Effects of surface roughness on magnetic flux leakage testing of micro-cracks

Magnetic flux leakage (MFL) testing owns the advantages of high inspection sensitivity and stability, but its testing results are always affected by surface roughness. The relationship between the surface roughness ($R_a$) and detection signals for surface-breaking cracks is mainly discussed. The existence of roughness magnetic compression effect (RMCE) in present MFL testing is specially pointed out and its relevant theory is also analyzed, which manifest themselves in the compression of MFL signal in its peak value and the baseline drifts mixed with noise. An experimental investigation on surface comparators with different arithmetic average height ($R_a$) and artificial notch size, is performed to analyze the effects of surface roughness on detection signals of cracks. The detection limit (DL) of micro-crack is analyzed by comparing the $B_y$ noise-signal ratio ($S_y$) and peak-peak signals of the cracks. Meanwhile, $S_y$ increases with the $R_a$ and $R{S}_m$, in this case, relatively shallow defects cannot be clearly distinguished at determined rough surface. Afterwards, a series of simulations are designed and performed to verify the effects of surface roughness on characteristic $B_y$ of the electromagnetic field, and a theoretical DL of micro-crack is presented as: $DL=2.88R_a+7.00$. Furthermore, the optimal lift-off value is selected for the micro-cracks' detection to weaken the negative magnetic compression effect. MFL signals cannot reflect the accurate sizes of the cracks on rough surface due to the RMCE and its relevant phenomenon. The discovery and results will benefit the quantitative evaluation of the MFL testing. (paper)