WP5 leader. Mainly contributing to WP1-WP3 and WP6-WP8.
NTUA will lead WP5, being involved in all WP5 tasks and will lead most of its tasks. NTUA will use expertise on both the development of control algorithms and systems and in data-driven mathematical modelling, to create, test and validate nanoQSAR and PBPK models and expose them as web services; design experiments for collecting rich dynamic data from the NM production platforms; generate adaptive models for the on-line prediction of toxicity scores and NM characteristics as functions of manipulated process variables; design novel Model Predictive Control algorithms considering the special requirements of the NM production processes; develop a simulation platform for testing, optimising and tuning the MPC algorithms before applying them on the real production platforms. NTUA will also be involved in WPs 1,2,3 and particularly in the integration of the control algorithms in the NM production systems at all TRL levels 4 to 6: NTUA will work specifically on software development, data collection and integration with screening technologies and infrastructure and with system actuators for feedback control implementation. NTUA will participate in data modelling tasks in WP4 and on the integration of the SABYDOMA database with the on-line production and control system, based on the expertise it has developed through the involvement in the eNanoMapper project. Finally, NTUA will also take part in WP7 – Dissemination and exploitation.
The NTUA group has proven track records relevant to the SABYDOMA project. Prof. Sarimveis leads the Unit of Process Control and Informatics in the School of Chemical Engineering and as the name of the Unit suggests, it combines all the modelling expertise and background required to lead WP5 and support the development of the online control technology through TRLs from 4 to 6 in WPs 1-3. The Unit has more than 20 years-experience in developing control algorithms with emphasis on the MPC theory and technology. MPC algorithms have been developed for various types of systems including linear stochastic, fractional and impulsive dynamics. Industrial process control is the main consumer of the algorithms, but applications in many other areas and disciplines have been published including energy systems, production and inventory control and optimal drug administration. MPC is closely connected to data-driven modelling and here the group has significant experience as it has developed numerous machine learning algorithm with emphasis on artificial neural networks, fuzzy modelling and evolutionary algorithms. It is important to emphasise that the experience of the group includes not only static, but also dynamic machine learning modelling, which is essential in MPC. The group has long time experience in nanoQSAR and PBPK modelling and has developed the Jaqpot platform (part of the European Open Science Cloud, EOSC) for creating, hosting, validating and sharing these types of models. Optimal experimental design is another area where NTUA can provide expertise, important for the SABYDOMA project, because these experiments will generate rich datasets, which will subsequently be used for creating the predictive models that are key components of the MPC algorithms.