I am modeling a 3D masonry building for the needs of my master thesis, using shell thick elements. The problem is that when I perform a response spectrum analysis, although I am able to see stresses in local axes, the principal stresses are zero. Is it possible to overcome this problem?