Model Parametrization-Based Genetic Algorithms Using Velocity Signal and Steady State of the Dynamic Response of a...

Area of Science:

• * Electrical Engineering and Control Systems
• * Computational Intelligence and Optimization

Background:

• * Dynamic models and parameterization are crucial for motors in engineering and industrial applications.
• * Traditional methods often require multiple motor signals (e.g., torque, current, speed), necessitating complex instrumentation.
• * Existing techniques face limitations in parameter estimation accuracy and measurement requirements.

Purpose of the Study:

• * To propose a novel method for parameterizing a motor's dynamic model using solely the speed signal and steady-state values.
• * To develop an approach that reduces the need for extensive sensor instrumentation, such as full current or torque measurement.
• * To reconstruct the motor's mechanical and electrical equations through evolutionary computation.

Main Methods:

• * Utilized evolutionary computation for reconstructing the motor's dynamic equations.
• * Employed the transfer function of the motor's speed signal as the primary input for analysis.
• * Validated the model's accuracy using Root Mean Square Error (RMSE) for speed and current signals.

Main Results:

• * Successfully reconstructed the motor's dynamic equations with an RMSE below 1% for both speed and current signals.
• * Demonstrated the feasibility of parameter estimation using only a single measured variable (speed).
• * Achieved accurate model parameterization without direct measurement of current or torque signals.

Conclusions:

• * The proposed method offers a significant advancement in motor dynamic model parameterization by simplifying measurement requirements.
• * This innovative approach enables accurate system identification using minimal sensor data, applicable to various motor control scenarios.
• * The technique provides a cost-effective and efficient alternative for motor parameter estimation in research and industry.