China
Africa
Explore chapters and articles related to this topic
Nondestructive Evaluation (NDE) of Materials and Structures from Production to Retirement
Published in Yoseph Bar-Cohen, Advances in Manufacturing and Processing of Materials and Structures, 2018
The oldest and the most commonly used NDT method is visual examination or visual testing (VT). The VT technique has the longest history in comparison to other NDE techniques like radiographic testing (RT), MFL, ET, and ultrasonic testing (UT)-based techniques. The visual inspection technique is popular because it is quick, less expensive, and often very effective. Almost 80% of industrial defects are detected and located by visual inspection. Visual inspection can be direct viewing, using line-of sight vision, or may be enhanced with the use of optical instruments such as magnifying glasses, mirrors, boroscopes, digital cameras, charge-coupled devices, and computer-assisted viewing systems (remote viewing). Corrosion, misalignment of parts, physical damage, and cracks are just some examples of conditions and discontinuities that may be detected by visual examinations (Newman, 1995; Pernkopf and O’Leary, 2003). VT is essentially a surface technique and is limited by the skill, expertise, and power of the examiner and the associated instrument or tools used by the examiner. However, uncertainty associated with the visual inspection technique is one of its primary limitations. The reliability of this technique depends on the minimum allowable defect size, among other factors. The primary disadvantage of VT is that very small defects and internal defects in an opaque object cannot be detected by this technique (Smith and Adendorff, 1991).
Product Quality
Published in G.K. Awari, C.S. Thorat, Vishwjeet Ambade, D.P. Kothari, Additive Manufacturing and 3D Printing Technology, 2021
G.K. Awari, C.S. Thorat, Vishwjeet Ambade, D.P. Kothari
Figure 10.9 identifies nondestructive test methods being applied to AM parts. Visual inspection is often employed to identify gross defects such as distortions, surface conditions, and gross anomalies. Camera and image inspection can accurately and rapidly capture and classify part features and compare them with a standard definition or part model. Cameras may also utilize magnified views of regions to verify conditions against a standard or population of parts. Bore scope inspection may be incorporated to verify clearances or deposit conditions within enclosed volumes or passageways. Multiple image cameras combined with software are capable of measuring distances and other characteristics of AM parts.
Coating Defects and Inspection
Published in Karan Sotoodeh, Coating Application for Piping, Valves and Actuators in Offshore Oil and Gas Industry, 2023
The first group of inspection tools are used for visual inspection. Visual inspection tools typically include a mirror, flashlight and magnifier. These handy tools enable the inspector to detect defects visually and verify the cleanliness and roughness of the surface. Figure 4.22 illustrates a mirror that is used to check the surface and identify any welding defects on the metal surface prior to applying coating. Figure 4.23 shows a visual inspection of welding using a magnifier. Some welding defects, like spatter and porosity, should be removed from the metal surface before applying the coating.
Fault diagnosis of visual faults in photovoltaic modules: A Review
Published in International Journal of Green Energy, 2021
Naveen Venkatesh S, V Sugumaran
The initial step for identifying faults in a PVM is through inspection by plain sight. Visual inspection is a simple method to identify some failures and defects. It also helps the observer to monitor external stresses and gives a proper insight into the PVM condition. This inspection is necessary to know the state of modules before installation and to monitor its performance after long time operation in the field where it is exposed to thermal stresses and environmental factors. Visual inspection is carried out based on standards given by International Electrotechnical Commission (IEC) which states that 1000 lux illumination is required during an investigation and faults visible to bare eyes are considered when inspected at different angles. The visual inspection helps in collecting a large amount of data which consists of a checklist for each panel; however, it has some drawbacks as it is time consuming and requires a lot of manpower (Aghaei et al. 2015). In recent times unmanned aerial vehicles (UAV) have been put into use in visual inspection to reduce the drawbacks listed earlier. UAV installed with onboard sensors and digital cameras are used to capture images and provide data to the ground control unit for further detection of visible faults (Li et al. 2017). Table 3 depicts typical faults that can be detected using visual inspection in various components of a PVM.
Structural fault diagnosis of UAV based on convolutional neural network and data processing technology
Published in Nondestructive Testing and Evaluation, 2023
Yumeng Ma, Faizal Mustapha, Mohamad Ridzwan Ishak, Sharafiz Abdul Rahim, Mazli Mustapha
Historically, visual inspection techniques have been a mainstay of early damage detection systems. However, visual inspection is limited in its ability to identify only apparent faults and is considered a local method. To overcome these limitations, a model-based method has been proposed for detecting sensor faults. In recent years, the emergence of data acquisition technologies and advancements in computing capacity have enabled data-driven methods, such as machine learning and deep learning methods, to become the mainstream in damage detection and identification. Data-driven methods can be [17] categorised into two main types: unsupervised learning methods and supervised learning methods. Supervised learning methods, such as convolutional neural networks (CNNs), require both normal and abnormal data for training. However, unsupervised learning methods, such as support vector machines (SVM) and K-nearest neighbour (KNN), only utilise normal data for training. Over the past few decades, unsupervised learning methods have been widely used for damage detection. For example, Cha proposed a density-peak-based fast clustering algorithm (DPFCA) for structural damage detection and localisation, which was evaluated on a laboratory-scale steel structure under different damage scenarios. The results indicate a satisfactory performance [18]. In another study, an unsupervised deep learning method based on a deep auto-encoder with a one-class support vector machine was proposed to detect damage using vibration signals. The performance of this method was evaluated on a 12-story numerical building model and a laboratory-scaled steel bridge, and it was found to achieve high accuracy compared to state-of-the-art methods [19].