0
$\begingroup$

I am into a project that simulates a moving truck with a tank holding water inside it, this is example (https://www.youtube.com/watch?v=jqpl4ME6rRY).

I studied the mechanics of the motion of the truck, and the motion of the water inside a static tank.
the problem is the mutual influence between the motion of the fluid and the motion of the truck. so there are two main problems I am stuck with:

  1. the influence of the motion of the truck to the motion of the fluid inside it.
  2. the influence of the motion of the fluid to the motion of the truck.

You can suppose that the truck moves in a straight line parallel to X-axes, with left and right move only.

for problem 1, as I am using SPH method to simulate the liquid, I am thinking of add the acceleration of the truck to the SPH equation for every fluid particle, I am not sure about it..

for problem 2, maybe I should compute the pressure force effecting to free surfaces of the truck (getting them from the pressure force of the particles) then add it to the acceleration of the truck, I am not sure also..

I hope my ideas to be correct and to share yours with me. Thank you.

$\endgroup$
  • $\begingroup$ How complex/accurate do you want your model to be? Will you be modelling only static cases, i.e. constant acceleration, or do you want to study the dynamics the the truck changes its acceleration? $\endgroup$ – Klas Lindbäck Dec 14 '16 at 13:27
  • $\begingroup$ The second one, the dynamic cases $\endgroup$ – Tareq Habbab Dec 14 '16 at 14:19
  • $\begingroup$ What is SPH? This is not a standard term in physics. I presume you mean Smoothed Particle Hydrodynamics? $\endgroup$ – sammy gerbil Dec 14 '16 at 20:25
  • $\begingroup$ Yes I mean Smoothed-particle_hydrodynamics. $\endgroup$ – Tareq Habbab Dec 14 '16 at 20:51
1
$\begingroup$

Since you cannot model the response of the driver, it seems to me you will just have to assume a motion for the truck. (By "the truck", I mean the truck and container but not the fluid itself, the fluid it carries is being treated as a separate subsystem that is subject to the constraints of the motion of its container.) That means part 2 will be limited to finding the forces of the fluid on the truck for given truck motion, rather than solving for the full motion of the center of mass of the system. The SPH can then handle part 1 easily.

So I'd recommend that you select the acceleration of the truck part of the system in advance, call it A(t) to make it a given function of time. The effect of that on the fluid is simple, because you go into the truck frame (to fix the container walls), and just apply a fictitious effective gravity of -mA(t) to any particle of mass m in your SPH. Then run the SPH, and like you say, the pressure at the walls is what will produce forces on the truck. Then you can say for any given A(t), what forces will the truck experience from the fluid, and you can assess how that might feel to the driver, but you can't determine how it will change the movement of the truck (i.e., change A(t)), because you can't know how the driver will respond.

You might also want to consider the drastic differences that will appear if there is any significant amount of air in the container, such that the fluid can slosh back and forth. That would require a two-fluid treatment in the SPH, so it might be too difficult to include.

$\endgroup$
  • $\begingroup$ @TareqHabbab On top of what has been said above, you could assume the "driver" is a very simple pre-programmed robot that only knows how to hit the gas and brakes and doesn't really react to the environment. Then the pressure forces can be used to update the dynamics of the truck as well (e.g. a simple model could assume the integrated pressures on the fluid container act as torques and forces on the center of mass of the truck). The total system then will include these reactions from the moving fluid as well as the pre-programmed acceleration. $\endgroup$ – neocpp Jan 2 '17 at 3:13
  • $\begingroup$ @TareqHabbab This I believe is your original suggestion in the question. However, if you wanted to add reaction due to the driver, a simple idea would be to use a feedback controller (PID is common) with a delay. This would need some tuning, as it should encompass the effects of driver reaction time, engine response, wheel characteristics, etc. But there's a good chance that even a simplistic response model would be sufficient given that these trucks usually have a lot of inertia and the driver state is usually either "I want to continue to go 55 mph" or "I want to stop". $\endgroup$ – neocpp Jan 2 '17 at 3:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.