<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
 <channel>
  <atom:link href="http://www.comsol.nl/community/forums/general/rss/thread/3859.rss" rel="self" type="application/rss+xml"/>
  <title>COMSOL Forums: pb of variable step for solve stiff ODE</title>
  <link>http://www.comsol.nl/community/forums/general/thread/3859/</link>
  <description>Most recent forum messages</description>
  <pubDate>Tue, 16 Mar 2010 06:14:08 +0000</pubDate>
  <image>
   <title>COMSOL Forums: pb of variable step for solve stiff ODE</title>
   <url>http://www.comsol.nl/shared/images/logos/comsol_logo.gif</url>
   <link>http://www.comsol.nl/community/forums/general/thread/3859/</link>
  </image>
  <item>
   <title>Re: pb of variable step for solve stiff ODE</title>
   <link>http://www.comsol.nl/community/forums/general/thread/3859/#p10087</link>
   <description>Hi Isabelle&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Your translation is OK ;)&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
First my references when I look into these subject is to go back to a couple of books I like, as they are close to COMSOL math notation:&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
&amp;quot;Introduction to computation and modeling for Differential Equations&amp;quot;, by L. Edsberg, Wiley, 2008&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
And as COMSOL cookbook well written with math references is the:&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
&amp;quot;Multiphysics Modelling with Finite Element Methods&amp;quot; by B.J. Zimmerman, World-Scientific, 2006&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Further, if I understand your well, your issue is in the transient solver settings, it has, at least in my 3.5a, many options, mostly I let COMSOL choose, and even if I say&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
t=range(min,step,max) &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I understand that the solver decides itself upon the steepness of the gradients in the model to select single or multiple intermediate points, but these are not normally stored in my solution vector, unless you define &amp;quot;Time steps from solver&amp;quot;, these options are selected in the time stepping sub-tab. I have noticed that solver step mode &amp;quot;free&amp;quot; might skip short duration features, then I impose rather &amp;quot;strict&amp;quot; and ensure my time steps cover well these features, or I use intermediate, then it set at least one more point inbetween my defined time steps. This is rather well described in the user guide (guide.pdf) in the solver section &amp;quot;Time-Dependent&amp;quot; stepping, in the V3.5a&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
As for me in V3.5a the adaptatif is there by default, but it might be different with femlab if you have an older version&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Good luck&lt;br /&gt;&#13;
Ivar</description>
   <pubDate>Tue, 16 Mar 2010 06:14:08 +0000</pubDate>
   <guid isPermaLink="false">3859.1268720048.10087</guid>
  </item>
  <item>
   <title>pb of variable step for solve stiff ODE</title>
   <link>http://www.comsol.nl/community/forums/general/thread/3859/#p10069</link>
   <description>thank you Mr Ivar&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
actually, I do not know wellEnglish , why I make my automatic translation.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
answering your questions:&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I want to solve a system of several equations ODE type stiff , this type of equations, it does not define a constant step, but it turns adaptatif at each new solution.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
global equation is not an application itself, but a tool embedded in FEMLAB to help us solve equations of type f (u&amp;quot;,u',u) = 0.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
my problem is to declare adaptatif step in box of Time-Dependent???&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
thank you infinitely</description>
   <pubDate>Mon, 15 Mar 2010 21:35:07 +0000</pubDate>
   <guid isPermaLink="false">3859.1268688907.10069</guid>
  </item>
 </channel>
</rss>

