per atom stress calculation LIGGGHTS

Hello users,

I have been using LIGGGHTS for studying flow down a rough inclined plane. I wanted to validate the stress profiles (for example, normal stress \tau_zz, perpendicular to the base) with the momentum balance prediction (\tau_zz (z) = g*cos_beta*integral of (rho dz), where beta=angle, rho=bulk density) and found that the computed stress was way larger than the one predicted by the momentum balance equation. I initially calculated the per atom stress using "compute stress/atom" and then did the ensemble averaging over some time using " fix ave/spatial" . I guess the per atom stress calculation based on pairwise contact is not right. The contact stress should act only on the contact plane and ensemble averaging should be done based on the contact points instead of atoms' centroids [Reference: "The stress tensor for simple shear flows of a
granular material", J. Fluid Mech. 203 449-473 (1989)]. I look forward to your comments.

Thanking you,
Sandip Mandal