Background
In the absence of an in-person research camp, like we have hosted in the past, we would like to propose a research challenge to the community this year to continue our collaboration and encourage a survey of the current state of the art in nonlinear dynamic prediction capability.
The challenge is to do a blind prediction of the nonlinear dynamic response of a system that has not yet been fabricated. You will be given the CAD model and technical drawings, including material and surface specifications required to manufacture and assemble the system. This challenge corresponds to an engineering task typical of the daily work within industry, and is distinctly different from recent research thrusts, which have focused on calibrating models against measured properties of a fabricated prototype. To make this challenge test the limit of our predictive capabilities, we are intentionally designing it to have both a frictional and geometric nonlinearity. The design of the system is based upon industrial best practices (specifically in terms of bolt patterns and torques). With the increasing direction of industry to reduce (and eventually eliminate) physical prototyping, addressing this challenge becomes even more important.
The Benchmark System and Your Tasks
A schematic illustration of the benchmark system is given in Figure 1. The system consists of a thin plate mounted on two rectangular profiles. The system is assembled at room temperature and the analysis shall be carried out assuming room temperature as well. You are asked to predict the frequency and damping ratio of the lowest-frequency elastic mode as function of the amplitude. The amplitude is defined as the root-mean-square value of the transversal displacement at the plate’s center. Frequency and damping are to be determined at least up to two times the plate’s thickness.
We would like to keep this challenge as open as possible and not restrict participants with regards to their modeling choices. Use your best engineering intuition, for instance to choose contact models, to consider or neglect nonlinear effects due to bending-stretching coupling, to decide if it is appropriate to simplify the problem to a two-dimensional one, and to incorporate reasonable levels of uncertainty.
Similarly, by requesting amplitude-dependent natural frequency and damping (as opposed to, e.g., frequency responses), the participants may choose their definition of these properties and analysis techniques as diverse as quasi-static, Harmonic Balance or time integration procedures.
Figure 1: The benchmark system – preliminary design (see the full documentation for more details).
Scientific Goals and Long-Term Plans
The scientific goal of this challenge is to assess the community’s state of the art strategies for making blind predictions of the nonlinear dynamics of jointed structures. Commonalities and discrepancies among methods, unique innovations, and open challenges will be identified and documented in a review paper that participants will be invited to be co-authors on.
Following completion of this initial challenge, it is planned to manufacture several of the benchmark systems and test them as part of the 2022 research camp, to determine the true dynamics and its variance. As these structures have not yet been fabricated, it is reasonable to include uncertainty in your modeling and predictions as well, based on the provided manufacturing tolerances.
A perspective is a final challenge in which the experimental results are provided, and the community is asked to predict the dynamics of a modified structure (much like in industry where experiments are used to calibrate models of an initial design, which is then used to make predictions for the next version of the design).
Timeline
The anticipated timeline of this challenge is:
March 2021: Challenge announced to the community, along with a list of initial teams committed to participating
August 2nd, 2021: (Optional) Presentation of modeling strategy and initial results at Tribomechadynamics 2021
September 30th, 2021: Simulation results due
December, 2021: Meta-analysis and documentation
February, 2022: Presentation of results at IMAC
Summer, 2022: Measurements of the benchmark system
Available Data
The CAD model, technical drawings and further specifications are available for download.
Registration and More Information
For all teams interested in participating, please register with the organizing committee by emailing Malte Krack to indicate your interest. Once you register, supplemental information will be made available as appropriate, and information for submitting your solution will be provided.
If you would like more information, please contact the organizers:
| Matthew Brake
Rice University |
Malte Krack
University of Stuttgart |
Christoph Schwingshackl
Imperial College London |
Submission Information
To participate, an individual/team should submit:
- Predictions (including uncertainty bounds if applicable) of the linear natural frequency of the structure for the first five lowest-frequency modes
- Predictions (including uncertainty bounds if applicable) of the amplitude-dependent frequency, normalized with respect to the linear frequency, and amplitude dependent damping, as specified above for the lowest-frequency bending mode of the plate
- A clear description of the methodology (modeling, analysis procedure, assumptions/simplifications, uncertainty bounds on input parameters, etc.)
- Estimation of computational time
We request that the computational time for your simulations be documented (specifically in terms of pre-processing time, e.g., initial runs of a model in an FE solver, processing time, e.g., your main simulation loop in your nonlinear solver, and post-processing time if applicable). To appropriately scale reported computational times, a small sample problem will be provided to estimate the relative speed of the hardware employed for the challenge.
Instructions for submitting will be provided by the organizers to all teams enrolled in the competition.