Hi I have created a tangent and i want this tangent to be used as a force for the smoke.could some one share a simple file to start ))

any help?

you could either use the gas calculate dop to manipulate the velocity field or try gas advect to get the right motion.
I set up a basic scene that shows how to use both methods,
maybe Alejandro can contribute more to your question, he did a really nice smoke path follow test a while ago:

http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=21558 [sidefx.com]

anyway, check out my scene file,

regards,

Manu

hi Manu
Thanks for your help
I saw the file of alejandro,but little complicated as I was a beginner in DOPs

There is another newer technique in H12 that uses the Pump Shelf tool to take an object's velocity and affect any fluid simulation (Smoke, Pyro, Flip) to achieve a similar result. It is designed to be straightforward and all the pieces are put in place for you by the Pump shelf tool.

The Pump shelf tool couldn't be easier to inject your custom geometry point velocities in to a Flip, Smoke or Pyro simulation. Just craft your geometry velocity “v” attribute as you wish. With an existing fluid simulation in place, just select the Pump tool from the shelf, select the geometry, hit enter, select the fluid container to add the pump to, hit enter. Done.

A lot of work went in to H12 to avoid the workflow presented in the attached files above. Yes there are so many ways to do things in Houdini and yes these files are valid examples but given the H12 workflow, completely unnecessary to just get smoke to follow geometry velocity point attributes.

Just the ability to easily control the pump mask and to inject a bit of curl noise in to the vel fields is worth it to use the Pump shelf tool.

See the modified attached hip file with lots of notes to make it self-explanatory.

Specific comments on the previously attached files:
Custom Advection Method
The custom advection method is fine but not using the default vel fields to do the advection means that rest, temperature, fuel, and all the other fields will not be carried along with your custom velocity field unless you manage this yourself.
Advecting all these fields is straightforward and what you need to advect is clearly illustrated inside the Smoke Solver.
Unless you have a very clear reason to advect fields with other custom velocity fields, don't. The Gas Advect DOPs can be expensive. They will take more time and more memory.

The Gas Calculate DOP is not supported by the GPU and is somewhat complicated to use with all it's various options. Houdini12 introduces the Gas Linear Combination to combine fields in various ways in fluid simulations that is GPU friendly. In general if you don't see the “Use OpenCL” toggle then it won't be gpu friendly.

Custom filed with Gas Calculate Approach
This method was common in Houdini11 when there was no clear way forward to inject custom geometry fields in to the Simulation. The custom field is persistent and has a constant memory overhead while the Houdini 12 pump approach adds the temporary fields for you and tries to run as efficiently as possible allowing you a little more memory room to scale the simulation up.

hi Jeff, thanks for clarifying things, appreciated!
The only thing that keeps me from using those shelf tools is that they create all those extra bunch of nodes that one usually doesn't need and I experienced that it slows the simulation down quite a bit.
Anyway, I like how the source volume DOP works, I will try to adapt this workflow where one creates a tempfield first using the gas match field microsolver, loads the SOP velocity field with a SOP solver and then uses a couple of gas calculate DOPs to pre-multiply the velocity mask and copy the tempvel field onto the vel field.

I learned a lot looking through your file,

best,

Manu

I hear you wrt the extra nodes and the shelf in general but the number of nodes in this case is far fewer as far as the artist is concerned and the speed of set-up is way way faster and you end up with an even better sim as it would take another few minutes just to add the VOP networks to add the curl to the incoming velocity.

Arguably yes there are a few assets that get placed down that contain many more nodes but for the most part the contents are bypassed depending on the functionality you want.

You can always use the Performance Monitor to see which operators are taking the time and then debug the network. This will tell you what nodes are taking the longest to evaluate.

It is great that many of us want to dive in and construct our own simulation bits from scratch but at the end of the day, artists need to drive this stuff and artists don't care how few or many nodes there are as long as they have the controls and the look that they want in a timely fashion.
No way we'll take the control away but packaging up these tools are critical if you want artists to use them day-in and day-out imho.
Look to the shelf as a decent work-in-progress to this end.

yes Jeff, I totally agree with you that artists shouldn't need to dive into the big solvers to manipulate them on a much lower level and instead just press some buttons on a nice interface.

I actually have a question to the usage of the gas linear combination DOP.
I set up a quick network that uses the technique of the Source Volume DOP to import the SOP velocity and ran into a problem.
I attached a hipfile, the dopnet right now is using Gas Calculate DOPs (red colored nodes) which work fine but as soon as I replace them with gas linear combination DOPs the setups stops working. (red nodes next to the gas calculate nodes, not connected right now).

what's going on there? Btw the Source Volume DOP doesn't use gas linear combination either …

In this example and my test I just did only the copy method seems to be working with the Gas Linear Combination…

Yeah the Gas Calculate is used all over the place in there.

Where's that white flag emoticon when you need it… ops:

the file posted above actually failed due to human error -
Julien Depredurand was so kind to show me how to use the gas linear combination DOP properly, I attached the fixed file - and it's working.
Learned something new …

cheers!

I didn't notice the “29” multiplier in the first gas calculate… Once I saw the global scale preference was set to 0.1 for metres, I picked this up and the two DOPs work just fine and give the same result.