A Newton solution for 3 coupled nonlinear elements has failed to converge within the iteration limit MAXZNO = 50. e! n: H0 F5 @+ N1 V" W
The current residual is 3.77841495E+02, which exceeds the sloppy convergence tolerance EPSTOP = 1.00000000E-01. The first of% I& a/ w9 W- K9 G d6 S
the coupled, true, nonlinear elements (in order of data input) is located in row 7 of the nonlinear element table, and it& s$ C; ^5 L9 B$ @- ?& H @
connects node "X0007C" with node " ". The final coupled element is in row 9, and the matrix rank is 3. Finally, the . f+ @' i* d! l7 vsimulation time is T = 1.08700000E-05. Possible corrective actions include a decrease in the time-step size DELTAT, an increase0 d: z: F# z6 c& T: e0 ~
in the iteration limit MAXZNO, an increase in the divergence tolerance EPSTOP, and artificial splitting of the subnetwork in 4 ?" t; X9 T2 O' V+ ]5 Kquestion by distributed-parameter lines (to reduce the number of coupled elements). In addition, it is possible for the user to % E8 V2 ^- M$ A: d( pmake a solution impossible by unrealistically limiting the Newton corrections. The user should remember that there can be no , g: a/ P* E' u6 O+ s; asolution if his bound on arrester voltage is less than the true solution voltage. The bound has a default value of 1.5 per unit, ' E. V7 t% u) \- Valthough it can be changed by a"ZINC OXIDE" special request (columns 57-64, variable ZLIN(2)). Is this parameter involved? The x4 q" P3 j9 i0 tnumber of solution variables that are being limited this way is 0. If this value is positive, think hard about ZLIN(2).