/sci/ - Science & Math

I got a PhD in physics theory and of the 15 or so people who graduated in my semester, most are at banks, consulting or energy companies. Two are at CERN, one is at a similar institution. I'm probably the only one not doing HEP but something physics related, and e.g. a year ago I was writing Kalman filters.

>>11976448
Could you talk more about what you do? Does the knowledge of control theory supplemented in any way your physics related work?

I am a physics undergrad, and I got interested in the topic of control. I will take some classes on the subject this semester.

https://youtu.be/oBc_BHxw78s

it's a meme, it's on yotube and wikipedia. Both unreliable sources.

>>11976448
That's funny I was doing that in my engineering bachelors, I thought physics was superior to engineering?

>>11976516
To apply control you need to be an engineer, chemical, mechanical, electrical or aeronautical, anything else is a poor imitation of the subject, you'll probably learn some undergraduate level shit and not understand the subject at all.

>>11976620
The classes I plan to take are offered by the engineering department. There's nothing similar to control in the physics curriculum.

Papers like these:

https://www.sciencedirect.com/science/article/pii/S1474667017353880/pdf?md5=92eacfff5634355da4b468318cdda5b5&pid=1-s2.0-S1474667017353880-main.pdf
https://www.google.com/url?sa=t&source=web&rct=j&url=http://citeseerx.ist.psu.edu/viewdoc/download%3Fdoi%3D10.1.1.124.7043%26rep%3Drep1%26type%3Dpdf&ved=2ahUKEwisgoPJiYTrAhVXHrkGHbBABbcQFjAWegQIBhAB&usg=AOvVaw1KN25HuE2cUcv9AVF8qfHR&cshid=1596631225196

Made me question whether control is something that is applicable in the realm of physics. On the other hand, I can only find stuff written by those two authors >>11976421

>>11976421
So some physical questions are questions of stability, I believe turbulence is one. So if you like fluids there's probably some physics there to discover because we know so little about turbulence. Aeroelastic flutter is another example of a stability problem the structures engineers do when designing a plane. For complicated structures there's definitely more physics to discover in that field. I once knew a guy who did his phd thesis on quantum control, I guess it's a thing.

>>11976671
It seems that control theory is more inclined to be useful for experimental physics.

Control theory basically looks to shape a differential equation to a certain response via its input, I'm a control engineer-theorist student but I work in robotics, so I can merely give you some pointers for you to connect the dots.

>>11976421
If you apply control theory to anything, you're doing engineering.

>>11976671
Do you realize that designing something to be stable is not control theory, right?

>>11979195
If you could point a thing - or a short list of things - that encapsulates the power of control theory, I might be able to orient myself on the field.

Right now, I have a general picture of what control theory is, but I can't see the extension of it.

>>11979374
*Regulate and track a set point: So you have a system, usually a dynamical system hence the differential equation, and you want it to behave in a certain way.

Example would be a robot manipulator, that thing has certain dynamic behaviors that are described by the differential equations obtained from an Euler Lagrange model, you want the end-effector to track a trajectory or regulate itself at some fixed point, you inject an input (with feedback, aka with measurements of the output positions and velocities for example) that will shape the dynamics of the robot so it tracks the desired trajectory or fixed point.

Linear design is pretty straight forward and uses Laplace domain and Linear algebra.
Non linear stuff uses Lyapunov and Lyapunov-like theorems.

Another thing that is very important in control is stabilizing equilibrum points of system. For example, the cart-pendulum's pendulum up equilibrium point is an unstable one. You can input a control law so the equilibrium becomes an stable one.


The most important concept in control is stability, so I suppose that there are some models in physics that are described by unstable differential equations that could somehow be stabilized?

>>11979408
>The most important concept in control is stability, so I suppose that there are some models in physics that are described by unstable differential equations that could somehow be stabilized?

That sounds like "control of chaos", which was the paper that started the cybernetical physics field.

The thing is that the field is defined by a set of papers that used some aspects of control theory, and since then it been stale.

Thanks for your answer, I will look it up.