I don't see the range of suction pressures as being a major problem since in any case, the condensate will be a bubble point lqiuid but the range of flowrates (3 to 17 t/hr) is going to require some sort of recirculation system back to the condensate receiver and pump suction.
If your flow and pressures can all vary independently, then you'll need to design the pump to handle from -0.3 barg suction to 4.5 barg discharge at 17 t/hr flow (plus any design allowances you might need). The control system and recirculation system would have to then be designed for the other possible high suction pressure, low discharge pressure, low flow operating conditions.
As Quark suggested, you might also want to look at a self powered condensate pump. I've put one of these in and it worked just as it was supposed to. It's bound to be a lot cheaper than a pump system, especially for these flows. You need to have a source of pressure to push the condensate out during the discharge cycle higher than the 4.5 barg condensate pressure.
Is the range of suction pressures due to the load or cleaniness of an upstream heat exchanger?