A broad spectrum of applications require numerous simulations (such as optimization, calibration or reliability assessment for example) in various input sets. In that context, some simulation failures or instabilities can be observed, due for instance, to convergence issues of the numerical scheme of complex partial derivative equations. Most of the time, the set of inputs corresponding to failures is not known a priori and thus may be associated to a hidden constraint. Since the observation of a simulation failure regarding this unknown constraint may be as costly as a feasible expensive simulation, we seek to learn the feasible set of inputs and thus target areas without simulation failure before further analysis. In this classification context, we propose to learn the feasible domain with an adaptive Gaussian Process Classifier. The proposed methodology is a batch mode active learning classification strategy based on a Stepwise Uncertainty Reduction of random sets derived from the Gaussian Process Classifiers. The performance of this strategy will be presented on different hidden-constrained problems and in particular within a wind turbine simulator-based reliability analysis.