In the trend towards resource conservation, lightweight construction, and electromobility, there are increasing demands on the material properties and quality of critical components, in addition to the requirement for error-free production. Besides fatigue strength, a stress-oriented heat treatment of the surface layers of functional surfaces must ensure that components withstand high static and dynamic loads. The process-integrated, necessarily non-destructive AI based determination of surface layer properties, such as surface hardness and case hardening depth, is becoming increasingly important in quality assurance outside the laboratory.

An initial application is the testing of constant velocity joints, whose shafts in electric cars are now subjected to much greater loads than those known from internal combustion engines. Traditional sampling methods (destructive laboratory preparation for hardness depth determination) are reaching their limits, which an efficient 100% inline non-destructive method can overcome.

For ferromagnetic components, the resulting eddy current signals reflect the physical material properties of the steel, such as electrical conductivity and magnetic properties. These are directly related to the mechanical-technological material properties. Through a multi-frequency testing (iPMFT) with simultaneous harmonic analysis (iSHA), the measurement effects can be separated, and by automatically varying the test frequency, additional information from different surface layer depths can be obtained. A subsequent deep learning AI evaluation provides reliable predictions of surface hardness and the achieved case hardening depth.