Fault Detection for Electric Motors Using Vibration Analysis (Project 253)

[robertogl]

Contribute to the discussion by asking and/or answering questions, commenting, or sharing your ideas for solutions to project 253

[sar-che]

Hello!
This is Sara Chehaidi an electrical engineer who graduated last year from the Lebanese university.
I am willing to work on this project to further develop my skills in the electrical machinery and fault detection field.
Will be extremely happy to collaborate with you all!

[K-Yogesh-Mishra]

Hi everyone,
I’m Yogesh Kumar Mishra, a final-year Electronics and Communication Engineering student at Vasavi College of Engineering. I’m excited to collaborate with you all on this project !
Looking forward to working with you again, @robertogl, Sir.

[sar-che]

hello Yogesh,
My team and I are ready to collaborate!
If you would like to connect my email is: [email protected]

[sar-che]

hello all,
we were trying to simulate the effect of the shaft misalignment using the flexible shaft block as suggested in the project's description.
But the problem arises when checking for the angular velocity after placing the block and all concerned parameters, no effect at all is seen by the introduction of the flexible shaft block in the system for the "model bending flexibility" option even when trying different values for the parameters that might clearly affect the bending, as the external forces applied and the bending rigidity and many more. So is there any advice?
note that for the "model torsional flexibility" clear effects were shown on the angular velocity.
thank you in advance,
best,
Sara Chehaidi

[robertogl]

Hi @sar-che,
It would be easier if I could take a look at your model. However, I would check if the shaft block is connected correctly and if you are using a load. If you have not already, see this example.

[sar-che]

Thank you for your response,
we tried adding a load (rotational damper), but there are still no results.
sure, the screenshot below is for the flexible shaft connection where b input is taking a torque. for the block we have checked for the "model bending flexibility" (the example given shows the torsional modeling case) and trying multiple values for the parameters of the block. The modelling method used for the shaft is the lumped mass approach with a minimum number of segments equal to 2 (short shaft with no variation in the properties along its length). For the inputs Fr, W, M , V no values were entered. The value of the rotational damper's damping coefficient was also changed.

the output for the different values of the parameters is not changing and can be seen in the screenshot below (same output as if there is no flexible shaft block)

[sar-che]

Hello Mr. @robertogl,
Hope you're doing well!
Any advice from your side on the above issue is highly appreciated.

[jkluger]

Hi @sar-che ,

the Flexible shaft block models how torsional motion influence transverse motion, but it doesn't automatically model the reverse. To make the transverse motion affect the rotational dynamics in the line, use the PS signals output by the Flexible Shaft block (Fr, M, V, W. these are the force, moment, translational velocity and rotational velocity at each of the bearings that you specify in the block). You can use other PS blocks to do calculations using these signals (or just use these signals directly), to feed the PS input signals of other rotational blocks in the system. For example, you can make the Bearing block have an input signal N for load, and varying N will vary the rotational friction in the system.

In the screenshot that you shared, please be aware that a Rotational Motion Sensor acts like a free end (it doesn't transmit any torque). Think of it as a sensor that you place in parallel to 2 points ofyour system to meausre the relative velocities of those 2 points., so the C port of the damper block is not physically attached to anything. You can attach a Rotational Reference block to the damper C port. Then attach the Rotational Motion Sensor R port to the damper R port. Keep the Motion sensor C port connected to the Rotational Reference block.

Best,
Jocie
Simscape Developer

[sar-che]

Hello @jkluger,
First of all, we greatly appreciate your response. Regarding the proposed solution, we will proceed with adjusting the system based on your suggestions. We will keep you updated on the results and would appreciate the opportunity to reach out for further guidance if you are available.

Best regards,

[Sattar-mazouzi]

Hello, I'm Abdesattar I'm a PhD student in embedded electronics. My team and I are working on this project, I'm happy and excited to share information and progress and collaborate with you, thank you all.

[sar-che]

Hello Abedesattar, Great to hear from you. We are ready for collaboration as well!

…

On Mon, 3 Mar 2025, 02:10 Abdesattar Mazouzi, ***@***.***> wrote: Hello, I'm Abdesattar I'm a PhD student in embedded electronics. My team and I are working on this project, I'm happy and excited to share information and progress and collaborate with you, thank you all. — Reply to this email directly, view it on GitHub <#120 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BNKKHCTI6RXTLDW5KA5EGRL2SOMYJAVCNFSM6AAAAABSO2REQKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEMZWHEYTQMY> . You are receiving this because you were mentioned.Message ID: <mathworks/MATLAB-Simulink-Challenge-Project-Hub/repo-discussions/120/comments/12369183 @github.com>

[Sattar-mazouzi]

Hello Sara, I'm glad to hear that and happy to collaborate with you!

[Sattar-mazouzi]

Hi Community

@robertogl, @K-Yogesh-Mishra , @sar-che @jkluger

I'm concerned about using the accelerometer block to capture the motor's variation. The accelerometer block provides an output voltage that is proportional to the acceleration rate presented at the mechanical translational physical port R. We've used the Wheel and Axle to converter of mechanical rotational into mechanical translational motion to the input of the accelerometer as seen in the figure below:

• Is this an effective approach to capturing the vibration?
• If so, this vibration is captured along one axe supposed (X), any suggestion of how to measure vibration along three axes (X, Y, Z)?

Thank you

[jkluger]

Hi @Sattar-mazouzi,

The set-up in that image shows the acceleromter outputting a voltage proportional to the rotational acceleration of the shaft, because the Wheel and Axle block essentially just converts the rotation to a directly proportional translation, as shown in the equation on the documentation page. To measure the rotational acceleration without a Wheel and Axle block, you could use a Gyro block instead of the accelerometer block.

That image shows a rotational model of the shaft, but it looks like it's missing the off-axis (transverse) model of the shaft. The model needs some blocks that relate the shaft rotation to the shaft off-axis motion. You can add a Flexible Shaft block, which models one-way influence of rotation on excitation of the transverse degrees of freedom. The transverse motion is usually a more complex function of the rotational motion, rather than directly proportional to it, and you ened to model things like the effective transverse mass, stiffness, and damping. The Flexible Shaft block handles this, but you could try alternative custom solutions (like create your own Translational network with a Force Source block that has an input signal that is a function of the shaft rotation).

Once you have a model for the transverse dynamics, you can add the acceleration sensors. If you use the Flexible Shaft block, that block already outputs transverse velocity as Physical Signals at select locations (usually bearing locations, but you can add any 'free' location). You could differentiate that velocity to get acceleration. For a clean measurement, you might need to do some filtering or use a Spectrum Analyzer block. If you create your own translational network, you can use the Accelerometer block or an Ideal Translational Motion Sensor block (if the system has many vibrations, you may still need to do some filtering if using this block to get a clean acceleration measurement).

Hope this helps.

Best,
Jocie