Prashant K. Jha

I am an Assistant Professor in the Leslie A. Rose Department of Mechanical Engineering at South Dakota Mines, where I lead the CEAD Lab.

The question I keep coming back to

What must a model keep in order to be trusted, and where is it worth spending computation? Almost everything I have worked on is one half of that or the other. The mechanics half asks which physics controls a failure. The computational half asks what a saving costs, because a cheaper calculation is only a saving if the quantity you were asking about survives it.

How I got here

A master’s at the Indian Institute of Science with Chandrashekhar S. Jog, on monolithic compressible fluid–structure interaction, where I learned much of the nonlinear continuum mechanics and finite element method everything since has rested on. Then a PhD at Carnegie Mellon with Kaushik Dayal, coarse-graining electric-field interactions with materials: the habit set there was mechanics first, mathematics second, implementation third, and it has not changed since. Then a postdoc at LSU with Robert Lipton, turning nonlocal fracture theory into solvers that run and theorems about what they converge to. Then the Oden Institute with the late J. Tinsley Oden, where the question shifted from is this model correct to is this model good enough for the decision being made with it.

That last shift is the one that organizes the group now. It is why PeriDEM resolves deformation and fracture inside each grain instead of treating a particle as rigid, why the corrector operator exists at all, and why the papers report the cases where the methods fail.

How I work

I would rather publish a method with its failure modes stated than a method that looks universal. Code meant for anyone else to use is released with documentation, and the error control is part of the result rather than an appendix to it. The graduate FEM course is being written as an open textbook in public, as the course runs, for the same reason: the working is the teaching.

Who I want to hear from

Students who want to build solvers rather than only run them, and who are willing to check their own results. I am interested in collaborations where an experiment can test a model’s assumptions, distinguish competing mechanisms, or expose a consequential model error. There is more on what the lab is looking for.

The full record of appointments, publications, funding, teaching and service is on the CV.