Emerging technologies in non-destructive testing (NDT): 8 innovations to know
Head of Innovation
Digital Transformation, AI, IoT, Machine Learning, and Cloud Technologies. Nearly 15 years driving innovation
The emerging technologies in non-destructive testing (NDT) that matter most are phased-array ultrasonics with the total focusing method, digital radiography and industrial CT, guided waves and permanent sensors, non-contact methods such as laser ultrasonics, robotic and drone inspection, AI-assisted defect recognition, and connected data with digital twins. Together they move NDT from periodic checks to continuous, data-driven inspection.
Key takeaways
- Phased-array ultrasonic testing (PAUT) and the total focusing method (TFM) give sharper images of welds and composites than conventional single-element ultrasound.
- Digital radiography and industrial CT replace film, give instant results and show internal defects in 3D, which is essential for additively manufactured parts.
- Robots, crawlers and drones take inspectors out of confined spaces and off rope access, and they repeat the same scan path every time.
- AI works best as a first-pass reviewer that flags indications for a certified inspector, not as a replacement for one.
- Standards such as ASTM E2339 (DICONDE) for data and ISO 9712 for personnel certification decide whether new methods are accepted in regulated work.
What is changing in non-destructive testing?
NDT is shifting from manual, periodic inspection with paper or film records to automated, digital and often continuous inspection. The physics (sound, radiation, electromagnetism, heat, light) stays the same; the sensors, automation and data handling change.
Three pressures drive this. Ageing infrastructure in energy, process industry and transport needs more inspection with fewer qualified people. New materials such as composites and 3D-printed metal parts have defects that older methods struggle to find. And asset owners want inspection results in the same systems as maintenance and production data, so they can plan instead of react.
1. Phased-array ultrasonics and the total focusing method
Phased-array ultrasonic testing uses a probe with many small elements that can be fired in sequence, steering and focusing the beam electronically. One probe position produces a cross-sectional image instead of a single A-scan trace, which speeds up weld and corrosion inspection.
The total focusing method goes further. The instrument records every transmit-receive combination (full matrix capture), then focuses at every point in the region of interest during processing. The result is a sharper image and better sizing of small or awkwardly oriented flaws. Codes have followed: ISO 23864 covers automated TFM for weld testing, and ASTM E2491 is the guide for evaluating the performance of phased-array instruments and systems.
Trade-off: TFM data sets are large and processing is heavy, so instruments, storage and analysis software need planning before you scale up.
2. Digital radiography and industrial computed tomography
Film radiography is giving way to computed radiography (reusable imaging plates) and digital detector arrays. Images are available in seconds, can be enhanced and measured on screen, and are stored digitally without chemical processing.
Industrial CT takes hundreds or thousands of projections and reconstructs the part in 3D. It reveals porosity, inclusions and internal geometry that no surface method can reach. For additive manufacturing it has become a standard tool: ASTM E3166 is a guide to NDE of metal additively manufactured aerospace parts after build, and CT is one of the methods it covers.
Trade-off: CT is slow per part and limited by part size and density. Most plants use it for first-article inspection and sampling rather than 100% inspection.
3. Guided waves and permanently installed sensors
Guided-wave ultrasonics send low-frequency waves along a pipe from one collar position, screening tens of metres for wall loss in one test. It is used on pipe racks, road crossings and insulated lines where full access is expensive.
The bigger change is permanent monitoring. Ultrasonic thickness sensors bonded to the pipe, fibre-optic strain and temperature sensing, acoustic emission sensors and low-cost MEMS vibration sensors send readings continuously. Inspection becomes a trend line rather than a snapshot every few years. This is where NDT meets IoT predictive maintenance: corrosion rates and crack growth can be tracked and acted on before a limit is reached.
4. Non-contact and near-contact methods
Several methods inspect without couplant or physical contact, which helps on hot, moving, delicate or complex parts:
- Laser ultrasonics: one laser generates ultrasound in the surface and another detects it, so you can inspect hot or curved parts from a distance.
- Electromagnetic acoustic transducers (EMATs): generate ultrasound directly in conductive materials without couplant, useful for hot steel and high-speed lines.
- Eddy current arrays: many coils in one probe cover a wider area per pass with a C-scan style image, used for surface cracks and corrosion in aerospace and tubing.
- Active thermography: heats the surface with a flash or lamp and watches how heat flows, which shows delaminations and disbonds in composites.
- Terahertz imaging: inspects non-conductive materials such as foams, coatings and some composites.
5. Robotics, crawlers and drones
Robotic platforms take inspectors out of hazardous places and make scans repeatable:
- Drones carry high-resolution and thermal cameras to wind turbine blades, flare stacks, bridges and roofs.
- Magnetic crawlers move along tank walls, vessel shells and pipelines carrying ultrasonic or eddy current probes.
- Remotely operated vehicles inspect subsea structures and flooded tanks.
- Robot arms in factories follow programmed scan paths around complex parts, so every part is inspected the same way.
Repeatability matters as much as safety. When the scan path, probe angle and settings are identical each time, changes in the data reflect changes in the asset, not in the inspector.
