non uniform pressure

Hi,
you have to insert a coordinate system oriented as Cosmos expects. You may use the global coordinate system only if it is already in the "needed" orientation.
If the gradient is in the wrong direction, often it is much simpler to reverse the sign(s) in the coefficient(s) than to reverse the coordinate system.

Regards

Cheers, but how do you start at zero one end and max pressure at the other

Hi,
I don't know if things have changed since release 2004, but back in this time a gradient was defined as a formula in the coordinate system, such as
p(x,y)=a*x^2+b*y^2+c*x*y+d*x+e*y+f
So, assuming "x" is your direction about which the pressure varies, and assuming also that the pressure doesn't vary also along "y", and that the variation law is linear (for example, hydrostatic gradient...), we have:
a=0
b=0
c=0
d=rho*grav/1000 [Pa/mm]: this is the gradient slope
e=0
f=0: this is the pressure value when x and y are both equal to zero, i.e. in your case it's the "start" pressure, null at x=0.

Regards

Thanks mate, that works, but I found that I had to put the origin at the zero pressure end. At the other end it didn't matter if I revesred the Y sign it still put high pressure at the top.

Hi,
of course! In fact:
- if you reverse the sign of the "Y" coefficient, you reverse the way in which the gradient "increase" (or "decrease")
- you can offset the gradient wrt the origin of the coord sys, that's the meaning of the "f" coefficient (the one which gives the pressure value at (X;Y)=(0;0)). That means that, if you have an hydrostatic pressure gradient with 5 [m] head, rho=1000 [kg/m^3] and g=9.81 [m/s^2], then
d=1000*9.81 [Pa/m] (X pointing DOWNWARDS!!!)
f=0 IF and ONLY IF the origin of the coord sys is at the zero pressure level
f=d*5 IF the origin of the coord sys is at the bottom of the gradient (the "max pressure" level)

WARNING: my assumption is that this coefficient "f" exists (I'm basing all this only on memory, 'cause I don't use Cosmos any more). If the equation doesn't have this "f", then it's "like if" f=0 and so yes you have to put the origin of the coord sys where at the "zero-level" of the gradient.

Regards