In this paper, the effect of using a simple back interpolation (BI) scheme with the preconditioned conjugate gradient (PCG) solver for the solution of global stiffness equations in 3-D analysis is considered. Numerical results obtained show that if a series of properly constructed adaptive meshes are available, the use of the BI scheme can effectively reduce the number of iterations needed to achieve a converged solution. More importantly, it is found in many cases for the last few adaptive meshes generated, the number of iterations needed is independent of the number of equations of the meshes. Consequently, the computational cost needed to solve the global system is only proportional to the number of degrees of freedom of the meshes which implies a significant savings in CPU time. The same set of numerical results also show that the total cost of the adaptive refinement can be minimized by gradually reducing the relative error of the solution so that the target accuracy is achieved in a reasonable number of refinements.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £75 +P&P)