Machine learning is going to revolutionize the NDT world. Trueflaw's FlawML is an automatic flaw detection system based on machine learning with human level performance, which will lower the time needed for analysing data to a fraction. In machine learning, data is king. Which is exactly the reason, why Trueflaw has such unique capability in this new field: because we can manufacture thermal fatigue cracks in real samples, we can produce the vital training data for the ML system.
FlawML systems are tailored (and trained) for each application case separately. To take advantage of FlawML, you need to have digital data. This can be, e.g., digital X-ray or data from mechanized ultrasonic or eddy current inspection. Both flawed and unflawed data are required.
Trueflaw's eFlaw technology allows us to train modern, sophisticated networks with small amount of flawed data. We can also manufacture physical flaws for generating additional data.
Please contact Iikka Virkkunen (firstname.lastname@example.org) for further information.