hexagon logo

For Modified Newton-Raphson method and Full Newton-Raphson method

How does Marc uses these two kind of methods during calculation? Does these two kind of methods ​have a effect on final result?
In my opinion, these two methods  only influence the convergence rate and  if the calculation convergenced then the result should be similiar. I do a comparation test by changing this option in Marc, and the result shows big difference.
Thanks.
  • Hi Ddenzel,
     
    In the Full Newton-Raphson method, The tangent stiffness matrix is reformulated for each iteration (recycle).
     
    In the Modified Newton-Raphson method, the tangent stiffness matrix is formulated only once per increment.  
     
    If contact or sudden material nonlinearities occur, reassembly cannot be avoided. The modified Newton-Raphson method is effective for large-scale, only mildly nonlinear problems. When the iterative solver is employed, simple back substitution is not possible, making this process ineffective. In such cases, the Newton-Raphson method should be used instead.
     
    If you are expecting contact and material nonlinearities, then use the default N-R method.
     
    Regards,
    Kiran
  • Try smaller convergence tolerance and see if the results improve.
  • Thanks,I changed convergence tolerence , used a absolute value of residual and smaller step,the result shows similar with the one used full Newton-Raphson. It need more cycles to get convergence in each increment than full Newton-Raphson mehthod.
    By the way,is the tang​ent stiffness matrix calculated only once or recalculate after serveral cycles in one increment when using Modified Newton-Raphson method?