Question Are the approximations used in TRALIN / SPLITS conservative?
Answer The approximations used in the Tralin / Splits (namely, using the impedance per unit length computed in Tralin along with the Span-Scaling feature of the circuit model in Splits) lead to conservative induced voltages when compared to the results obtained with a finite element approach, such as the one used in HIFREQ. The difference is quite small, however, being on the order of 3 to 4%. Moreover, the results improve when the length of the line being modelled increases.
The above conclusions are based on extensive testing. They can be understood from the fact that TRALIN assumes that the lines are infinitely long when computing the self and mutual impedances, which tends to overestimate these impedances. For a given current, these larger impedances yield larger GPRs.