Dynamic Simulation (Rotation-Translation with Gravity)

[inventor97VJT]

I'm attempting to simulate a heavyweight flywheel array that is powered by a falling Deadweight (under the force of gravity).  The Deadweight falls, and the energy is transferred from the linear motion into rotational motion of a driveshaft through a pulley.  To simplify the model for Dynamic Simulation, I've removed pretty much all the unnecessary components/geometry.  I've constrained the flywheel array (which includes the driveshaft) to an axis to rotate around.  The Deadweight is concentrically constrained to a vertical shaft through which it falls/travels.

I've created a Motion Constraint (rotation-translation) between the deadweights linear motion, and the rotation of the flywheel array that is equivalent to the ratio between the Pulley (not pictured or included in the model for simplification).  The relationship is approximately 3 revolutions per meter (so 45 total revolutions of the flywheel array as relates to the 15m fall distance of the Deadweight).

In the Assembly Environment, everything works as expected.  If I "drag" the deadweight along it's fall axis, the flywheel array behaves as expected.  HOWEVER, once I enter the Dynamic Simulation environment, and attempt to simulate the manner in which Gravity would cause the Deadweight to fall (linear), thusly transferring energy (rotational) to the flywheel array, things seem to break.  As you can see in the video below, the relational rotation rate of the flywheel array is not proportional (as in the constraint settings) to the distance traveled by the Deadweight during Dynamic Simulation...  The Deadweight itself travels the 15m distance within >2 seconds (which is correct), but the Flywheel Array is barely moving at all.

Anyone have any ideas why?  Any help is much appreciated! 

Here's the pack-and-go model: FLYWHEEL SIMULATION ASSEMBLY ZIP 
NOTE: THE ASSEMBLY FILE IS FOUND UNDER:
\Workspaces\Workspace\TECHNICAL DATA\V1\ASSEMBLIES\V1_DEADWEIGHT_ASSEMBLY_MODEL (sim testing.iam

[JDMather · Accepted solution]

@inventor97VJT 

You have the wrong Joints for this assembly.

You should have Revolution, Prismatic and Rolling:Cylinder on Plane Joints.

[inventor97VJT]

@JDMather Thank you so much for helping.  I went and watched a half dozen or so videos on your youtube channel (stellar work btw!), and after getting my head around things a bit more clearly, I was able to make it work nicely!  In fact (funny story) in reviewing the Output Graph, I noticed a strange unexpected "wobble" that (in closely reviewing the model) turned out to be an axially misaligned flywheel (lol).  After correcting the geometry and rerunning the simulation, the velocity curve was nice and clean, and that tiny wobble resulted in a 30% loss in power over the 45s cycle!  Crazy (see below)!  Just thought you'd find that amusing 😆

[inventor97VJT]

@JDMather 

On a separate note (at the risk of looking such a generous gift horse in the mouth 😬), I'm setting up the second side of the system now, and was able to get everything setup and working, similar to the first side (a couple revolute joints, a belt joint, and no need for the prismatic in this case), and it kind of works...  While my assembly interacts correctly in dynamic motion mode (components move with applied force when dragged), the "initial forces" aren't behaving correctly in the actual simulation... In the model, I have the same flywheel array from the previous post linked to a generator via a chain drive. Both have revolute joints at their respective mockup shafts, with a belt joint connecting them.

In this case, the generator is a real-world object with certain empirical (known) performance data.  The one that's most important to me (in terms of simulation) is the Drag Torque Curve... 

As you can see, the drag torque (functionally acting as 'resistance' to the angular motion of the flywheel array) is equal to ~1000nm at ~1800rpm (the intended operational speed of the generator).  The Chain Drive has a 3:1 ratio (FW Array = ~500-600rpm, Gen = 1500-1800rpm). In an effort to simulate the decay of velocity of the flywheel array when the generator is engaged (and thus imparting drag torque in opposition to the flywheel array motion), such that I can determine the approximate time the flywheel inertia will keep it within a defined operational velocity window, I'm running into issues that are (most likely) just my own simple ignorance in not understanding how to properly use the simulation functions surrounding initial loads, and other similar features.  Hoping you can help (and that you're so inclined).

I'm trying to set an initial velocity (Flywheel Array) of 500rpm, and am setting the Joint Torque of the Generator side to 1000Nm... Although this torque decreases with speed, I'm only trying to model the scenario (for now) to graph the velocity decay with this static drag torque value in place.  I'm not sure this is the most physically correct way to do it (alternatively, a friction value for the generator joint?  Some dynamically defined curve like the table above in one/the other of these two approaches?)... I'm definitely open to your thoughts on this if you're open to providing them...

In any case (ACTUAL question lol), upon activating the simulation, the assembly counter-rotates as expected (the joint torque setting seems to counter-rotate the jointed assembly as expected, if not really realistically, since it's a constant value, which is fine)...  However, when I manually set the initial velocity of the flywheel's revolute joint to 3000deg/s (500rpm), the flywheel starts abruptly, comes to an instantaneous stop, and then begins counter-rotating with the action of the other joints constant-torque setting.  Any ideas on why this is happening?  How do I properly set an "initial velocity" for a component in the simulation, such that I can observe it decaying under the other loads present as function of time in the output graph?

Again, thank you VERY much for the help.  It's sincerely appreciated.  I'm a quick study, but I've been banging my head against the wall trying to figure out what I'm not seeing here.  The documentation for these features is woefully sparse/non-communicative.

Here's the model: DRAG TORQUE MOCKUP ASSEMBLY (DROPBOX) 

Here's a couple-minute video of me running it:
VIDEO (ON DROPBOX VIDEO SHARE) 

[JDMather]

@inventor97VJT 

You have to use the Input Grapher to define the motion.

https://www.youtube.com/playlist?list=PLp5izJt_zvN08mpY4UcYrzJV2N6QOLAIR

https://www.youtube.com/playlist?list=PLp5izJt_zvN0GnwEUFR1zE06nQvDAWxz2

[inventor97VJT]

Thanks Again @JDMather!  After tweaking with it a bit this morning I was able to get it to work!  The simulation is definitely affected in a (relatively) expected way by the Joint Torque formula I'm using...

My only question (in your opinion/experience), is this an appropriate (or rather, 'best') approach to accurately simulating the effect of a known Drag Torque imposed by a generator in operation?

I've used this formula to relate to the velocity of the joint, in an effort to closely approximate the Drag Torque value as a function of said speed: (((V[1] (Revolution:1)*.16))/2)*-1

This essentially is derived from the known Drag Torque/vs/RPM from the generator spec (about 1/2 the RPM in Nm), and I'm multiplying it by .16 to convert to RPM from DEG/s, then multiplying by -1 (to make it an opposing force)...

Do you think this is an accurate way to simulate the approximate decay in Angular Acceleration imposed by the generator?

Is there a better or more accurate way to simulate this that I'm unaware of?  Friction setting of some sort?

Thoughts?