testing for DC motor inertia

"I have an MG set (DC motor coupled with the AC generator), and I want to measure the inertia of the shaft."

If it is the total inertia of all of the shaft and attached rotating parts you might be able to measure the power input and compare to the running power input to arrive at a reasonable figure for the overall inertia.

May we ask why this inertia is of interest?

My project is to create a speed controller for the DC motor in both hardware and simulation. First, I want to use a PID controller for this purpose. However, this is a really old machine, thus, I could not find its spec anywhere. And I kind of need its data such as torque constant, inertia... for the simulation. The problem I have right now is the AC gen shaft and DC motor shaft are coupled, and I do not many expensive equipment to measure. I am not sure if there is any easy way to measure the inertia of this DC motor

Thank you for your help

Do you have any nameplate data?

Frame size, volts, amps, etc.

For stability calculations for a PI or PID controler you usually the per normal mechanical time constant of the entire system you will be driving.

If you have the motor HP, amps, rated voltage etc then you can use the DC drive, set the current to say 100%, and measure the time it takes to accelerate from 0 to 100% speed. Your mechanical time constant is this time (inertia if you take the rated torque and re-calculate). If you you use 50% current and 50% speed it is also the accel time. If you shoose 50% current and 100% speed then it is t/2 for your time constant "J".

i.e. The inertia term in the PI control will be (1/Js) where J is in seconds per normal.

I'm pretty sure, as several others have said , that what you really need is the inertia of the rotating system, not just the motor.

Assuming that there is at least a small amount of shaft accessible between the motor and the generator, wrap several turns in a single layer of very flexible cord (fishing line?) around the shaft. Tie a series of weights on the cord, and measure the time required for the weight to fall a known distance. The smallest weight that will just keep it moving tells you your low-speed (negligible air) friction force/torque. The lever arm is half the shaft diameter plus half the cord diameter. Force (accelerating weight - friction weight) multiplied by lever arm is torque. High school physics should do the rest.

Hi,

you can measure runup time (velocity versus time and then plot and deduct the time constant: intersection of initial slope with infinitely reached velocity).

Then add some considerable amount of known inertia and measure once more.

Then calculate.

Alternative: first measure the motor constant: torque/current.

Then measure the slope of the runup curve: acceleration.

Inertia is the ratio of torque/acceleration.

Compensate for friction loss by measurement of current at "no load" condition after acceleration is completed. The current draw in this steady velocity condition is to be converted to friction torque. This will give too high a value as any air-friction is higher at high speed than at runup.

Good luck and success

RHABE

Thank you for all of your help

I thought about that way too, however, as I said before, this was a old MG set, I could not find any of its data (online + factory). I also just have 1 tacometer was attached to the AC gen and the shafts were coupled. If I use the ratio of torque/acceleration, the inertia of the AC gen shaft will be include in my calculation which I do not want to. This problem really gives me a headache

There is an old rule of thumb, take the measurements you have, estimate what you don't have, apply a little math, and the buest guess is probably in the correct order of magnitude.

i.e.: give us the motor / generator dimensions, diameter and length of rotors, diameter of shafts, conductor sizes, estimated hp, size of protection devices, rating of DC drive if there is one, ALL the data you can provide. I find it difficult to believe none of the pieces have any data, generator rating, volts etc.

If you have none of this data and cannot make any reasonable guesses, give up, it won't work.

One way to check the overall inertia of the rotating mechanism is to measure the elapse time between stopping the power supply and the complete stop of the mechanism. It could be measured (and recorded) in minutes. More this time is longer, more the inertia is high.

Not very accurate. You don't know the friction. A coast to rest is simply a function of inertia and friction losses. The higher the efficiency of the bearings the longer it takes to coast to rest. You want the per-normal time constant with the rated torque.