- Simscape Onramp: https://matlabacademy.mathworks.com/details/simscape-onramp/simscape
- Getting started on gears: https://www.mathworks.com/help/sdl/gears.html
- Modelling gears with losses: https://www.mathworks.com/help/sdl/ug/model-gears-with-losses.html
- Example for simple gear: https://www.mathworks.com/help/sdl/ug/simple-gear.html
- Powertrain Blockset: https://www.mathworks.com/help/autoblks/ug/getting-started-with-powertrain-blockset.html
Help with generating dataset for machine learning
2 views (last 30 days)
Show older comments
I want to make a real life gear train to generate a dataset for machine learning. I want a real world scenario with included losses. How should I go about it, I'm a beginner in simulink. Specific instructions will be really helpful
0 Comments
Answers (1)
Jaynik
on 8 Mar 2024
Hi,
Since you are a beginner to Simulink, to make a gear train model, you can start with the "Simscape Onramp". It will help in understanding how to create mechanical models in Simulink.
Once you understand the basics, to create a model from scratch, you can start by creating a basic gear train model. "Simscape > Driveline" offers gear blocks. You can start by creating a model using multple "Simple Gear" blocks and chaining them together. You will also need "Mechanical Rotational Reference" block under "Simscape > Foundation Library > Mechanical > Rotational Elements" as it serves as a ground for mechanical rotational systems. You can then add the "Rotational Friction" block found in "Simscape > Foundation Library > Mechanical > Rotational Elements". These blocks can be placed in series with each gear to represent the frictional losses at that gear stage.
To drive your gear train, initially you can add a "Ideal Torque Source" block from "Simscape > Foundation Library > Mechanical > Mechanical Sources". Later, to simulate real-world operational noise, you can add a "Torque Noise Source". To collect data, use "Ideal Rotational Motion Sensor" and "Ideal Torque Sensor" blocks. Place them as per the data requirement. Once you have created the model, set the Configuration Parameters and select the appropriate solver. A variable step solver like "ode45" is often suitable for mechanical systems.
To log the data, you can right-click on a signal line and select Log Selected Signals. The logged data will be available in the MATLAB workspace. Depending on your machine learning application, you may need to preprocess the data.
If you want to create complex models and have a fair idea of modelling automotive vehicle, you can also use the "Powertrain Blockset" as it can simplify the process of modeling complex powertrain systems.
You can refer the following documentation links to get started:
I hope this helps!
0 Comments
See Also
Categories
Find more on Powertrain Blockset in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!