MES of flat-belt drives-ask for help

[lihouxin065]

MES of flat-belt drives  ,the result Appears to be wrong，could anyone help me ?

[lihouxin065]

there is the archive file of the model

[John_Holtz]

Hi @lihouxin065 

One thing that I see as wrong is surface group 3 on the driven roll is missing a surface. There is no contact between the belt and the driven roll on the half where you need it!

You might also need to do some of these:

1. Give the rollers (and the joint meshes inside them) and initial rotational velocity equal to 10 rev/sec.
2. Use a smaller time step (larger capture rate). (10 rev/1 sec)*(360 deg/rev)*(1 sec/200 steps) = 18 degrees per step. That is a large rotation angle for 1 step. I was thinking on the order of 2 to 5 degrees of rotation per step.

More importantly, what are you hoping to get from the analysis? Will the analysis provide those results? The analysis treats the belt as if it is molded to that "oval" shape. Once the roller rotates 180 degrees, the curved portion of the belt is no longer in contact with the roller. The tension in the belt is trying to "flatten" the curved portion and "bend" the flat portion. Will the results of trying to do that be accurate for you purpose?

[lihouxin065]

Thanks for your reply，This is just a preliminary  test model. I need to do a simulation task to get the time-varying curve of the average strain of the sensitive area of the strain-type load cell that supports the conveyor belt frame during the conveying process of the conveyor belt，Therefore, I need to learn how to implement the belt Simulation modeling  first.

[lihouxin065]

@John_Holtz Thanks for your reply，This is just a preliminary test model. I need to do a simulation task to get the time-varying curve of the average strain of the sensitive area of the strain-type load cell that supports the conveyor belt frame during the conveying process of the conveyor belt，Therefore, I need to learn how to implement the belt Simulation modeling first.

[lihouxin065]

@John_HoltzMy simulation settings are as follows, but the driven wheel does not rotate at the same speed as the driving wheel as expected, what am I doing wrong?

Model schemeconstraints of tow point on the joint( Driven pulley)constraints of tow point on the joint(Driving pulley)rotation displace of Driving pulleybelt cannot move in Z directionTension force on the joint on Driven pulleygravity in -Y directionsurface contact with smooth coulomb friction between belt and pulleysMES analysis parametersPulleys not rotate at the same speed as expected

[lihouxin065]

@John_Holtz  I have fixed the surface contact problem （two surface missed ）

[John_Holtz]

Hi @lihouxin065 

I do not fully understand what this means: "I need to do a simulation task to get the time-varying curve of the average strain of the sensitive area of the strain-type load cell that supports the conveyor belt frame during the conveying process of the conveyor belt". I am not convinced that you need to perform a full MES analysis where you try to simulate the physical motion of the belt. 

Stated another way, what effect does the rotation of the rolls and motion of the belt have on the "sensitive area"? Is there a faster way to simulate the effect without going through the time (and trouble) of moving the rolls and belt? 

The driven roll is not rotating at 10 rev/sec because it takes a lot of force to accelerate a 26 N roll from 0 rev/sec to 10 rev/sec over a time period of 0 seconds. (Answer: it takes an infinite force for an infinite acceleration.) There is not enough friction in the belt to do that! The results show that the belt is slipping on both rolls.

Assuming you do not use an alternative method to simulate the loads, then

• If the start-up of the rolls is important to the analysis, then you should accelerate the driven roll from 0 to 10 rev/sec over a time period of T.
• If the start-up of the rolls is not important, then you should give the rolls an initial angular velocity ("Setup > Loads > Initial Velocity") and give the belt an initial velocity. (You would need to split the belt into 4 parts to give a -X velocity to the top, +X velocity to the bottom, and angular velocity to the half wrap.)

Unless the belt is a thick piece of plastic that resists bending, I suggest that you use membrane elements for the belt. Membrane elements have no resistance to bending. They only have tension loads. (By using bricks and a plastic material, the stress required to bend and unbend the belt will not be realistic.)

[lihouxin065]

@John_Holtz what I want to do was like this ,could you give me a demo to do it in Simulation Mechanical 2017?

here is video link:Applications of Contact — Lesson 4 - ANSYS Innovation Courses

[John_Holtz]

It looks like they use a prescribe displacement to move the driver roll to create the tension in the belt, and then they user a prescribe rotation to rotate the driver roll. I suggest that you try the same approach.