it depends what kind of vibration mode You want to use (a thickness resonance or a radial one). You can easy have an equivalent circuit measuring U an I along frequency.
As for the voltage and current used the main idea is not to let the transducer to exceed Curie temperature bacause of depolarization. If You work in a pulse mode You can apply short bursts or short pulse (Dirac delta) of hundreds of volts and expect a current in a range of a few ampers.
I don`t think that the transducer will exceed Curie temperatur in a system for flow measurement regarding transmit voltage (but in high-power ultrasound), but of course you have to take care with higher temperatures e.g. > 150°C of your flow medium. This can lead to depolarisation for your PZT.
I would start by hooking an everyday signal generator to the source and a wideband low-level amplifier to the receiver. Experiment with signals in the 1-10V range. Try tone bursts of a few cycles of a sine wave, near the transducer's resonance if possible, but otherwise below it. If the transducers are well-coupled to the fluid it might not take much source voltage to produce a useful signal.
Note that you may see some direct electrical coupling, depending on how well shielded the system is. Direct coupling shows up as a received signal that occurs essentially at the same time as the source pulse. You can usually just ignore it, but sometimes you have to improve the electrical shielding of the source and receiver cables.