HFSS Power Flow Calculations

Status
Not open for further replies.
A

ashish.mw

Full Member level 2
Joined
Jul 18, 2011
Messages
121
Helped
15
Reputation
30
Reaction score
14
Trophy points
1,298
Location
India
Activity points
2,236
Hi all,

I have simulated a rectangular waveguide WR284 in HFSS with the WavePort 1 in excited state and input power of 1W. I wish to calculate the net power flow through input port (read Surface1) and output port (read Surface2).

I am using HFSS Field Calculate to find the power flow through the above two surfaces which is given by the closed surface integral of Poynting Vector.

The formula for power flow computation goes like:

Scl : Integrate(Surface(Surface1), Dot(Real(Poynting), SurfaceNormal))

and

Scl : Integrate(Surface(Surface2), Dot(Real(Poynting), SurfaceNormal))

But the above one yields very strange values that I could not comprehend.

Power through surface1 = -1.0113 W
Power through surface2 = 1.0128 W

I read somewhere that power flow computation works in case of incident plane wave excitation but I am using waveport. Can somebody confirm on this?

Please help in understanding the phenomenon.

Thanks.
 

Hi
I have tried to simulate the model in the way you have described.
I used the WR284 ratings and then defined two PORT1 and PORT2 as waveports solution frequency 2.998 GHz and I have implemented the same power flow formula in HFSS calculator and I get
Power through Port 1 1.03887 W
Power through Port2 -1.03887 W
Which makes sense as all the power from incident port is absorbed by other port.
 

Thanks for reply.

Can you please explain where from does this value of 1.03887 W comes into picture. When I look at the Edit Sources, I can see that PORT1 is coupling 1W of power. So what is this 1.03887 W?

What about the concern that inorder to do Poynting vector integration, one need to have Incident Plane Wave excitation? Is is necessary? Does this waveport method actually correct?

My main aim is to use HFSS to compute what amount of power comes out when 1W is fed in (for more complex shapes like bends, twists etc. WR284 is the simplest one to start with). Is it possible to do?

Thanks
 

 


The waveport should work fine. A few thoughts on your problem:

How converged is the simulation? I would be sure to extra-converge it to make sure that everything is modelled properly, and the extra power is not just numerical error.
How many modes are you exciting? It should be only one.
The negative sign is fine, that should just be direction with respect to the surface normal.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…