CompBat will focus on developing tools for discovery of new prospective candidates for next generation flow batteries, based on machine learning assisted high-throughput screening. Density functional theory calculations will be used to obtain data on solubilities and redox potentials of different molecules, and machine learning methods are used to develop high-throughput screening tools based on the obtained data. The results of the high-throughput screening are validated with experimental results. Target molecules will be bio-inspired organic compounds, as well as derivatives of the redox active specialty chemical already manufactured in bulk quantities. Stability and reversibility of the molecules will also be investigated by DFT calculation, experimental investigations and machine learning methods, for a selected group of interesting molecules. Numerical modelling of flow battery systems will be performed with finite element method, and with more general zero-dimensional models based on mass-transfer coefficients. The models will be verified experimentally, and the modelling will generate a data-set to allow prediction of the flow battery cell performance based on properties of the prospective candidates obtained from high-throughput screening. This data is used then to predict the flow battery system performance from the stack level modelling. Freely available cost estimation tools are then adapted to estimate the system performance also in terms of cost. This approach will allow prediction of the battery performance from molecular structure to cost. Furthermore, the concept of using solid boosters to enhance the battery capacity will be investigated by developing models to simulate the performance of such a systems, and validating the models experimentally with the candidates already reported in the literature.