Sreevatsav prasath

Robotics • Aerospace • Films • But mostly just a Stupid kid trying

[Sreevatsav prasath @Vats_42]

To build stuff like good one day!🤞

[Marwa ElDiwiny @MarwaEldiwiny]

Passive dynamic stability.

Intelligence in mechanism: no sensors, no feedback control.

[Sreevatsav prasath @Vats_42]

Need to start dreaming equations ASAP, else ngmi

[Sreevatsav prasath @Vats_42]

Turns out the partial wall slip amplifies the slip per unit kn but also at the same time flattens the velocity profile’s slope as it grows

The underlying degeneracy of the problem still persists

Gonna check if using curvature based higher order slip model can separate kn and σv

[Sreevatsav prasath @Vats_42]

Welcome to the sequel of the 3 am rambling, not at 3 am (hopefully)

Picking up from the earlier endpoint, I kept the σv fixed and swept over 100 evenly spaced kn values from 0.001 to 0.2 to build a better dataset.

Hoping that we could recover the kn from a velocity profile.

[Sreevatsav prasath @Vats_42]

Onto the next attempt:

Instead of inverting one profile, I’m gonna try sweeping kn at a fixed σv, and see if that brings enough information for the MLP to predict the params.

Thank you for attending my 3 am rambling!

[Sreevatsav prasath @Vats_42]

My expectation was that when the σv=0.5, the kn would be predicted 3x more precisely than the fully accommodating case (σv=1.0).

Based on the conversion factor math but at noise=0.5% it was actually only 1.6x, this piqued my interest.

[Sreevatsav prasath @Vats_42]

Onto the next attempt:

Instead of inverting one profile, I’m gonna try sweeping kn at a fixed σv, and see if that brings enough information for the MLP to predict the params.

Thank you for attending my 3 am rambling!

[Sreevatsav prasath @Vats_42]

I was studying and trying to understand couette flow with slip condition,

then as a fun experiment I wanted to see if the knudsen number (kn) and the tangential momentum accommodation coefficient (σv)could be predicted given a velocity profile

A nice inverse problem for ML

1/n

[Sreevatsav prasath @Vats_42]

Both σv and kn don’t determine the velocity profile independently but are related by the maxwell slip length relation.

There just isn’t enough information to distinguish σv and kn.

[Sreevatsav prasath @Vats_42]

Sheet metal - the answer to every design question, regardless of what question it is. 👍

[em m0shouris @emm0sh]

sheeted in 1/2" thick aluminum and weighing in at 930 pounds, a contender for the world's most uncomfortable chair

[Sreevatsav prasath @Vats_42]

How can one claim engineering is not art??