who we are      products      services      partners      users      contact
 

 

 

 

 

download







Boosting CFD


Because the memory was extremely rare and expensive in the early developments of computational fluid dynamics, only algorithms with small demands on computing power could survive. In this context, the best strategy was the Patankar-Spalding SIMPLE algorithm that segregates the momentum and continuity equations, in spite of a well known poor convergence as the number of cells increases.


Today industrial computational fluid dynamics stands at the crossroads. The mean amount of affordable memory on workstations would allow to refine models but the resulting time for convergence would be excessive.


Then the door was opened to other ways of thinking. Couldn't we compute faster using that underemployed memory? This is precisely the idea of MIGAL: the algebraic coupled multi-grid solver that will boost your CFD applications.

Speed-up


Because of its strong velocity-pressure coupling technique and its multi-grid approach MIGAL produces great robustness and large speed-up for many flows. As shown below, when plugged into the commercial software Phoenics, MIGAL impressively reduces the elapsed time for convergence.





Memory requirement


Of course such a speed-up has a cost. The storage of the coefficients of the coupled equations and the additional storage for the multi-grid procedure increase the minimum amount of memory. But this enlargement of the memory requirement is not an obstacle for 2D flows and is already affordable for 3D flows.





In order to approximate the additional memory required for MIGAL one should use the following rules in Mb/Point :

 


 

Staggered

Co-locative


2D

4.E-4

3.E-4

3D

7.E-4

5.E-4

 

 

Application range


Initially designed for the velocity-pressure coupling in both staggered and co-locative arrangements, MIGAL has been continually improved over several years to become a robust software able to face industrial flows.


Today, MIGAL can also apply its multi-grid feature to single variable equations as well as many kinds of coupling like the k-epsilon turbulent model or the velocity-temperature coupling.


If you want to know how MIGAL could fit your own CFD developments please, feel free to contact us.