If you cannot find any data about boundary condition that seems implausible,the best way is to start with the very high numerical stiffness value, then by comparing it with the experimental research that you are trying to verify, you can easily figure it out was that value appropriate or not. If its load-deflection diagram matches well with your analytical analyse, then you do not need to change it. On the other hand, if its slope was higher than experimental work, you must decrease the numerical value of your nonlinear springs. Although It is an iterative process, I do believe it saves more time.