Prev: Adding a "global" title to a group of subplots
Next: Lowpass Butterworth Filter
From: david s on 23 Oct 2009 09:49
Hi,

I get the following error while i try to run my simulation.
??? Derivative input 1 of 'test/x9dot//Integrator' at time 10 is Inf or NaN. Stopping simulation. There may be a singularity in the solution. If not, try reducing the step size (either by reducing the fixed step size or by tightening the error tolerances).

does anybody know what is the problem in this case .
thanks
From: Arnaud Miege on 26 Oct 2009 05:47

"david s" <david.sabine760(a)gmail.com> wrote in message
news:hbsc8v$lmf$1(a)fred.mathworks.com...
> Hi,
>
> I get the following error while i try to run my simulation.
> ??? Derivative input 1 of 'test/x9dot//Integrator' at time 10 is Inf or
> NaN. Stopping simulation. There may be a singularity in the solution. If
> not, try reducing the step size (either by reducing the fixed step size or
> by tightening the error tolerances).
>
> does anybody know what is the problem in this case .
> thanks
>

Well, yes. The derivative input of the integrator block mention is Inf or
NaN at t = 10s, as the error message suggests. This indicates that your
model is numerically unstable. Which solver are you using? What absoluate
and relative tolerances are you using? What is the model of? Is it
numerically stiff as far as you can tell?

Arnaud


From: david s on 26 Oct 2009 06:59
"Arnaud Miege" <arnaud.miege(a)nospam.mathworks.co.uk> wrote in message <hc3r7l$el3$1(a)fred.mathworks.com>...
>
> "david s" <david.sabine760(a)gmail.com> wrote in message
> news:hbsc8v$lmf$1(a)fred.mathworks.com...
> > Hi,
> >
> > I get the following error while i try to run my simulation.
> > ??? Derivative input 1 of 'test/x9dot//Integrator' at time 10 is Inf or
> > NaN. Stopping simulation. There may be a singularity in the solution. If
> > not, try reducing the step size (either by reducing the fixed step size or
> > by tightening the error tolerances).
> >
> > does anybody know what is the problem in this case .
> > thanks
> >
>
> Well, yes. The derivative input of the integrator block mention is Inf or
> NaN at t = 10s, as the error message suggests. This indicates that your
> model is numerically unstable. Which solver are you using? What absoluate
> and relative tolerances are you using? What is the model of? Is it
> numerically stiff as far as you can tell?
>
> Arnaud
>
Hi Arnaud,
Thanks for your response . I tried to use the ode45 solver and the ode23s solver but i got the same error message . also i used relative tolerances =1e-3 and absoluate tolerances = auto . My model is a fuel cell model , it is composed of 6 inputs (the pressure of hydrogen ,the pressure of oxygen, the pressure of water, the humidity and the temperature) ,of 5 state variables ,and of a 2 outputs (the voltage and the current ).
If it is possible i can send you by email my simulink model that i did and the differential equations that describe this state space model .
thanks in advance

David
From: Arnaud Miege on 26 Oct 2009 07:46

>>
> Hi Arnaud,
> Thanks for your response . I tried to use the ode45 solver and the ode23s
> solver but i got the same error message . also i used relative tolerances
> =1e-3 and absoluate tolerances = auto . My model is a fuel cell model
> , it is composed of 6 inputs (the pressure of hydrogen ,the pressure of
> oxygen, the pressure of water, the humidity and the temperature) ,of 5
> state variables ,and of a 2 outputs (the voltage and the current ).
> If it is possible i can send you by email my simulink model that i did and
> the differential equations that describe this state space model .
> thanks in advance
>
> David
>

Try using ode15s or ode23t instead. Try setting the absolute tolerance to
something else other than auto, e.g. 1e-4. As the error message suggests,
try tightening the error tolerances to see if it improves (keep absolute at
least one order of magnitude smaller than relative). If it still doesn't
work, you might need to look at your model and see if there are elements
that could be modelled in a different manner or changed (for example, to get
rid of algebraic loops).

You can also try to contac technical support:
http://www.mathworks.com/support/contact_us/index.html

HTH,

Arnaud


From: david s on 26 Oct 2009 12:57
"Arnaud Miege" <arnaud.miege(a)nospam.mathworks.co.uk> wrote in message <hc4266$5j9$1(a)fred.mathworks.com>...
>
> >>
> > Hi Arnaud,
> > Thanks for your response . I tried to use the ode45 solver and the ode23s
> > solver but i got the same error message . also i used relative tolerances
> > =1e-3 and absoluate tolerances = auto . My model is a fuel cell model
> > , it is composed of 6 inputs (the pressure of hydrogen ,the pressure of
> > oxygen, the pressure of water, the humidity and the temperature) ,of 5
> > state variables ,and of a 2 outputs (the voltage and the current ).
> > If it is possible i can send you by email my simulink model that i did and
> > the differential equations that describe this state space model .
> > thanks in advance
> >
> > David
> >
>
> Try using ode15s or ode23t instead. Try setting the absolute tolerance to
> something else other than auto, e.g. 1e-4. As the error message suggests,
> try tightening the error tolerances to see if it improves (keep absolute at
> least one order of magnitude smaller than relative). If it still doesn't
> work, you might need to look at your model and see if there are elements
> that could be modelled in a different manner or changed (for example, to get
> rid of algebraic loops).
>
> You can also try to contac technical support:
> http://www.mathworks.com/support/contact_us/index.html
>
> HTH,
>
> Arnaud
>
thanks again for you,
i found that in the equation 5 of case (1) in my s_function m-file which is :
dx5/dt=[(-1.2e10*u(1)*x(5))/( u(6)*log(abs(1.92e5*u(1)))+log(abs(1- 6432.3*u(1))))] - 1e6*u(1);
when i cancel the last term of the equation which is ( 1e6*u(1) ) and try the simulink there were not any error message and i get a good output ,so i think that the problem is in this equation .
where u1 is a pulse signal or a constant (u1 = 2.5e-6) and u6 = 353
have you any idea about how can i play on this equation ?
thanks
 |  Next  |  Last
Pages: 1 2
Prev: Adding a "global" title to a group of subplots
Next: Lowpass Butterworth Filter