Thanks to the advances in information technology, modern industrial systems are becoming increasingly intelligent and autonomous; thus their requirements for, e.g., correctness, availability, traceability and reliability, are also increasing. Monitoring, analysis and diagnosis of such industrial systems became pivotal and fueled the development of virtualization and simulation solutions such as digital twins. In a nutshell, digital twins are virtual representations of actual systems or processes that serve as real-time digital counterparts for, e.g., prediction, analysis, testing, and simulation. Developing digital twins is a complex process. On the one hand, it includes developing digital twins at different levels of abstraction of the system to allow one to focus on different relevant aspects (e.g, behavioural, logical, physical). On the other hand, it must ensure the correctness of digital twins with respect to the system specifications and the respective level of abstraction, and the federation enabling the communication between digital twins conceived as exchange and resume of models. This project aims to develop a model-based framework addressing the above-mentioned challenges by i) automating the creation of digital twins for the simulation, monitoring and testing of functional and non-functional properties ii) continuous validating digital twins to meet the required properties and iii) developing a multidomain and automated digital twin toolchain for the verification and validation of complex industrial systems based on digital twins. We foresee that this project will positively impact the efficiency of such systems by reducing their time to value and by increasing their final quality.