diffmpm-bridge

Structural identification through differentiable MPM

Simulation in Taichi MPM

The simulation is built with taichi MPMsolver from taichi_elements. The project starts with a simple scenario by dropping a wheel with an initial angular velocity $\omega_0$.

To run the mpm simulation you can simply run python run_mpm.py with an argument -o "path/to/output/folder". It will output .png files for each dt. It will run 500 frames for simulation, and store a .png file for each frame in output directory.

To make a .gif of these files, you can run python make_gif.py -i "path/to/png/folder" -o "path/to/output.gif". You can use --fps to control the frame rate.

Experiments

• 2d_plate: 2D 1:1 plate subject to point force

• 2d_beam_rolling: 2D 1:2 beam subject to rolling load, E optimized

• 2d_rolling_enlarge: 2d_beam_rolling, larger scale to fit frame, E optimized

• 2d_beam_f: 2D deep beam, testing density compensation

• 2d_beam_p-e_ratio: 2D deep beam, testing P/E ratios

• 2d_cantilever: 2D deep cantilever beam subject to point load, nonhomogeneous elastic field

• 2d_multistiffness: 2d_rolling_enlarge, nonhomogeneous elastic field, E optimized

• 2d_multistiffness_force: 2d_multistiffness, E-F both optimized

• 2d_beam: 2D 1:8 beam, faster rolling load

• 2d_beam_noise: 2d_beam with noise injected into signal

• 2d_beam_noise_avg : Uniform noise level applied to all particles, determined by avg signal power

• 2d_beam_discretization : 2d_beam_noise_avg with modular discretization

• 2d_beam_gradient : 2d_beam_discretization but true experiment is a gradient of E's

• 2d_beam_damage : 2d_beam experiment with damage identification, fill layer + masonry layer

• 2d_beam_damage_uniform : 2d_beam experiment with damage id, uniform undamaged layer