Thanks for the response. I think I need to do some homework. I called the VFD manufacture (Dan Foss) and they said it was not unusual to parallel motors as long as they had individual overload protection (they did not specify what type). I think the motors are constant torque and were purchased as one unit with the fans attached. Unfortunately I am only familiar with conventional overloads. Are the individual thermistors internal to the motor or maybe an electronic overload? The motors are not VFD rated.
The VFD can be sized as Variable Torque or Constant Torque. The motor is just a motor. The LOAD is what determines if the application is VT or CT, and the difference is only related to the over current CAPACITY of the VFD (not to be confused with the MOTOR thermal overload protection). Most fans are going to be Variable Torque loads, meaning that the mechanical load on the motor shaft is dropping down as the speed drops, because the fan blades are no longer moving as much air.
The VFD has motor OL protection built-in, but as has been said, this cannot differentiate between the two motors, it only sees the entire (combined) load. Hence the need for separate OL relays.
OL relays monitor the motor current and determine, based on predictive modeling, what amount of current represents a potentially damaging temperaure in the motor. OL relays come in two main flavors;
1) Conventional, a.k.a. thermal or Bi-metal (there are other types of conventional OLs, but they are less common now). These have been around for years and are still the most common type because they are the cheapest and work very well for the most part.
2) Solid State overloads, which use current transducers to provide information to a microprocessor that has the motor damage curves programmed in. As a general rule, SSOLs offer a lot more protection features and a wider range of adjustment so are now superior to Conventional OLs except in one case, that being this one where you are down stream of a VFD going to multiple motors. SSOLs have a difficult time differentiating real (active) current from harmonic current and most ore not rated for use in what are called "non-linear loads" such as VFD outputs.
A 3rd option is one that uses Thermistors, a resistive device that is made to change resistance at a specific threshold of temperature. The thermistor sensor devices MUST be embedded into the motor windings, which for the most part means that the motor mfr must do so when the motor is made. The thermistors are then wired in series to a Thermistor Trip Relay or a Thermistor Trip Input to a controller such as the VFD. Instead of monitoring current then, they monitor the actual physical temperature of the motor.
The problem with Thermistor OL protection however is that as I said, the thermistor devices must be built-in to the motor. In countries where IEC motors are used, this is a very common feature of motors, even if they are never used. In the US and Canada however, it is not and someone would have needed to special order the motors with them inside. Some manufacturers are now including thermistors in their "Inverter Duty" motor offerings, but that is by no means universal. Their inclusion would be noted on the nameplate so if your motor's nameplates don't say they are thermistor protected, they are not. Thermistors can be added in the field by a good motor rewind shop, but it is not practical for the average user to add them.
So bottom line, IF your motors happen to have thermistors in them already (as noted on the nameplate), then use them, they are perfect for what you want to do here. If not (most likely), then use basic every day conventional bi-metal overload relays, one for each motor, with the trip contacts wired in series with each other in a way that trips the VFD. Usually if the VFD has a Thermistor Trip input (the Danfoss will), you can wire the conventional OLs to that circuit. This way if either OL relay trips, the VFD will shut down and the display will read "External OL Trip" or something to that effect, rather than just shutting down without indication.