Discussion:
[COMSOL_Users] This error drives me crazy in k-e turbulence (trainsient)and moving mesh module
viroose1
2009-05-11 03:50:35 UTC
Permalink
HI all~
i am working with simulating the rotating Fan.
i am using k-e trainsient turbulence module and moving mesh module.
after all settings done,i try to simulate the model.
but this error occurs every try..T.T


Error: 6199
Attempt to evaluate real square root of negative number.
- Function: sqrt
Failed to evaluate expression.
- Expression: nojac((min((gijgij_chns^-0.25),((U_chns*rho_chns)/((eta_chns+etaT_chns)*48)))*(0.5*min((ck_chns*sqrt((((((res_p_chns*tauc_chns)/(nojac(sqrt((u2int_g1_chns/vol_g1_chns)))*nojac((rhoint_g1_chns/vol_g1_chns)))))^2+((res_u_chns*taum_chns))^2)+((taum_chns*res_v_chns)^2)))),U_chns))))
Failed to evaluate temporary symbolic derivative variable
- Variable: ***@VDN${***@1}
- Defined as: (((ux*g21_chns)+(uy*g22_chns))*(rho_chns*(-nojac((min((gijgij_chns^-0.25),((U_chns*rho_chns)/((eta_chns+etaT_chns)*48)))*(0.5*min((ck_chns*sqrt((((((res_p_chns*tauc_chns)/(nojac(sqrt((u2int_g1_chns/vol_g1_chns)))*nojac((rhoint_g1_chns/vol_g1_chns)))))^2+((res_u_chns*taum_chns))^2)+((taum_chns*res_v_chns)^2)))),U_chns)))))))
Failed to evaluate expression.
- Expression: diff((((-2*(eta_chns+etaT_chns)*ux+p)*test(ux)-(eta_chns+etaT_chns)*(uy+vx)*test(uy)-rho_chns*(u*ux+v*uy)*test(u)+gls_chns+shock_capt_chns)-(rho_chns*ut*test(u)))*(dvol_ale),{***@1})


all settings that i have done are Constants
( option- constant - i put in rho, eta, rpm)
and other settings are set by modules( ex: scalar variables..).

how can i find solution in this situation?
what means that messages?
Ravindra Dwivedi
2009-05-11 05:52:14 UTC
Permalink
Hi viroose1,
 
There are at least two reasons why you are getting this error.
 
1) either, you mesh is not dense enough at the critical points (depends on the geom. )
 
2)or, you need to start from a very low value of the governing parameter of your model. If you start from a very high value  of the governing parameter then COMSOL's turbulence models are not very stable.  
 
Ravindra,
 
 
 
 
 
 
 
 


--- On Mon, 5/11/09, viroose1 <***@yahoo.co.kr> wrote:


From: viroose1 <***@yahoo.co.kr>
Subject: [COMSOL_Users] This error drives me crazy in k-e turbulence (trainsient)and moving mesh module
To: ***@yahoogroups.com
Date: Monday, May 11, 2009, 3:50 AM








HI all~
i am working with simulating the rotating Fan.
i am using k-e trainsient turbulence module and moving mesh module.
after all settings done,i try to simulate the model.
but this error occurs every try..T.T

Error: 6199
Attempt to evaluate real square root of negative number.
- Function: sqrt
Failed to evaluate expression.
- Expression: nojac((min(( gijgij_chns^ -0.25),(( U_chns*rho_ chns)/((eta_ chns+etaT_ chns)*48) ))*(0.5*min( (ck_chns* sqrt((((( (res_p_chns* tauc_chns) /(nojac(sqrt( (u2int_g1_ chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))
Failed to evaluate temporary symbolic derivative variable
- Variable: shock_capt_chns@ VDN${test@ 1}
- Defined as: (((ux*g21_chns) +(uy*g22_ chns))*(rho_ chns*(-nojac( (min((gijgij_ chns^-0.25) ,((U_chns* rho_chns) /((eta_chns+ etaT_chns) *48)))*(0. 5*min((ck_ chns*sqrt( (((((res_ p_chns*tauc_ chns)/(nojac( sqrt((u2int_ g1_chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))) ))
Failed to evaluate expression.
- Expression: diff((((-2*( eta_chns+ etaT_chns) *ux+p)*test( ux)-(eta_ chns+etaT_ chns)*(uy+ vx)*test( uy)-rho_chns* (u*ux+v*uy) *test(u)+ gls_chns+ shock_capt_ chns)-(rho_ chns*ut*test( u)))*(dvol_ ale),{test@ 1})

all settings that i have done are Constants
( option- constant - i put in rho, eta, rpm)
and other settings are set by modules( ex: scalar variables..) .

