ABB ACS800 are VFDs, yes. It is also unlikely that there are contactors on VFDs, unless maybe there is only one drive and two motors? That's not clear. Probably not though.
Assuming one drive per pump, no other circuit elements, then it's a real mystery. But you must understand the way drives can work, especially DTC (Direct Torque Control) drives like the ACS800. The output of the VFD can be set for a current limit setting. In order to accomplish that limitation of current in a circuit, the drive must alter the voltage and frequency being applied, regardless of what you are telling it to do. That's the only control it really has, it cannot control the actual current that the load demands unless it essentially changes the nature of the load, which in the case of a pump, means lower the amount of pumping it is doing by lowering the speed of the pump. But if the pump shaft is not turning, then the motor is in a Locked Rotor condition and the current will attempt to go to 600% of the FLC. Drive will stop that at whatever the current limit setting is, but that means driving the speed down, and that also results in driving the RMS voltage down as well. The result: high current, little voltage, also little speed, but you may not be able to see that.
If I had to guess as to why this is happening when it does NOT happen if only one pump is running, is that this is ultimately a mechanical problem. Many types of vertical pumps, including submersibles, have a "thrust bearing" arrangement to absorb the upward thrust created by the spinning motor and the lift that creates. But at the same time, it DEPENDS on the hydrodynamics of that lift in order to get off of the lower bearing, or sometimes there IS NO lower bearing at all, the liquid flowing is the lower bearing. If you have two pumps starting together, and one starts slightly faster than the other ad there is no means of back flow prevention, it's head pressure is forcing the lagging pump downward onto its bottom resting point, it never lifts off of it, never develops the lift into the thrust bearing and the rotor remains locked. Drive #2 then keeps dropping the speed to maintain the current limit, which exacerbates the problem and perpetuates it until the drive trips off line.
Solution? Back flow prevention on each pump output, no current limit, synchronized ramping of both drives together, some combination of the above.
Possibilities other than that, one pump is pumping backward. It can be hard to tell because there is flow either way, just more flow in one direction. But if that were the case, the problem would only happen in one direction, meaning A then B, not not B then A, at least I think so, it depends on a lot of other things.
If it IS just one VFD and two pumps and you are using contactors to select one or two, then it may be that one contactor is closing sooner, the direction is wrong, any nomber of other issues could be going on there. If that is the case, clear that up before we waste any more time on this.