6. AI-assisted defect recognition
Machine learning models, mostly convolutional neural networks, are now trained to find and classify indications in radiographs, ultrasonic C-scans, eddy current data and visual images. They review large volumes of data quickly and consistently.
The sound way to use them today is as a first pass. The model flags likely indications and the certified inspector confirms, rejects or sizes them. This keeps the inspection within code requirements and gives the inspector more time for difficult calls. Making it work depends on:
- Data quality: consistent acquisition settings and calibration across equipment and sites.
- Labelled data: enough confirmed examples of each defect type, including rare ones.
- Validation: blind tests on known defects and probability of detection (POD) studies before anyone relies on the model.
- Change control: every model update is versioned and re-validated, like any other inspection procedure change.
The same approach powers automated visual inspection on production lines, where cameras and computer vision check surfaces and assemblies at line speed.
7. Connected data, digital twins and augmented reality
Inspection data is moving from local instrument files and PDF reports to central platforms, often in the cloud. One searchable history per asset makes trend analysis possible and stops knowledge leaving with retiring inspectors.
A digital twin combines the asset's geometry, inspection history and operating data. Engineers use it to plan inspections, simulate probe coverage and decide where to look next. Augmented reality headsets or tablets can show the inspection plan and earlier findings on the real asset, which helps technicians find the right spot and position the probe correctly.
8. Data standards and interoperability
Connected NDT only works if data can move between vendors and survive for decades. ASTM E2339 defines DICONDE (Digital Imaging and Communication in Nondestructive Evaluation), based on the DICOM standard from medical imaging, for storing and exchanging NDE images with their metadata. Vendors and users are also working on open formats for ultrasonic data.
For asset owners, the practical rule is simple: put data format and export rights into your procurement requirements. Inspection records often have to be kept for the life of the asset, which outlasts most software products.
Standards, certification and the workforce
New methods are only useful in regulated work once codes accept them. Construction and in-service codes such as ASME BPVC Section V set the requirements for nondestructive examination methods, and ISO and ASTM standards are revised as digital techniques mature.
People remain the constraint. Personnel certification under ISO 9712 (common in Europe and India) or employer schemes based on ASNT SNT-TC-1A (common in North America) still applies to whoever signs the result. Training now has to cover digital instruments, data handling and the limits of AI tools alongside the physics.
How to adopt new NDT technology: a practical plan
- Start with a problem, not a technology. Pick an inspection that is slow, unsafe, costly or missing defects today.
- Check code acceptance. Confirm the governing code and customer specifications allow the new method, or what qualification it needs.
- Run a pilot against known defects. Compare detection and sizing with your current method on reference blocks and real parts.
- Plan the data. Decide formats, storage, retention, access control and how results reach maintenance or quality systems.
- Train and qualify people. Budget time for certification and procedure writing, not only equipment.
- Measure. Track inspection time, findings, false calls and rework before and after, then scale what works.
How Opsio helps
Opsio is a cloud and AI services company, not an NDT inspection body, so we work alongside your certified inspectors and NDT vendors. We build AI visual inspection and computer vision systems, cloud platforms that store and analyse inspection data, and IoT asset monitoring that links sensor readings to maintenance planning. Teams in Karlstad, Sweden and Bangalore, India support projects in the Nordics, EU and India.
Frequently asked questions
What are the newest NDT technologies?
The most active areas are the total focusing method with full matrix capture, industrial CT, permanently installed monitoring sensors, laser ultrasonics and EMATs, robotic and drone inspection, and AI-assisted analysis of inspection data.
Will AI replace NDT inspectors?
Not in the foreseeable future. Codes require qualified, certified personnel to interpret and sign results. AI reduces the time spent on routine screening and helps consistency, but a human still makes the accept or reject decision.
What is NDT 4.0?
NDT 4.0 is the industry term for applying Industry 4.0 ideas to inspection: connected instruments, digital data, automation, digital twins and analytics. In practice it means inspection data becomes part of the plant's wider data and maintenance systems.
Which NDT method is best for 3D-printed parts?
Industrial CT is the most common choice because it shows internal porosity, lack of fusion and internal geometry in 3D. Surface methods such as penetrant or eddy current, and ultrasonics on simpler shapes, are used alongside it.
How do NDT and condition monitoring differ?
NDT looks for defects in materials and welds, usually at set intervals. Condition monitoring tracks how a running machine behaves (vibration, temperature, oil) to predict failure. Permanently installed NDT sensors are blurring the line between the two. See our condition monitoring guide.
Related reading
About the Author
Head of Innovation
Jacob leads innovation at Opsio, specialising in digital transformation, AI, IoT, and cloud-driven solutions that turn complex technology into measurable business value. With nearly 15 years of experience, he works closely with customers to design scalable AI and IoT solutions, streamline delivery processes, and create technology strategies that drive sustainable growth and long-term business impact.
Editorial standards: This article was written by a certified practitioner, named in the byline above. We revise it when the technology or regulation it covers changes; the dates above show the latest update. Opsio maintains editorial independence — we recommend solutions on technical merit, not commercial relationships.