Technical Papers and Presentations

Sulfidation corrosion of Nickel based super-alloy turbine blades is one form of corrosion that adversely affects blade performance. It results in a depletion zone involving magnetic corrosive byproducts with a high relative permeability [1]. Magnetometers are sometimes used in non-destructive testing, to detect the presence of these zones, and determine when a blade should be removed from service and repaired. However, without modeling, the measurements do not provide enough details on the size or depth of the depletion zones.

We developed a finite element model, shown in Figure 1, of the Foerster Magnetoscop probe 1.070 [2]. The probe measures the relative permeability in accordance with IEC 60404-15 and ASTM A342M. We modeled this as a magnetostatic problem since there are no electric field or high frequency effects on the probe or its surroundings. The model included the central magnet, the blade including regions of high permeability, and the surrounding air. The precise method used by the probe to calculate the relative permeability is not available. Therefore, we assumed that the probe measures the average of the horizontal magnetic flux density at two locations [3]. Our first steps were to verify that this measurement procedure is accurate, and to develop a relationship between the probe reading and the measured horizontal magnetic flux density. We accomplished this by comparing the simulation model output to experimental permeability measurement data involving different distances between the probe and calibration blocks of known permeability. Figure 2 shows the agreement between simulation and experimental measurements for different distances between the probe and a calibration block with a relative permeability of 1.0195. With this relationship verified, we simulated cases of actual blades with increasing diameters and depths of depletion zones to determine combinations of minimum/maximum depletion zone dimensions that can be detected by the probe.

We also explored the use of the boundary element method for this magnetostatic problem. As the boundary element method requires only a surface mesh, it was much more computationally efficient and we found it to be as accurate as the results from our finite element model.

References

1. Bornstein, N.S., Reviewing Sulfidation Corrosion—Yesterday and Today, JOM (1996) 48: 37-39

2. Foerster. (n.d.). Magnetoscop 1.069: Product Information. Retrieved from https://www.fluxgatemagnetometer.com/assets/foerster/media/Downloads/Magnetoscop%201.069/1069GBEN_RE V0114.pdf.

3. Foerster. (n.d.). Magnetoscop 1.070. Retrieved fromhttp://www.foerstergroup.com/en/usa/products/magnetoscop-1070/