I'm interested in your thoughts on how to locate the position and depth of a tension crack in a method of slices/limiting equilibrium computer analysis. The various programs offer lots of latitude in deciding how/where a tension crack is modeled. I've come across a few schools of thought on how to do this and hope to learn a few more:
1. Change the tension crack depth and location until the minimum Fs is found
2. Place a tension crack to eliminate negative (tensile) interslice normal forces. Get forces from free body diagrams or force output tables.
3. Set the tension crack depth equal to some percentage of the failure height. ie Cousin's charts limit crack depth to half the slope height.

I've used #2 over the years, but was told recently that this is not the right way to model a tension crack as the interslice forces in the LE analysis are not realistic since the analysis method forces each slice to have the same Fs. I get this, but also disagree with the logic to a certain extent.

Thoughts??