Repository navigation
Edge-edge discontinuity of ESP energy #280
Description
Hi @fsichetti, I see in #259 that you have released your implementation of ESP. Congrats for the great work!
To understand this work better, I have set up a scene of a triangulated box, which applies EPS forces to a nearby edge, or another rigid box (which I treat as a rigid body and sum up the forces and torques). I have tried to produce a visualization of the force field by ESP using Polyscope. One thing that I notice is that the when an endpoint of an edge is move past the other edge, there is a sudden jump of force applied to the vertices. The attached videos show this situtation, and the script should be able to reproduce it (the only dependency should be polyscope).
May I know wheter you think this is exactly how ESP should work or I am doing something incorrect? How do you understand this phenomenon? Do you see this as an issue and do you see any fix for it? Thanks a lot in advance!
https://github.com/user-attachments/assets/d490c1d7-b0eb-41fa-9b64-bb29f60a4b96
https://github.com/user-attachments/assets/c79deae5-6cdb-4941-a30e-1fe9b6f1caf1
Activity
fsichetti commentedon Oct 29, 2026
Hi @xiangdonglai, this is a formulation issue with the EE quadrature that dominates when meshes are very coarse. The potential is not discontinuous but changes very abruptly. We have a fix that seems to drastically improve things, I will share it in the next few days as a PR. Thanks again for reporting the issue!
xiangdonglai commentedon Oct 29, 2026
Thank you! @fsichetti Looking forward to the improved formulation.