Usually oscillating transient waves are caused by a combination of capacitance and inductance. In this case, the inductance may be motor loads, or it may just be the 20 miles of wire back to the POCO.
The capacitor stores energy in an electric field, and the inductor stores energy in a magnetic field. When a sudden change occurs (like switching capacitors into a circuit), energy can get traded back an forth between the capacitors and inductors. This happens at the resonant frequency of the circuit. Eventually, the loads and losses in the circuit dampen the oscillations, and the energy trading becomes "steady state". Steady state is basically what the power triangle and Volts-Amperes-Reactive describe.
Steve
It seems like we could just switch the capacitor off without much effect on the rest of the circuit, but its not quite that simple. In the steady state condition, the capacitor would always be charging and discharging. When we switch it off, we are switching off a current flow. If we assume there are inductors in the circuit, they are going to push that current somewhere else - somewhere like into stray capacitances. Since the stray capacitances are much smaller than the capacitor bank, the transient will probably die out much quicker than when you switch a capacitor on. But I am guessing that the voltage peaks could actually be higher when switching a capacitor off. 