New technologies in the field of building and damage detection lead to an automation of inspection processes and thus an increase in efficiency. However, an adequate digitalisation of the recorded building data into a BIM model is currently not possible without problems. One main reason for this is the lack of specifications for a digital model that can represent recorded damages. Thereby, a primary problem are uncertainties and fuzzy data in the information modelling, which usually does not occur when applying BIM for new buildings. Fuzzy information, such as the classification of detected damages or the assumption of further hidden damages, is currently evaluated manually by experts, which often requires a complex evaluation of contextual information in a multitude of distributed building documents. An automated evaluation of detected damages based on the building context is applied or implemented in practice. In this thesis a concept for the representation of structural damages in a digital, generically structured damage model is presented. The developed concept offers solutions for problems of current damage modelling, e.g. the management of heterogeneous documentation data, versioning of damage objects or processing of the damage geometry. The modular scheme of the damage model consists of a generic core component, which allows a general description of damages, independent of specifying factors, such as the type of construction or building material concerned. For the definition of domain-specific information, the core component can be supplemented by corresponding extension schemes. As a preferred serialisation option, the damage model is implemented in a knowledge-based ontology. This allows an automated evaluation of the modelled damage and context information using digitised knowledge. For the evaluation of fuzzy damage information, a knowledge-based evaluation procedure is presented. The developed damage evaluation system allows a classification of detected damages as well as the conclusion of implicit evaluation information relevant for further maintenance planning. In addition, the method allows the assumption of undetected damages that can potentially occur inside the structure or in places that are difficult to reach. In the ontological assessment, not only damage characteristics are considered, but also information regarding the building context, such as the affected component or material type as well as existing environmental conditions. To illustrate the developed specifications and methods, the whole concept is applied to two test scenarios.