Logical ontologies (for example, in OWL) are used for several tasks ranging from structured databases to artificial intelligence. These tasks usually involve a form of reasoning. Due to the logical nature of the ontologies, errors in them can carry relevant consequences to their usage. These faults can be due to a human error by an ontology engineer, or because the ontology was generated by an error-prone procedure such as ontology merging or natural language processing, among other reasons. In this PhD project, we propose to explore a framework for detecting, explaining and repairing faults in an existing ontology. To do so, we build on existing work by myself in meta-ontology fault detecion, work in reformation and in belief revision. We aim to combine these three approaches and to complete, extend and improve them to make the resulting combination a practical and well-founded solution.