due to fact that each motor has its own invert er, one method is sending command of e.g. 100rpm to the first electromotor and 90rpm to the second electromotor. in this way the connecting shaft will rotate at 100rpm but the torque would be applied on the shaft. Am I right?

What is "Mechateronics"? ;-)

a combination of electrical, computer and mechanical engineering

Get a differential gear system. Just exactly what you have at the back of you axle driven car or truck. Fix the two rigid support on a rigid frame. Couple the two motors on a coupler with steel pulley or friction drive solid tire like you have on a tricycle. Make your three phase connection of the two motor to run in the same direction( Synchronization)

Use Anti-slip belt to tap the power on the pulley and antivbration Damper coupler

Dear Adel,

some of the ansewrs given scare me to death :-)

Seriously - if you have a proper control in place like field orented control, synchronizing the two machines using the inverters is not a big delt. Just make sure, not to send a speed set-point but a torque and both the machines will povide the same torque to the common shaft.

Basic synchronisation, exact match of resolver outputs and in best case a well-done alignment of d-q-positions of the machines rotors prior to fixing the mechanical coupling will help to do the trick.

In the lab, if two identic machines/resolvers/converters are in use, this is done before coffee break :-)

The other alternative is to get a thru shaft on one of the motors. In this case the motor will be able to be mechanically coupled to the other motor on the one end and mechanically coupled to a load on the other end.

Dear Mrtin, due to fact that both electromotors are AC motors (and not DC) and controlled by invert er, we can use just frequency control which has direct relation by speed, so how is it possible to use torque set-point?In other word, we want to apply different speed set point and in this way both machines shaft will go under torque, we know that in this way, both electromotors will experience high current (and thus torque), but this is not important because our test takes just 200 seconds 

Dear Adel,

I expected the drives to be AC anyway.

If you change your control scheme from the very basic f/U to a more advanced Field Oriented Control (FOC) that is very common in controlling AC machines, you get independent access to both, speed and torque.

Actually this control is so basic, that most inverters you can buy will allow to change to this scheme by means of configuration.

From your description I take it, you're not going to drive a common load with two machines, but consider one machine to be under test and the other one to act as a load, right?

Dear Adel,

If you are using frequency controlled drives without any speed feedback then it is best to provide a common frequency signal to both drives. One will be the drive motor and the other the load motor. Both will act as motors when running up to any desired speed. If you have any imbalance in the settings you may adjust the load drive reference until one it motors neither motors or regenerates which will indicate true synchronism.

Once synchronised when you wish to load the drive motor simply lower the reference frequency of the load motor so it begins to regenerate in which case it will load the drive motor as it will develop a negative electrical / shaft torque. Under frequency control the load torque will depend upon the torque slip curve of the motor and you will have to design the reference frequency signal of the load motor accordingly. Generally fairly linear though. Also as the drive motor slip increases it speed will drop so its best to have the drive motor under closed loop speed control

In this case it is best to also couple the dc links of each inverter so that the energy regenerated from the load motor is returned and under load test only system losses will be drawn from the mains supply. It will also avoid requiring a mains regeneration unit. Thus a 300kW system will draw maybe 30 kW only! Yes the drive motor will provide say 300A and the load motor will regenerate say 295A and the supply will draw on the difference. This is the concept of the common DC bus.

If your drive system allows FOC then as Martin notes run the drive motor under speed control and the load motor under current control (or torque control). As the shafts are coupled only a single resolver or speed sensor is required. Here you would set Id = Imag and Iq of the load motor to zero and the drive motor would run both up to whatever speed you desire at no load. When you wish to load the motor simply set Iq for the load motor to whatever level you require (-Iq) and it will load the drive motor and regenerate all energy back into the dc link. In this way the load motor may apply any type of load characteristic to the drive motor.

I have designed several test beds like this, just like Martin notes, and they work great although not before coffee time! :)

Dear Martin and Diana, thank you very much for your answers. you are right! one motor is driver (starter) and the other one is driven (load). I will apply your methods and let you know the results

You can use hydraulic coupling as in convertor of automatic transmission