatic and precise 3D defect segmentation and characterization. It also facilitates search-guided sampling strategies. This paper presents the implementation of skin imaging for primary engineering surfaces. The results, supported by several industrial case studies, show high efficiency of skin imaging in providing models of the real manufactured surfaces.

SPEED:Semantic Prior and Extremely Efficient Dilated Convolution Network for Real-Time Metal Surface Defects Detection

Automatic defect detection on the metal surface is a vital task for product inspection in industrial assembly lines or production processes. Owing to miscellaneous patterns of defects, inter-class similarity, intra-class difference, and fewer defect samples, achieving accurate and automatic detection remains a big challenge. What's more, since the rising demand for production efficiency, real-time detection is increasingly desirable. This article proposes a semantic prior and extremely efficient dilated convolution network, named SPEED, for pixel-wise detection on the metal surface, which aims to address the aforementioned issues. The architecture of SPEED involves the following: (i) a semantic prior (SP) branch, with shallow layer and prior mapping module to capture low-level details; and (ii) an extremely efficient dilation (EED) branch, with lightweight bottleneck to obtain high-level context. Furthermore, an aggregation module is designed to fuse both types of feature representation. Additionally, different level features of bottleneck are fused to improve the segmentation performance. Experimental results on three metal surface defect datasets indicate that the proposed method outperforms the state-of-the-art approaches in terms of the mean intersection of union, model parameters, FLOPs, and FPS. More specifically, SPEED achieves 92.34% mIoU on NEU-Seg, 88.65% mIoU on Severstal Strip Steel, and 63.91% mIoU on MT Defect.

Automatic histogram-based defect detection in glass fibre reinforced polymer composites using terahertz time-domain spectroscopy reflection imaging

We developed a method for automatic detection of defects in glass fibre reinforced polymer (GFRP) based on a terahertz time-domain spectroscopy (TDS) system. Two GFRP specimens with round defects of different sizes (inclusions and delaminations) were measured by means of a high-speed TDS reflective scanner with an electronically controlled optical sampling scheme. The algorithm provided a series of C-scans, which were then pre-selected to identify only those with features potentially associated with defects. A thresholding method based on Gaussian fitting of the histogram was developed. The results showed that 87 % of defects were detected for the specimen with inclusions and 100 % for the specimen with delaminations. The algorithm also provided information about the depth of defects, as each C-scan was assigned to a time slice. The results demonstrate that the TDS-based system, along with the proposed histogram-based method, can automatically detect defects in GFRP. © 2017 Elsevier Inc. All rights reserved.

Barriers and drivers for implementation of automatic fault detection and diagnosis in buildings and HVAC systems: An outlook from industry experts

This study aimed to assess the current status of Fault Detection and Diagnosis (FDD) implementation in building and Heating, Ventilation and Air Conditioning (HVAC) systems in the building industry. Semi-structured qualitative interviews were conducted with 29 experts from different HVAC company types in the building industry. In addition, a literature review was performed to investigate academic research on FDD implementation. The study identified barriers and drivers to implementing FDD systems, these included; technological and technical, economic and business, users, social and societal, and regulatory. An Automatic Fault Detection and Diagnosis (AFDD) implementation matrix was developed to evaluate FDD implementation in building systems, and all interviewed companies were classified based on their FDD knowledge, services, and type. Results show that expert-rule systems are still prevalent in the industry. The literature review revealed a scarcity of FDD implementation studies in academic research due to challenges in testing and validating results in actual building operation conditions. Lastly, this study discusses the key findings: 1) FDD does not sell, 2) Lack of actively engaging and promoting FDD services, 3) FDD seems to be an academic definition, 4) The bottlenecks: The fault handling process and user’s mindset towards FDD, and 5) Governmental regulations and legislatives drive the implementation focus.