how can i find solution in this situation?
what means that messages?
viroose1
2009-05-12 04:31:04 UTC
Permalink
Thanks Ravindra :-)

I don't understand your 2) method..
governing parameter means..like a density or viscosity ?
what i input the condition is only Constant..( rho eta rpm.. rpm has low values (3[1/min]) and rho, eta has air's properties. )
in boundary conditions, there is only open boundary..
so i think that there are no high values of settings..T.T
Post by Ravindra Dwivedi
Hi viroose1,
?> There are at least two reasons why you are getting this error.
?> 1) either, you mesh is not dense enough at the critical points (depends on the geom. )
?> 2)or, you need to start from a very low value of the governing parameter of your model. If you start from a very high value?of the governing parameter then COMSOL's turbulence models are not very stable. ?> ?> Ravindra,
?> ?> ?> ?> ?> ?> ?> ?>
Subject: [COMSOL_Users] This error drives me crazy in k-e turbulence (trainsient)and moving mesh module
Date: Monday, May 11, 2009, 3:50 AM
HI all~
i am working with simulating the rotating Fan.
i am using k-e trainsient turbulence module and moving mesh module.
after all settings done,i try to simulate the model.
but this error occurs every try..T.T
Error: 6199
Attempt to evaluate real square root of negative number.
- Function: sqrt
Failed to evaluate expression.
- Expression: nojac((min(( gijgij_chns^ -0.25),(( U_chns*rho_ chns)/((eta_ chns+etaT_ chns)*48) ))*(0.5*min( (ck_chns* sqrt((((( (res_p_chns* tauc_chns) /(nojac(sqrt( (u2int_g1_ chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))
Failed to evaluate temporary symbolic derivative variable
- Defined as: (((ux*g21_chns) +(uy*g22_ chns))*(rho_ chns*(-nojac( (min((gijgij_ chns^-0.25) ,((U_chns* rho_chns) /((eta_chns+ etaT_chns) *48)))*(0. 5*min((ck_ chns*sqrt( (((((res_ p_chns*tauc_ chns)/(nojac( sqrt((u2int_ g1_chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))) ))
Failed to evaluate expression.
all settings that i have done are Constants
( option- constant - i put in rho, eta, rpm)
and other settings are set by modules( ex: scalar variables..) .
how can i find solution in this situation?
what means that messages?
John Paul Bir Singh
2009-05-11 12:12:23 UTC
Permalink
Hi,

I was getting a very similar message but for a stokes flow problem (low reynolds number flow). And the error occured only in version 3.5a and not in 3.4. In the end it was found that in 3.5a crosswind stabilization was "on" by default which was causing this error. I am not sure if the same (or something similar) applies to your case.

My guess will be try to play with stabilizations- switch them on/off and see if that changes anything.
Best
John




________________________________
From: viroose1 <***@yahoo.co.kr>
To: ***@yahoogroups.com
Sent: Sunday, May 10, 2009 11:50:35 PM
Subject: [COMSOL_Users] This error drives me crazy in k-e turbulence (trainsient)and moving mesh module





HI all~
i am working with simulating the rotating Fan.
i am using k-e trainsient turbulence module and moving mesh module.
after all settings done,i try to simulate the model.
but this error occurs every try..T.T

