Skip to content

toohidsharifi/Machine-Learning-Course-for-Engineers

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 

Repository files navigation

🤖 Machine Learning for Engineers: Electrical Machines Design and Modeling

A professional development course on building high-speed 'digital twins' to replace time-consuming FEA simulations and accelerate your engineering workflow.

MATLAB Ansys


▶️ About The Course

Do computationally expensive and time-consuming FEA simulations slow down your design cycle? This course provides the solution. We will teach you how to build and deploy powerful machine learning 'digital twins' that capture the complex physics of electrical machines and deliver performance predictions in seconds, not hours.

By mastering this skill, you will:

  • Gain Deeper Insights: Rapidly analyze thousands of design variations to develop a more intuitive and agile workflow.
  • Boost Innovation: Replace simulation wait times with instantaneous feedback, allowing you to focus on engineering excellence.
  • Unlock the Future: Build the foundational modeling capability required for automated, multi-objective design optimization.

✅ What You Will Learn

Upon successful completion of this course, you will be able to:

  • ☑️ Formulate raw engineering data into structured datasets suitable for machine learning.
  • ☑️ Analyze datasets to identify and handle outliers, noise, and other data quality issues.
  • ☑️ Apply data preprocessing techniques like normalization and feature scaling.
  • ☑️ Implement feature selection strategies to identify the most influential design parameters.
  • ☑️ Develop and train artificial neural networks (ANNs) to predict the performance of electrical machines.
  • ☑️ Validate the accuracy and generalization of a model using cross-validation strategies.
  • ☑️ Deploy a finalized model as a high-speed 'digital twin' for parametric studies.
  • ☑️ Articulate the advantages and limitations of data-driven models versus traditional methods.

🗓️ Course Schedule (10 Sessions)

This course is built around a hands-on, mentorship-driven approach. You will work with a Synchronous Reluctance Motor (SynRM) dataset for practical exercises and a Permanent Magnet Synchronous Motor (PMSM) dataset for a comprehensive capstone project.

Session 1: Introduction to Machine Learning Modeling & MATLAB Setup

Goal: Understand the "why" behind ML-driven engineering. Set up the MATLAB environment and introduce the course workflow.

Session 2: Dataset Creation from FEA Software

Goal: Learn data sampling methods (e.g., Latin Hypercube) to design an effective parameter sweep in FEA software and create a structured dataset.

Session 3: The Engineering Dataset: Loading and Visualization

Goal: Import the raw dataset into MATLAB and perform exploratory data analysis and visualization.

Session 4: Data Preprocessing in Practice

Goal: Apply essential data cleaning techniques, handle outliers, and implement normalization and scaling.

Session 5: Practical Feature Selection

Goal: Understand and implement methods to identify the most influential design parameters, reducing model complexity.

Session 6: Neural Networks Part 1: Multi-Layer Perceptron (MLP)

Goal: Learn the theory and practical application of MLPs for regression tasks and build an MLP model in MATLAB.

Session 7: Neural Networks Part 2: Radial Basis Function (RBF) Networks

Goal: Explore RBF networks as a powerful alternative, understand their strengths, and build a model.

Session 8: Training and Validation: The Iterative Loop

Goal: Master the process of training a model and using a validation set to monitor for overfitting.

Session 9: Model Evaluation and Cross-Validation

Goal: Use robust methods and key performance metrics (RMSE, R-squared) to evaluate the final model's accuracy and generalization.

Session 10: Deploying Your Digital Twin & Course Conclusion

Goal: Use the final, validated model as a high-speed "digital twin" for analysis. Review course concepts and discuss next steps.

🛠️ Prerequisites & Requirements

This course is designed for engineers seeking to leverage data-driven methods. No prior experience in AI is necessary.

  • Knowledge: A basic understanding of electrical machine design concepts and analysis.
  • Skills: Basic programming fundamentals in MATLAB.
  • Software: MATLAB (R2018a or later) and Ansys Electronics Desktop (version 2021 R1 or later).

🎓 Instructor

This course is led by Tohid Sharifi from the ComProgExpert R&D Team.

✉️ Enrollment & Course Details

  • Start Date: Decided based on requests
  • Format: 10 online sessions, 1.5 hours per session
  • Platform: Google Meet
  • Total Fee: 1200 €
    • An initial payment of 600€ is required to confirm enrollment.
    • The remaining balance of 600€ is due after the completion of Session 5.

To enroll or for more information, please contact us:

About

Learn to replace slow FEA simulations with high-speed machine learning 'digital twins'.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published