What are the aspects of artificial intelligence in microwave nondestructive testing technology, and what are the application prospects?
Non-destructive testing (NDT) is defined as the practice of evaluating variations in various properties of a material, such as delamination, corrosion, cracks, and fatigue, including internal defects or metallurgical conditions, without in any way interfering with the integrity of the material or its suitability for service.
Commonly used NDT methods are wide range and well established in the industry, such as thermography, ultrasonic inspection, eddy current inspection, X-ray and magnetic particle inspection, and the choice between these technologies is based on their advantages and disadvantages, taking into account safety, operating costs, and their efficiency on the material being inspected.
Ultrasonic testing cannot penetrate highly porous materials and requires a lot of data interpretation. While laser ultrasound performs non-contact detection without couplant, its performance degrades when inspecting highly porous materials.
Laser ultrasound simultaneously generates various waves (such as shear waves, longitudinal waves, Rayleigh waves, and Lamb waves), which complicates signal analysis, eddy current detection is widely used in industry for surface crack and corrosion detection, but this method is not effective when detecting low-loss dielectric materials due to the penetration limitation of low-frequency electromagnetic waves.
Traditional thermal imaging procedures are based on infrared (IR) radiation, using an infrared camera to observe and capture the radiation emitted by the test material to represent the temperature distribution of the test material in a visible image, infrared thermography (IRT) is divided into passive thermography, passive IRT depends on the temperature change between two media that are not in thermal equilibrium.
Active IRT relies on an external excitation source, such as a flash, ultrasound, or laser, to change the temperature of the inspected material, and the defect region will produce a different temperature distribution than its adjacent area, resulting in the shape of the defect.
Due to its fast inspection speed, high imaging resolution and high defect detection sensitivity, thermal imaging is considered one of the most widely used techniques in non-destructive testing, and IRT technology is usually limited to active technology in NDT.
Pulsed thermography (PT) is used to examine CFRP composites in active mode. In the first phase of the experiment, the CFRP specimen is pulsed heated using two Balcar xenon flash lamps while the specimen is observed through an infrared camera.
If the specimen is free of defects, heat will be transferred to the specimen, the temperature distribution will drop evenly, if there is a defect, the uniformity of the temperature distribution will be affected, and this change will be recorded using an infrared camera for further processing.
During the inspection, the camera is cooled to 196.15°C to mitigate the effect of the incoming temperature on the thermal detector, using PT for two types of inspections; Reflection and transmission, reflection inspection is achieved by positioning the infrared camera and excitation source in the same direction as the sample being examined.
On the other hand, when the sample to be inspected is between the infrared camera and the heat source, both technologies can effectively check for defects in locating the inspected material, however, a controlled inspection environment is required, such as controlled ambient temperature and infrared camera cooling, and the static position of the infrared camera is required to measure the temperature amount of the inspected material at a specific time.
In the second phase of the experiment, ultrasonic waves are used as an excitation source, called ultrasonic infrared thermography (UIT), ultrasonic transducers excite the sample under test at a frequency of 20kHz, mechanical sound waves are propagated through the surface, and if the surface is defective, friction is generated through cracks.
As a result, friction increases the temperature around the crack area, which will be recorded by the infrared camera, and the advantage of this technique is that the crack itself is a heat source, so UIT successfully locates CFRP defects that are more pronounced than PT results.
Should strictly select the appropriate IRT excitation source to avoid damage to the tested material, in fact, the excitation source changes the temperature of the measured medium, should first study the ability of the medium to be excited, need to perfectly control the detection environment, so that the temperature is evenly distributed on the tested material, due to the change of ambient temperature, the temperature is difficult to control, especially for the detection of in-service materials.
Laser thermography (LT) is used to inspect glass laminated reinforced epoxy resin (GLARE), a complex structure widely used in aerospace applications that applies a matte black paint to the surface of the sample being tested to unify infrared emissivity.
The moving beam of the laser is focused with a spot of 1.5 mm for heating the material to be tested, the infrared camera is used to obtain the temperature distribution of the sample to be tested, if there is a defect such as delamination, the surface temperature of the defect area will change, and the standard division (SD) is used for post-processing of the defect representation.
SD is applied to the temperature distribution of the current laser spot, the reference area follows the point on the left, the reference region is involved because the position of the internal defects of the composite does not correspond to the laser spot, the technology shows very good delamination detection ability, small size delamination is difficult to evaluate, because it causes small perturbations on the difficult-to-observe reference area.
The trend of using MNDT to detect these structures is emerging because of the advantages of microwave-based techniques in addressing the challenges of traditional technologies, which still have many limitations, such as blurry spatial images and large computational intensity.
OERW-based techniques face several challenges, such as distance variation, optimal frequency points, and poor image quality, as researchers focus their research on sensor-based enhancement and lack the implementation of soft computing techniques in MNDT applications.
In traditional MNDT technology, OERW is widely used for non-destructive testing and has shown promising results in defect detection, localization and depth estimation of various materials such as metals, CFRP, GFRP, TBC, and dielectric components.
