Robust Bayesian Sample Size Determination for Avoiding the Range of Equivalence in Clinical Trials