Details on Hysteretic Property Measurements
[Ivanov99]Hysteretic parameters were measured using a Magniscope, an instrument designed and used in the Metals Development Lab at Iowa State University. This instrument allows "local" measurement of the B-H characteristics for ferromagnetic materials. The B-H curve is measured by magnetizing a small volume of the sample, using a yoke. The field intensity is measured, with a Hall probe located in the middle of the yoke. The probe is oriented to pick up the horizontal component of the field. The flux density is measured using a coil wound on the yoke. Parameters such as coercivity, remanence, and hysteresis loss are estimated from the B-H curve. The depth of measurement is roughly equivalent to half the distance between the poles of the yoke. A half-inch probe was used; the penetration depth was, therefore, approximately 0.25 inch.
Measurements were made on the surface opposite the defect. The scanning area was 3 by 3 inches, divided into a 12 by 12 grid, with the defect located in the middle of the scanned area. The measurement procedure included demagnetization, registration of a single hysteresis loop, and demagnetization again. Care was taken so that the orientation of the magnetic field remained constant. Measurements were taken with the field oriented in two perpendicular directions. Measurements were also made on a circular grid with eight divisions along the circumference and six divisions along the radius, resulting in 48 measurement locations. The magnetization field for circular measurement was radially oriented. This was done in order to maintain the symmetry of the residual stress field.
The resulting sets of data were processed and are shown below. The data were compared with the stress distribution patterns obtained from the structural finite-element model discussed earlier. For example, the scan shown in the first figure represents the distribution of coercivity (Hc) around a defect corresponding to a 30 kip load. The second figure shows the calculated residual stresses for the same defect.
The small variation of coercivity around the metal-loss defect in the first figure is a result of measurement and instrumentation error and does not indicate a variation of the coercivity of the material. No variation should be expected because the area is free of stress. The pattern around the pressed-in gouge exhibits a very large variation, on the order of 25 percent. This variation represents the residual stress in the sample due to the mechanical damage. The third and fourth figures show the distribution of remanence and hysteresis loss around the same defect. Similar results are observed.
The results suggest that the residual stress can be linked to magnetic parameters, such as coercivity, remanence, and hysteresis loss. All of the circular scans showed patterns of the shape similar to the expected stress distribution in the test samples. The sensitivity of a parameter to residual stress can be estimated from the relative change in the observed pattern. Remanence is more sensitive than coercivity, but hysteresis loss was most sensitive.
