Sensitivity-enhanced NDE computations for SHM applications
Abstract
Research Problem and Objectives:The implementation of the Digital Twin methodology promises to provide improved prediction accuracy and minimizelife-cycle costs using the as built geometry and properties, and updating the model with as-experienced informatio n from thephysical twin. A critical component is the incorporation of an on-board vehicle health management system to mirror the li fe ofthe physical twin. To that end, we propose to develop a unique modeling capability to assess, to optimize, and to deploy effec tiveon-board damage detection sensor systems.Technical ApproachDevelopment:The proposed research is to develop a hypercomplex sp ectral finite element formulation and computer software to predict highlyaccurate displacement time histories and derivatives of th e time histories with respect to sensor location, sensor materialproperties, structural material properties, geometries, and bounda ry conditions for wave propagation problems. Thehypercomplex differentiation method has been shown to be highly accurate within mul tiple domains; however, the method hasnot been implemented in the field of wave propagation. The fundamental considerations will be mathematical, numerical, andcomputational to have a usable and effective method. The technical challenge will be to systemically h ypercomplexify thesoftware, manage the additional data, provide the necessary supporting hypercomplex library routines, and minimiz e theunderlying computational expense required to solve the equations, verify, and validate the software. Once developed, thesoftw are will be general and the derivative to compute will be defined by the user through the input data.Typical parameters that will b e investigated include: a) material properties, b) sensor properties, c) sensor location, and d)geometric properties. Three categor ies of sensitivities will be computed: i) fully characterize a sensor system (run one time andreused in other scenarios), ii) quali ty control, e.g., which piezo or bond properties are important, iii) POD sensitivities, e.g.,distance to damage, and type of damage .VerificationVerification of the accuracy of the computed wave propagation results and sensitivities will be proven through a comb ination ofcomparisons with known or manufactured solutions and with numerical differentiation estimates. Existing relevant experime ntaldata will be used for comparison, e.g., estimate numerical sensitivities using multiple experimental results.Experimental Vali dation (Option)The proposed effort will be to perform design of experiments at nominal and off-nominal conditions to validate the s ensitivitiespredicted by the software. Variations in material properties (aluminum, steel, titanium, and composites), sensor proper ties (scalesensor properties by temperature), and geometric properties will be explored consisting of, nominally, a set of 2 tests for eachcondition using a 4-pt bend specimen.Anticipated OutcomesThe anticipated outcomes are the ability of the Navy to develop optimal NDE/SHM systems and to identify the key variablesgoverning the NDE systems. More importantly, this research will usher in a fundamental change in the science of NDE modelingthat will propagate to other researches and developers and will have profound i mpacts on all future sensor systems development.Impact on DoD capabilitiesThis capability will support the future naval capability mission to deliver robust management decision tools to shift away fromthe current deterministic, fixed-flight-hour-based maintenan ce practice to one based on the management of risk and reliabilityover the rotary and fixed wing fleet.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 20, 2021
- Source ID
- N000142112428
Entities
People
- Harry Millwater
Organizations
- Office of Naval Research
- United States Navy
- University of Texas at San Antonio