What can be the reason of unexpected high damping during a simulation?
The damping of transient phenomena during a transient simulation is due to 2 causes: Physical damping or numerical damping.
- The physical damping is due to the behaviour of the modeled electrical system. The attached example shows a network with a single busbar and a single R-L-C shunt. The shunt is short-circuited showing swings on its own afterwards.
The swinging of the shunt is described by:
freq = 2 * pi * sqrt((4 * L/C - R^2)/(2 * L)) = 1000Hz
Damping time constants = 2 * L/R = 24ms
The simulation shows a (correct) damping result of transient behaviour about 5% in 72ms. The calculation case physical damping in the attached example has all numeric calculation options correctly set, and the simulation therefore
shows the correct transient behaviour of the swinging shunt.
- The numerical damping is due to the numerical integration, and may be larger than 0.0. Numerical damping may be due to too large step size, or due to too small damping factor. The effects of too small damping factor is shown in the study case SmallDamp. The resistance of the shunt has been set to zero, which should give undamped behaviour. The damping factor for the simulation, however, has been set to 0.8 on the advanced options page of the initial condition command. The result is a damped behaviour.
The effects of too large step seize is demonstrated in the study case SmallStep. The resistance is again zero, but the step size has been increased by a factor of 100. The result is a damped behaviour. The damping factor for the simulation is normally set to a value close to 1.0, this value gives a normal (undamped) trapezoidal method, a value of 0.0 gives a (damped) Newton method. The damping is normally set to a value minor than 1.0 to avoid numerical oscillations.