Robust Control Design Procedure Based on Particle Swarm Optimization and Kharitonov’s Theorem with an Application for PMSMs
This paper proposes an automatic procedure for robust control design applicable to power converters based on particle swarm optimization and Kharitonov's Theorem. The main benefit is to provide control gains that have a theoretical certificate of robust stability and also accomplish multiple pe...
| Autores: | , , , , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2020 |
| País: | Brasil |
| Recursos: | Associação Brasileira de Eletrônica de Potência (SOBRAEP) |
| Repositorio: | Eletrônica de Potência (Online) |
| Idioma: | inglés |
| OAI Identifier: | oai:ojs2.journal.sobraep.org.br:article/271 |
| Acesso em linha: | https://journal.sobraep.org.br/index.php/rep/article/view/271 |
| Access Level: | acceso abierto |
| Palavra-chave: | Kharitonov’s theorem Particle swarm optimization Permanent magnet synchronous motors Power converters Robust control |
| Resumo: | This paper proposes an automatic procedure for robust control design applicable to power converters based on particle swarm optimization and Kharitonov's Theorem. The main benefit is to provide control gains that have a theoretical certificate of robust stability and also accomplish multiple performance criteria in a design less dependent of human-machine interaction. Regarding the particle swarm optimization, each particle represents a controller candidate whose performance is evaluated by means of an objective function, using the vertices of a polytopic model of the plant and the four polynomials of Kharitonov's Theorem. The effectiveness of the proposed procedure is illustrated by means of a case study that considers the speed control of a permanent magnet synchronous motor subject to uncertain mechanical and electrical parameters. The designed controllers, obtained in an off-line way, yield good trade-offs between performance and robustness, as confirmed by simulation and experimental evaluations. Analyses show superior results with the proposed strategy compared to a genetic algorithm and to a design tool specialized for PID tuning, indicating its viability as an alternative for robust control design in power electronics. |
|---|