Semantic validation of physical models using role models

Abstract

The complexity of models for the simulation of physical systems is steadily increasing. This makes the effective validation of models for different design aspects crucial. One of the many important aspects is the structural correctness and the behavior due to design parameters which are of particular concern for the modeling of wind turbines. This article presents a design and implementation of a role-based validation framework. The framework allows for the creation of validation rules for different design aspects. This is done by role models that are used to define restrictions for an aspect by roles and rules. Multiple role models can be combined to cover all design features during model development. Restrictions on how models can interact with each other can be defined, which broadens language-specific restriction capabilities. The resulting rules can then be tested on arbitrary models based on the Eclipse Modeling Framework, for which mapping between elements of the role model and elements of the validated modeling language must be provided. In the domain of wind turbines, this approach is evaluated by application to two kinds of modeling languages (Modelica and UML2). Role models and rules have shown to be easily described with the frameworks role model language and role model definitions are successfully re-used by the definition of mappings for both kinds of modeling languages.