Some vibration problems can be described as resonances; there is an exciting phenomenon with given frequency (e.g. a vortex in the draft tube) and a system with given natural frequency (e.g. the generator pole wheel). If the frequencies match (and the vibration mode shape permits) then there may be high amplitudes (in our example of the active power swing). This kind of trouble is theoretically predictable, if the frequencies can be determined.
Many practical problems, however, do not adhere to this ‘textbook’ concept. Some vibrations are self-excited, thus they will per definition occur at a natural frequency (because the damping of the pertinent mode drops to zero). Other vibrations are already too strong in absence of resonance. For another type, there will resonance in any case (for instance, because the penstock shell has an abundant number of natural frequencies). In all those cases the concept of resonance is not really helpful.