Simurgh: AI-Powered Framework for Fast and Accurate Industrial X-ray Computed Tomography

Invention Reference Number

90000193

High-resolution X-ray computed tomography (XCT) is an important technique for the inspection of industrial components. XCT is typically used off-line to inspect a subset of manufactured parts and with substantial cost and labor. Accelerating measurement speed while retaining accuracy could not only allow for inspection of a significantly greater number of parts at a lower cost and labor, but also enables using XCT for in-line inspection to rapidly identify defects in each part as it is manufactured. ORNL researchers have developed a deep learning (DL) based approach to rapidly perform high-quality reconstructions from sparse (i.e. very fast) XCT measurements. This approach uses available digital model (e.g. computer-aided design or CAD) of the parts along with physics-based information and generative adversarial neural network (GAN) to produce accurate 3D reconstructions. As an important area of study in additive manufacturing (AM) inspection, and using experimental XCT data of metal parts, ORNL has demonstrated enhanced defect detection capabilities while dramatically reducing the scan time. Recently, Simurgh was among the finalists for the prestigious R&D 100 award.

Benefits

Increased speed: Delivers scan times of >12 times faster than traditional methods.
Enhanced accuracy: Improves defect detection capabilities by four times. Simurgh has reliably demonstrated that it can identify flaws as small as 50 µm–100 µm depending on material and related standard high energy industrial X-ray system capabilities with significantly faster scan times.
Cost reduction: Lowers operational costs by reducing scan times, the need for extensive post-processing and manual input, and associated labor costs.
Versatility: Applicable across various industries including aerospace, automotive, biomedical, electronics, and advanced manufacturing.
Scalability: Facilitates high-throughput evaluation and testing, making it ideal for Industry 4.0 applications.

Applications and Industries

Additive manufacturing (AM): Qualifies complex AM parts by providing rapid, high-resolution CT scans to detect flaws and ensure part integrity.
Aerospace and automotive: Enhances safety and performance through accurate defect detection in critical components.
Biomedical devices: Improves the quality assurance of medical implants and devices, ensuring patient safety.
Energy industry: Assists in the evaluation of components used in energy production, enhancing reliability and efficiency.
Casting: Identifies defects in casted parts, crucial for automotive and industrial quality.
Electronics Manufacturing: Verifies integrity of electronic components and assemblies, focusing on multilayer structures and solder joints.