Data analysis is a key function in structural health monitoring (SHM). To develop algorithms for SHM, one needs realistic data. A library of SHM data is introduced with simulations of vibration measurements resulting in three challenging SHM cases: (1) a beam structure with environmental and operational influences, (2) a beam with a non-linear breathing crack, and (3) moving loads on a beam, modelling traffic on a bridge. In all these cases, the excitation is varying and unknown. Also, the environmental or operational variables are random and unknown. A different modelling strategy is used in each case. Also the damage scenarios are different. Data analysis examples for damage detection and localization are presented utilizing redundancy of the sensor network. There are endless possibilities to utilize the benchmark data.