Real-time neural inverse optimal control for a linear induction motor

ABSTRACT

A discrete-time neural inverse optimal control is designed for a three-phase linear induction motor (LIM) in order to control its position. This controller is optimal in the sense that it minimises a cost functional. A recurrent high-order neural network, trained with the extended Kalman filter, is employed to obtain a mathematical model for the LIM with uncertainties. A super twisting-based state estimator provides an estimate of the unmeasurable state variables of the system. This control scheme is applied in real time in an LIM prototype which achieves trajectory tracking for a position reference.

KEYWORDS:

• Inverse optimal control
• linear induction motor
• neural identification
• discrete-time nonlinear systems
• real-time implementation

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. DS1104 is a registered trademark of dSpace GmbH.

2. IGBT Power Electronics Teaching System is a registered trademark of SEMIKRON.

3. LabVolt is a registered trademark of Lab-Volt Systems Inc.

4. Matlab is a registered trademark of MathWorks Inc.

5. Simulink is a registered trademark of MathWorks Inc.

6. DSpace R4 is a registered trademark of dSpace GmbH.

7. ControlDesk is a registered trademark of dSpace GmbH.

Additional information

Funding

The authors thank the support of CONACYT-SEP Mexico, through projects CB-256769 and CB-257200. “Proyecto apoyado por el Fondo Sectorial de Investigacion para la educacion”