Thetis: A Booster for Building Safer Systems Using the Rust Programming Language

Rust is a new system-level programming language that prioritizes performance, safety, and productivity. However, as evidenced in many previous works, unsafe code fragments broadly exist in Rust projects. The use of these unsafe fragments can fundamentally violate the safety of systems developed using the programming language. In response to this problem, we propose a novel methodology (Thetis) to enhance the safety capability of Rust. The core idea of Thetis is to reduce unsafe code, encapsulate unsafe code using safety rules, and make it easier to verify unsafe code through formal means. The proposed methodology involves three main components. In the context of Rust itself, Thetis combines replacement and encapsulation for Interior Unsafe segments, minimizing unsafe fragments and reducing unsafe operations and their range. For systems developed using Rust, new ACSL formal statutes are applied to reduce the unsafe potential of the encapsulated Interior Unsafe segments, enhancing the safety of the system. Regarding the development life cycle in Rust, Thetis introduces automatic defect detection and optimization based on feature extraction, improving engineering efficiency. We demonstrate the effectiveness of Thetis by using it to fix defects in BlogOS and ArceOS. The experimental results reveal that Thetis reduces the number of unsafe operations in these OSs by 40% and 45%, respectively. The use of Miri to detect and eliminate defects in ArceOS reduces the likelihood of undefined behavior by about 50%, which effectively demonstrates that the proposed method can improve the safety of the Rust system. In addition, performance test results from LMbench show that the performance loss caused by Thetis is only 1.076%, thereby maintaining the high-performance characteristics of the Rust system.

Blockchain-based trusted accountability in the maintenance of medical imaging equipment

Preventative maintenance is imperative for ensuring the longevity and reliability of diagnostic medical imaging equipment. However, conventional maintenance systems face challenges related to centralization, fragmented information, and ineffective tracking mechanisms. In response to these challenges, this research presents a blockchain-based solution that introduces a robust framework for managing preventative maintenance tasks. The proposed solution leverages blockchain technology, smart contracts, IPFS, and timer oracles to promote transparency, accountability, and traceability in maintenance operations. To ensure the integrity and availability of maintenance data while preserving the confidentiality of sensitive information, a detailed system architecture is outlined. Sequence diagrams and algorithms are developed to define the flow and logic of the solution, focusing on automatic fault detection and diagnosis. The smart contracts, which are made publicly available on GitHub, were rigorously tested to guarantee their intended functionality, thus addressing software reliability concerns. Additionally, the solution includes a maintenance incentive mechanism, actively motivating stakeholders to engage in preventative maintenance activities, and establishing accountability. Security testing confirms the absence of vulnerabilities in smart contracts, aligning with the scope of automatic fault detection and diagnosis. Furthermore, potential generalizations of the solution are discussed, and future directions emphasize the importance of continuous improvement and innovation in maintenance practices.

Automatic Identification Method of Defects in Concrete Structures Strengthened with Composite Materials Based on a Multi-Scale Model

Due to the inconsistent geometric characteristics of reinforced concrete structures with composite materials, some echo signals have a lower degree of clarity, resulting in problems such as long time, small range, and low accuracy in identifying structural defects. Therefore, in order to improve the accuracy and efficiency of defect identification in composite material reinforced concrete structures, a multi-scale model based on the automatic defect identification method for composite material reinforced concrete structures is proposed. Using the interface connection method, based on continuous distribution coupling and the uniform weighting coefficient calculation method to weigh the interface nodes, we determined the interface connection mode and analyzed the stress–strain relationship. The impact echo method was used to analyze the degree of high-frequency stress wave scattering caused by internal non-uniformity in composite reinforced concrete, and the fast Fourier transform conversion algorithm was used to generate amplitude–frequency curves and resolved thickness or defect depth. The experimental results show that the multi-scale finite element simulation structure of this method is clear and accurate, and although there are errors, they do not affect the simulation results of the structure. The impact echo detection effect is obvious; it can automatically identify defects in composite reinforced concrete structures and quickly and accurately identify defects in different states and positions of composite reinforced concrete, with a recognition accuracy of 98%.

