Modeling the Wind Turbine Benchmark with PWA Hybrid Automata (Experience Report) - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

Modeling the Wind Turbine Benchmark with PWA Hybrid Automata (Experience Report)

Nikolaos Kekatos
  • Fonction : Auteur
Marcelo Forets
  • Fonction : Auteur
Goran Frehse

Résumé

The wind turbine benchmark is part of the ARCH benchmark repository. It entails closed-loop requirements and encompasses nonlinear and hybrid dynamics. Owing its origin to industry based applications, the benchmark modeling is done with MATLAB/Simulink. Formal verification tools, however, do not operate on simulation models but on formal models, such as hybrid automata. Particularly efficient verification algorithms are known for systems with Piecewise Affine (PWA) dynamics. In this vein, we construct a PWA model of the wind turbine in the SX format, which formally describes a network of hybrid automata and can be used by several reachability tools. The model transformation follows a four-step approach with the aim of (i) adapting the Simulink model to obtain a verification model (ii) translating the Simulink blocks to equivalent blocks in SX format, (iii) conducting compositional, syntactic hybridization to obtain a PWA approximation of the dynamics of the nonlinear blocks, and (iv) performing model validation. We also report some preliminary experiments on the subsystems (network components) of the wind turbine that we conducted with SpaceEx.
Fichier principal
Vignette du fichier
main.pdf (1.43 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01508674 , version 1 (14-04-2017)

Identifiants

  • HAL Id : hal-01508674 , version 1

Citer

Nikolaos Kekatos, Marcelo Forets, Goran Frehse. Modeling the Wind Turbine Benchmark with PWA Hybrid Automata (Experience Report). ARCH17. 4th International Workshop on Applied Verification of Continuous and Hybrid Systems., Apr 2017, Pittsburgh, United States. ⟨hal-01508674⟩
247 Consultations
249 Téléchargements

Partager

Gmail Facebook X LinkedIn More