Dataset for optimized design parameters of three-phase induction motors with validation through machine learning.

Area of Science:

• Electrical Engineering
• Computational Electromagnetics
• Data Science

Background:

• Three-phase induction motors are crucial in industrial applications due to their efficiency and reliability.
• Optimized design parameters are essential for enhancing motor performance and longevity.
• A comprehensive dataset is needed for advanced motor design and analysis.

Purpose of the Study:

• To create a curated dataset of optimized design parameters for three-phase induction motors.
• To expand the dataset to 6000 instances, covering output ratings from 0.5 kW to 100 kW.
• To validate the dataset's reliability and utility for practical applications.

Main Methods:

• A Python-based computational framework was used to compute motor parameters based on standard electromechanical design equations.
• The dataset was expanded from 200 to 6000 design instances.
• Descriptive statistics and tree-based regressors (Decision Tree, Random Forest, Extra Trees) were employed for validation.

Main Results:

• The Extra Trees model demonstrated high accuracy, with R² ≳ 0.9996 and low errors (e.g., MAE ≈ 7.31 W for losses, MAE ≈ 0.0073% for efficiency).
• Residuals were concentrated near zero, confirming the dataset's internal consistency.
• The dataset is suitable for simulation, preliminary design, controller tuning, and predictive maintenance.

Conclusions:

• The physics-driven dataset provides a reliable resource for three-phase induction motor design and analysis.
• The dataset's accuracy supports its use in various engineering applications.
• Future work will incorporate nonlinear magnetic effects, thermal constraints, and experimental validation.