An effective torque-based method for automatic turn fault detection and turn fault severity classification in permanent magnet synchronous motor

An effective torque-based method for automatic turn fault detection and turn fault severity classification in permanent magnet synchronous motor

Method of automatic fault location in 6-35 kV electrical networks

The paper analyses the known means of fault location in electrical networks, identifies their disadvantages and proposes its own method of solving this problem. The model of electric network is built and investigated by MATLAB.

Automatic seismic fault identification based on an improved U-Net network

Automatic seismic fault identification based on an improved U-Net network

Contrastive learning and dynamics embedding neural network for label-free interpretable machine fault diagnosis

Industrial machinery often produces vibration signals that can serve as indicators of underlying faults. However, these signals often need to be labeled, presenting a challenge for accurate and interpretable fault diagnosis. While supervised learning methods, such as deep neural networks, have been applied for fault diagnosis, they need help in effectively distinguishing between different vibration-related faults. In response to this issue, our study introduces an innovative approach for automatic fault diagnosis through the application of the Bootstrap Your Own Latent and Dynamical Systems Model Discovery algorithm (BYOLDIS). This method not only addresses the challenge of unlabelled signals but also provides readily interpretable results. The proposed methodology consists of three fundamental steps. First, we derive a matrix of differential equations to capture the dynamic behavior of faulty bearings. Second, we employ a contrastive learning network alongside a time-delay embedding matrix to reconstruct the coordinates of the fault-dynamical system. Lastly, we construct a library of fault machine dynamic polynomial equations, incorporating prior constraints based on physical models. To assess the effectiveness and robustness of our proposed method, we conducted both simulations and experiments. The results of these case studies affirm that BYOLDIS can accurately diagnose bearing faults and offer dynamic explanations for the diagnostic outcomes. This suggests that BYOLDIS holds substantial promise as a diagnostic tool for processing unlabelled vibrational data.

Defect detection of glassivation passivation parts wafer surface with random texture and different brightness

Automatic defect detection for glassivation passivation parts (GPP) wafer surface becomes an extremely challenging task, due to the interference of random texture, disturbance of low-contrast pseudo defects within the image, and difference of different brightness between images. In this paper, we propose a novel defect detection scheme for GPP wafer surface with random texture and different brightness. First, an automatic segmentation method center expansion idea-based for the region of interest (ROI) of die image is presented to eliminate the interference from edge background and improve the efficiency of defect detection. Then, a feature point set extraction method based on adaptive gain and error diffusion is proposed so that the defect feature between images with different brightness can be highlighted by adaptive gain, and the candidate defect feature point set under random texture feature is able to be extracted by error diffusion. Finally, the density-based spatial clustering of applications with noise considering grayscale constraint (GC-DBSCAN) is designed to identify the true defect clusters from the candidate defect feature point set and accomplish the detection of various types of surface defects. Experimental results show that the proposed method can completely implement the extraction of crack and dirty defects and eliminate the false detection caused by random texture and different brightness, which is very efficient and superior to other methods.

A lightweight network for photovoltaic cell defect detection in electroluminescence images based on neural architecture search and knowledge distillation

Nowadays, the rapid development of photovoltaic(PV) power stations requires increasingly reliable maintenance and fault diagnosis of PV modules in the field. Due to the effectiveness, convolutional neural network (CNN) has been widely used in the existing automatic defect detection of PV cells. However, the parameters of these CNN-based models are very large, which require stringent hardware resources and it is difficult to be applied in actual industrial projects. To solve these problems, we propose a novel lightweight high-performance model for automatic defect detection of PV cells in electroluminescence(EL) images based on neural architecture search and knowledge distillation. To auto-design an effective lightweight model, we introduce neural architecture search to the field of PV cell defect classification for the first time. Since the defect can be any size, we design a proper search structure of network to better exploit the multi-scale characteristic. To improve the overall performance of the searched lightweight model, we further transfer the knowledge learned by the existing pre-trained large-scale model based on knowledge distillation. Different kinds of knowledge are exploited and transferred, including attention information, feature information, logit information and task-oriented information. Experiments have demonstrated that the proposed model achieves the state-of-the-art performance on the public PV cell dataset of EL images under online data augmentation with accuracy of 91.74% and the parameters of 1.85M. The proposed lightweight high-performance model can be easily deployed to the end devices of the actual industrial projects and retain the accuracy.

