Efficient Modeling of Nonlinear Scattering of Ultrasonic Guided Waves From Fatigue Cracks Using Local Interaction Simulation Approach

This paper presents an efficient modeling technique to study the nonlinear scattering of ultrasonic guided waves from fatigue damage. A Local Interaction Simulation Approach (LISA) is adopted, which possesses the versatility to capture arbitrary fatigue crack shapes. The stick-slip contact dynamics is implemented in the LISA model via the penalty method, which captures the nonlinear interactions between guided waves and fatigue cracks. The LISA framework achieves remarkable computation efficiency with its parallel implementation using Compute Unified Device Architecture (CUDA) executed on GPUs. A small-size LISA model is tailored for the purpose of extracting the guided wave scattering features. The model consists of an interior damage region and an exterior absorbing boundary. The interior damage region captures various types of fatigue crack scenarios, while the exterior absorbing boundary surrounds the damage model to eliminate boundary reflections. Thus, the simulation of guided wave scattering in an infinite media can be achieved utilizing a small-size local LISA model. Due to the parallel CUDA implementation and the small-size nature, this local LISA model is highly efficient. Selective mode generation is achieved by coupling/decoupling excitation profiles with certain wave mode shapes, which allows the study of sensitivity of different wave modes to a certain fatigue damage situation. At the sensing locations, mode decomposition is performed on the scattering waves, which enables the study of mode conversion at the damage. Fourier analysis allows the extraction of scattering features at both fundamental and higher harmonic frequencies. A numerical case study on nonlinear scattering of guided waves from a fatigue crack is given. The higher harmonic generation and mode conversion phenomena are presented using the wave damage interaction coefficients (WDIC), from which the sensitive detection directions can be inferred to place sensors. This study can provide guidelines for the effective design of sensitive SHM systems using nonlinear ultrasonic guided waves for fatigue crack detection.