Brandt's method is explained in his book (best to buy it in print but I think there is a hooky pdf download somewhere if you look) and looks a bit like this
The method I usually use applies the load at right angles to this (i.e. parallel to the axle), at the spoke crossings.
In theory it is similar to the stress-relief loading applied by some wheel building machines eg
where the side wheels are pushed against the spokes, but in reality some wheels that are built in such machines still break spokes, so either there is an intrinsic difference in stress relief so produced, or the conditions used are substantially different in some way.
cheers