Error: 6199
Attempt to evaluate real square root of negative number.
- Function: sqrt
Failed to evaluate expression.
- Expression: nojac((min(( gijgij_chns^ -0.25),(( U_chns*rho_ chns)/((eta_ chns+etaT_ chns)*48) ))*(0.5*min( (ck_chns* sqrt((((( (res_p_chns* tauc_chns) /(nojac(sqrt( (u2int_g1_ chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))
Failed to evaluate temporary symbolic derivative variable
- Variable: shock_capt_chns@ VDN${test@ 1}
- Defined as: (((ux*g21_chns) +(uy*g22_ chns))*(rho_ chns*(-nojac( (min((gijgij_ chns^-0.25) ,((U_chns* rho_chns) /((eta_chns+ etaT_chns) *48)))*(0. 5*min((ck_ chns*sqrt( (((((res_ p_chns*tauc_ chns)/(nojac( sqrt((u2int_ g1_chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))) ))
Failed to evaluate expression.
- Expression: diff((((-2*( eta_chns+ etaT_chns) *ux+p)*test( ux)-(eta_ chns+etaT_ chns)*(uy+ vx)*test( uy)-rho_chns* (u*ux+v*uy) *test(u)+ gls_chns+ shock_capt_ chns)-(rho_ chns*ut*test( u)))*(dvol_ ale),{test@ 1})

all settings that i have done are Constants
( option- constant - i put in rho, eta, rpm)
and other settings are set by modules( ex: scalar variables..) .

how can i find solution in this situation?
what means that messages?
Ravindra Dwivedi
2009-05-12 06:40:38 UTC
Permalink
Hi viroose1,
 
Earlier I guessed that you are solving your governing equations in dimensionless form. But, you are  not. So, what is the Ra value based on you rotation speed and fluid and flow domain parameters?
 
I would like to point out that k-epsilon ( specifically) and k-omega (to certain extent) are not good for cases where there is extreme streamline curvature. Please refer to "Turbulence modeling with CFD" by wilcox ( I guess that this is mentioned in the 6th chapter).
 
Ravindra,
 
 
 
 
 


--- On Tue, 5/12/09, viroose1 <***@yahoo.co.kr> wrote:


From: viroose1 <***@yahoo.co.kr>
Subject: [COMSOL_Users] Re: This error drives me crazy in k-e turbulence (trainsient)and moving mesh module
To: ***@yahoogroups.com
Date: Tuesday, May 12, 2009, 4:31 AM








Thanks Ravindra :-)

I don't understand your 2) method..
governing parameter means..like a density or viscosity ?
what i input the condition is only Constant..( rho eta rpm.. rpm has low values (3[1/min]) and rho, eta has air's properties. )
in boundary conditions, there is only open boundary..
so i think that there are no high values of settings..T. T
Post by Ravindra Dwivedi
Hi viroose1,
?> There are at least two reasons why you are getting this error.
?> 1) either, you mesh is not dense enough at the critical points (depends on the geom. )
?> 2)or, you need to start from a very low value of the governing parameter of your model. If you start from a very high value?of the governing parameter then COMSOL's turbulence models are not very stable. ?> ?> Ravindra,
?> ?> ?> ?> ?> ?> ?> ?>
Subject: [COMSOL_Users] This error drives me crazy in k-e turbulence (trainsient) and moving mesh module
Date: Monday, May 11, 2009, 3:50 AM
HI all~
i am working with simulating the rotating Fan.
i am using k-e trainsient turbulence module and moving mesh module.
after all settings done,i try to simulate the model.
but this error occurs every try..T.T
Error: 6199
Attempt to evaluate real square root of negative number.
- Function: sqrt
Failed to evaluate expression.
- Expression: nojac((min(( gijgij_chns^ -0.25),(( U_chns*rho_ chns)/((eta_ chns+etaT_ chns)*48) ))*(0.5*min( (ck_chns* sqrt((((( (res_p_chns* tauc_chns) /(nojac(sqrt( (u2int_g1_ chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))
Failed to evaluate temporary symbolic derivative variable
- Defined as: (((ux*g21_chns) +(uy*g22_ chns))*(rho_ chns*(-nojac( (min((gijgij_ chns^-0.25) ,((U_chns* rho_chns) /((eta_chns+ etaT_chns) *48)))*(0. 5*min((ck_ chns*sqrt( (((((res_ p_chns*tauc_ chns)/(nojac( sqrt((u2int_ g1_chns/vol_ g1_chns)) )*nojac(( rhoint_g1_ chns/vol_ g1_chns)) )))^2+((res_ u_chns*taum_ chns))^2) +((taum_chns* res_v_chns) ^2)))),U_ chns))))) ))
Failed to evaluate expression.
all settings that i have done are Constants
( option- constant - i put in rho, eta, rpm)
and other settings are set by modules( ex: scalar variables..) .
how can i find solution in this situation?
what means that messages?
Loading...