Intelligent fault diagnosis for air handing units based on improved generative adversarial network and deep reinforcement learning

Data-driven Automatic fault detection and diagnosis (AFDD) for air handling units (AHUs) is crucial for ensuring the stable operation and energy consumption of the heating ventilation air-conditioning (HVAC) system. However, traditional machine learning methods often underperform when confronted with insufficient training sample data, especially when lacking samples from the fault types. Based on the issues of insufficient samples from the fault types and imbalanced training dataset, this study proposes a novel AFDD approach using transformer integrated conditional Wasserstein generative adversarial network and deep reinforcement learning (TCWGAN-DRL) to synthesize the fault data and select high quality synthetic data samples. Firstly, we utilize the proposed TransCWGAN to synthesize fault samples. Then, reinforcement learning is utilized to select high quality synthetic samples. Finally, the filtered samples and the real fault samples are merged to form the training dataset for conventional supervised learning classifiers. Experimental results demonstrate that the enriched training dataset can effectively improve the AFDD results and outperforms recently published existing methods, for instance, compared to the suboptimal model, our method exhibits an increase in fault recognition accuracy of 4.9%, 3.66%, and 4.02% when the number of real fault samples is 15, 20, and 30, respectively.

Novel MobileNet based Multipath Convolutional Neural Network for defect detection in fabrics

Automatic fabric defect detection and classification is the most important process in the textile industry to ensure the fabric quality. In the existing systems, a learning based method is used for detecting defects in plain weave fabrics. In this paper, a novel MobileNet based Multipath Convolutional Neural Network (MMPCNN) architecture is proposed for detection and classification of simple and complex patterned fabric defects. In the proposed MMPCNN architecture, MobileNet model is used in the first path. In this, Gabor filter bank is used instead of conventional filters in the first convolution layer. A simple convolutional neural network architecture with Gray Level Co-occurrence Matrix (GLCM) features as an input is used in the second path of the MMPCNN architecture. Gabor filters are more useful for analyzing the texture with different orientations and scales. Each Gabor filter parameter has its own impact on analyzing the texture and extracting the information from the texture. Therefore, in this paper, the use of Gabor filter parameters in MMPCNN architecture is analyzed. The proposed model is experimented on the TILDA textile image database and it is able to achieve 100% accuracy with reduced trainable parameters for fabric defect detection and classification.

An automated defect detection method for optimizing industrial quality inspection

Detecting defects in industrial-quality inspection is an important task, but defect detection remains challenging due to limited size, various types of defects, and imbalanced samples in images. In this paper, we propose a novel automatic defect detection network that efficiently detects defects. We introduce the Automated Sample Assignment (SCID) algorithm, which automatically divides the distribution of positive and negative samples based on similarity scores using the Gaussian mixture model. The Automated Sample Assignment (SCID) algorithm solves the problem of sample imbalance in defect detection networks caused by manually setting the intersection over the union (IOU) threshold and assigning positive and negative labels to samples. Then, instead of the pooling method used by most backbone networks, we use a loss fusion pool named LE-Pool to retain more detailed information. Finally, we use the soft weighted attention mechanism of the detection head (S-head) to identify noise information. Experimental results on multiple defect detection datasets show that the proposed method is superior to other detection methods in terms of accuracy.