QUESTION IN GENERATION OF BODY-FITTED MESH

James Zhu

When programming a code in generation of body-fitted mesh for a cirque using the Laplace equation,I met a problem: the first,the outputted x and y coordinates are wrong,clearly,the point almost all locate in the secant,especially in the two endpoints(the secant coinsidence with the x-axis) the second,I use multigrid method to solve the system of elliptical equation with the FMG+FAS,5 mesh level are used,on the forth 4 level from the coarsest grid(5) to the finer grid(2),the V-cycle converge very fast,but in the finest grid,the V-cycle converge in the beginning,but donot converge finally(the residual from the Gauss-Seidel iteration is 1.6...,too big),please help me!

thank you!