I have been analyzing water hammer using force time histories. With a cutoff frequency of several hundred Hz I have seen time history captured modal mass greater than 100%. When the cutoff frequency is not specified and the system is analyzed up to 999 modes, one transient extracts 26 modes and gives me Captured modal mass (%) : 104.19. The Frequencies sub-report does not show a captured mass total over 100% even when the Time History analysis summary does have a value above 100%. This doesn't make physical sense to me - why is the Time History analysis capturing more mass than exists in the model? For what it's worth, all supports in my model have rigid stiffness.