Predictive Model-Assisted Guided Wave Structural Health Monitoring
Abstract
Guided wave ultrasonic structural health monitoring (SHM) has potential to monitor, assess, and track the health of many structures over long periods of time. We address challenges in applying guided wave SHM by creating a predictive modeling framework that integrates large numerical guided wave simulations with experimental data to predict wave propagation and to exploit geometric complexities for damage localization and characterization. The predictive model framework employs dictionary learning and sparse coding algorithms to deconstruct and characterize wave propagation in a complex structure. We will validate our algorithms using experimental data from four different structures, including a three-dimensional stiffened plate, which partially mimics the inside of an aircraft wing.Completion of this project will result in new technologies to improve the sustainability and survivability of United States aircrafts and munitions. Truly predictive nondestructive testing and structural health monitoring models will improve inspection accuracies, minimize inspection costs, improve safety, reduce the need for excessive data, and create pathways for new system-wide monitoring methods. Our ability to interrogate inaccessible, interconnected regions of a structures reduces costs and improves safety by reducing the need to disassemblecomponents for inspection. Our predictive models will also reduce the need for acquiring prohibitive amounts of simulations with variations in unknown parameters, such as temperature, wave velocities, and reflection coefficients.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 02, 2017
- Source ID
- FA95501710126
Entities
People
- Joel B Harley
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Utah