For the 2019 Tribomechadynamics Research Camp, there will be three main projects:
Project 1: Advanced Nonlinear System Identification of Jointed Structures
One of the greatest challenges with the design and optimization of a jointed structure is the accurate characterization and quantification of its nonlinear properties. Over the last two decades, there has been significant investment in the area of nonlinear system identification, ranging from methods such as the zero crossing method, the Hilbert method, the short time Fourier transform, and the wavelet-based identification methods, amongst many others. These methods all primarily exist for understanding the response of a single mode or for data from a single point on a structure. Recent advances, though, have started to shed light on multi-modal responses and full-field responses of structures.
The focus of this research will be to push the boundaries of recently developed nonlinear system identification techniques by combining new methods and experimental techniques. In particular, the primary thrust of this project will be two investigations: 1) how can using full field digital image correlation data augment nonlinear system identification techniques, with a particular focus on localization, and 2) how can newer techniques, such as nonlinear normal modes and the spectral submanifold method, complement existing nonlinear system identification techniques and inform us about modal interactions?
It is expected that at the end of this project some guidelines and a much better understanding is available for how these novel methods can help with the understanding of the underlying nonlinear dynamic behavior of a structure.
Primary Goal: Integrate digital image correlation and nonlinear system identification methods.
Work: Experimental investigation (shaker/ring down testing); Digital image correlation; Matlab-based analysis; Nonlinear system identification
Fundamental Questions Addressed: Do mode shapes of jointed structures appreciably change with response amplitude? Can a nonlinear discontinuity be identified with DIC? Is there significant modal coupling in jointed structures and, if so, how can it be quantified?
Student Qualifications: Up to six students will be recruited for this project. Ideally, these students will have expertise in digital image correlation and the nonlinear system identification techniques to be investigated. There is potential for up to two students new to this field to participate.
Mentors: Christoph Schwingshackl (Imperial College London), Matthew Brake (Rice), Keegan Moore (Nebraska), and J.P. Noel (Liege)
Project 2: Jointed Structures with Geometric Nonlinearities
Lightweight design drives structures into vibration regimes where geometric nonlinearities can no longer be neglected. At the same time, the primary cause of mechanical damping is commonly the nonlinear dry frictional and unilateral contact interactions in joints. In this project, a thin beam with frictional clamping on both ends is considered, where the nonlinear bending-stretching coupling gives rise to considerable kinematic stiffening. The system is instrumented and mounted via a frame on a shaker.
Control-based continuation (CBC) and phase control using phase locked loops are applied to identify the frequency response and the nonlinear modal properties as function of the excitation/vibration level. Particular attention is paid to the non-invasiveness of the excitation and the nonlinear mode isolation quality, and how these can be improved by controlling the higher harmonics of the excitation.
Besides controlled (mainly) periodic excitation, random multi-sine excitation is applied to identify a nonlinear state space model. Further insight into the individual strengths and limitations of nonlinear modal analysis and nonlinear state space model identification shall be gained by numerical simulations based on a mathematical model of the described setup. This permits to control the strengths of individual nonlinearities and noise.
If the progress within the project allows this, it is also investigated how well the identified nonlinear modal characteristics and frequency responses are suited for model updating.
Primary Goal: Determine and understand opportunities and limitations of CBC, nonlinear modal analysis and nonlinear state space model identification for geometrically nonlinear jointed structures.
Work: Measurements with controlled/un-controlled excitation, numerical simulations and nonlinear system identification (MATLAB)
Fundamental Questions Addressed: How sensitive are the identified frequency responses and nonlinear modes to an (imperfect) control of higher harmonics in the excitation? How do the different methods compare in terms of measurement quality, robustness and effort? Can nonlinear state space model identification be improved for the case of unilateral contacts by considering non-polynomial terms?
Student Qualifications: Up to six students will be recruited for this project. These students should have theoretical knowledge in nonlinear dynamics and system identification, experience with methods for testing and/or numerical simulation of structures with nonlinearities.
Mentors: Malte Krack (University of Stuttgart), J.P. Noel (University of Liège), Ludovic Renson (University of Bristol), Matthew Brake (Rice University), and Paolo Tiso (ETH Zurich)
Project 3: Tribomechadynamic System Identification and Modeling
A question within the joint mechanics community is how to deduce accurate and reliable properties for the interfacial contact definition. For systems that exhibit significant evolution with damage (such as those that experience fretting), there are multiple experiments that can study the progression of damage in an interface. For instance, a common approach is to use a fretting test rig to measure the hysteretic properties of an interface pair (i.e. surface to surface) undergoing macroslip. At a different scale (a smaller contact patch but larger gross motion), a tribometer measures the frictional properties of a surface being worn by a reciprocating ball contact. On a significantly different scale (potentially large contact patches but microscale, or smaller, displacements), a sonotrode can induce microslip or small gross motions across an interface, but at a significantly higher frequency than typically studied in jointed structures (kHz instead of hundreds of Hz). This project is focused on characterizing the hysteresis from these different categories of experiments and seeing how appropriate each of the measured hysteresis loops is for predicting the system level response of a jointed structure. This project will build upon a previously developed wear evolving Bouc-Wen model that is formulated in terms of physical parameters, and a numerical framework developed to directly include the hysteretic model into predicting the response of an FEA model of a large system. The project will conclude with a comparison of the numerical predictions against experimental measurements of the same large system for both pristine and worn conditions to assess the suitability of the different measurement techniques for characterizing the system hysteresis and evolution of wear within the system.
Primary Goal: To understand the suitability of different hysteretic measurements for characterizing the frictional properties of an interface in a built-up structure.
Work: Tribological characterization (tribometer, sonotrode, profilometer, and SEM potentially); analytical modeling of hysteresis; numerical/FEA modeling
Fundamental Questions Addressed: How representative are different tribological tests for characterizing frictional interaction within a jointed structure?
Student Qualifications: Up to five students will be recruited for this project. Ideally, these students will have expertise in tribology and numerical methods. There is potential for up to two students new to this field to participate.
Mentors: Matthew Brake (Rice), Christoph Schwingshackl (Imperial College London), Zack Cordero (Rice)