This is not a bug. The shell element is in a state of plane stress, so when ANSYS reports principal stresses, there will be one which is expected to be, and is, zero. The three principal stresses, output as S1, S2, and S3, are ordered so that S1 is the most positive (tensile) and S3 is the most negative (compressive). At the top and bottom of the shell surfaces, where transverse shear stresses are zero, there will always be a zero principal stress (in the direction normal to the shell face). For anywhere else through the shell thickness where there are nonzero transverse shear stresses, this may not be the case. This may explain some of the differences between ANSYS and NASTRAN. We are not aware of how NASTRAN may treat principal stresses in shells; however, we have not observed anything abnormal in the results reported by ANSYS SHELL181.