I would chose a servo motor instead of a VFD.
A PLC is not fast enough and I don't know of any general motion controllers other that ours that have built in sine waves with linear and logarithmic sweeps and amplitudes can be changed on-the-fly.
In the test arena we are usually competing with MTS ( Minnesota Test Systems ). The advantage MTS has is that they will sell you a complete system and they know what they are doing. Our solution is usually our RMC75 controller, Lab View, and an integrator that can put all of that together. Our solution is cheaper and more flexible the integrator must be fairly knowledgeable. I am usually in the back ground verifying the design.
If you choose some other solution make sure they have a successful track record because most people that try this on their own screw it up.
Sinusoidal test systems are more difficult that they appear. Most designers grossly under estimate how much power is required or over estimate what they can do with the power they have.
The amplitude and mass or inertia was not specified but I will give an example of what people under estimate most often.
The power at the load is proportional to the frequency cubed. That is a 60 Hz oscillation will require 8 times more power than a 30 Hz oscillation unless the amplitude of the 60 Hz oscillation is reduced by a factor of 8. This calculation doesn't take into account the power lost getting it to the load.
It is good to have the system modeled before picking out equipment.