[Martins.Brics]

Hi Robin,

you are absolutely right. This means that falling down ball which rotates counter-clockwise will lean right as seen in the video. This means that everything is fine with your code, if original rotation matrix is used. Just by manipulating the rotation matrix I introduced an error and got lost with what is correct outcome of the Magnus effect. Somehow I thought that the falling ball in front of itself increases the density of media it falls in. Thus, by rotating counter-clockwise the ball at the bottom moves more media particles to rihgt than to left at the top. Therefore, do to 3rd Newtons’s law, the ball should move left. But that is obviously wrong. Bernoulli effect at the sides of the ball is lager. I just ignored it and thus predict wrong behavior. Somehow I was convinced that I am correct and was able to convince my colleges that I am right.

Sorry for wasted time!

Best regards
Martins Brics

[Martins.Brics]

Hi Robin,

sorry for long not response. After vacation I had to deal with another project.

But I finally understood why the effect suddenly disappeared. I changed the ration matrix from

rotationMatrix[0] = cos;
rotationMatrix[1] = -sin;
rotationMatrix[2] = sin;
rotationMatrix[3] = cos;

to

rotationMatrix[0] = cos;
rotationMatrix[2] = -sin;
rotationMatrix[1] = sin;
rotationMatrix[3] = cos;

and forgot that I did the changes. Changed indicis [1]->[2] and [2]->[1]. This is equivalent to rotation by negative angle. In version 1.4 I have to change to do the same trik in order not to observe visually reversed Magnus effect.

videos can be seen here

[Martins.Brics]

Dear Robin,

Thanks for replay. Indeed I applied positive Torque when got clockwise rotation.

the piece of modified code:
T torque=1.0e-7;
_vecOfExt_Torque.push_back(torque);
_vecOfHydroDyn_Torque.push_back(sumF[2+4*iInd]);
indicator[iInd]->setAlpha2( (sumF[2+4*iInd]+torque) / indicator[iInd]->getMofi() );

To determine how the angle is defined I repeated simulation with cube, where I can see rotation direction. with cube for a positive external torque I got that angle theta also increased so I assumed there is the same definition of angle – positive for clockwise rotation.
Later I will look at velocity fields, but next week I am on vacation.However, I am bit confused today as I got that cuboid2d as circle deflects in correct direction now. video

Best regards
Martins Brics

[Author]

Posts