Recent Advances in Computational Methodologies for Real-Time Hybrid Simulation of Engineering Structures

Predicting complex structures’ dynamic behavior is challenging, especially since full-scale testing of structures subjected to natural hazards is not always possible. Testing individual systems is insufficient, and consequently, engineering experimentation requires refinement. Hybrid simulation (HS) is a cost-effective and efficient dynamic testing technique that evaluates a systems’ performance with rate-dependent behavior. HS is also known as cyber-physical testing, dynamic virtualization, pseudo-dynamic testing, dynamic sub-structuring, and hardware-in-the-loop. In structural engineering, it consists of combining experimental-analytic simulations of structures subjected to dynamic loading. It seeks to: (1) leverage established understandings about the physical world to gain insight into the behavior of physical systems that have limited prior knowledge, and (2) study the coupling of physical and computational models to realistically include their dynamic interactions. The cyber-physical setup consists of dividing a structure into numerical and physical substructures and using actuators to achieve the coupling. The actuator dynamics generate a time delay in the overall system that affects the simulation’s accuracy and stability. Hence, tracking control methodologies strive to mitigate these adverse effects. Cyber-physical testing with linear, pre-determined models has been studied and established. Thus, recent studies seek to enable the most realistic conditions for such engineering experimentation through robust, nonlinear and adaptive control methodologies to address challenging cases involving damage, failures, or sharply changing dynamics. This paper presents a state-of-the-art review of recent tracking control methodologies for real-time hybrid simulation (RTHS), including identifying the limitations and challenges of modern implementations. Furthermore, this paper presents a comparative study evaluating control methodologies using a benchmark problem.