Reinforced concrete (RC) and steel moment resistance frame (SMRF) structures are common in seismic zones. However, damage assessment after earthquakes can be problematic and subjective. This research applies fully automated hysteresis loop analysis (HLA) structural health monitoring (SHM) to several experimental cases. It quantifies accuracy and robustness for realistic structures over multiple events to demonstrate the ability to accurately monitor structures long-term.Data is analysed from: (1) experimental, scaled 12-story RC structure subjected to 2 events; (2) 2x12-story scaled RC structures undergoing 4 events; and (3) a full-scale 3-story E-Defence test with 6 ground motions. Accelerations of each DOF are recorded with lower rate displacements. Nonlinear hysteresis loops are reconstructed for each DOF for analysis. Changes in identified elastic story stiffness in these loops are used identify damage location and severity, and tracked for multiple events. The error from final identified stiffness to initial identified stiffness in a subsequent event assesses the method’s ability to accurately and continuously monitor a structure. Elastic stiffness drops of 24%, 23% and 21% were identified for 2-4 th DOF of the 12-story structure for the small ground motion though no visible damage was recorded, with large drops over 50% from initial values after the strong ground motion. Similar results are obtained for the 2x12-story RC frame test case, and the E-Defence test. More importantly, final and subsequent initial stiffness values were within 10% for all cases, and all but 2 were within 5%, clearly showing the consistency and reliability of this method. Initial stiffness for the first event was within 6% of calculated values. Overall, results indicate the HLA method automatically, accurately, and robustly detects and assesses damage location and severity for realistic structures across multiple events without human input, which has not been previously demonstrated by other methods.