In order to improve the performance of a carbon capture process, novel solvents are being developed and uniquely adapted to different applications. The process performance is highly dependent on the choice of the solvent, given by a major share of the energy demand for regeneration. In addition, plant design and operation conditions both affect the process performance, thus energy efficiency and have to be investigated before designing a large-scale plant and optimally operating it. In this contribution, a framework is presented for the systematic characterisation of novel solvents in an industrial pilot-plant using a data driven modelling approach, which takes into account plant characteristics and industrial operation conditions. The framework is designed to determine optimal operation conditions regarding maximum energy efficiency with reduced number of experiments due to the fact experimental absorbent screening is time-consuming and costly. The proposed three-step framework is based on the assumption that for a novel solvent little to no thermodynamic knowledge is available. Therefore, Monoethanolamine (MEA) is taken as a baseline for the general performance of amine-based solvents. In the first step, data from physical simulations with MEA serve for the development of a surrogate model describing the general behaviour of a carbon capture process and is used for further solvent comparison. Followed by pilot-scale experiments under industrial operation conditions with the reference solvent (step two) and the novel solvent (step three), the surrogate model is adapted to experimental data to account for plant characteristics. By means of the surrogate model, optimal operation conditions regarding maximum energy efficiency are derived in 13 experiments for the reference solvent and in 16 experiments for the novel solvent. Finally, the optima allows for a fair solvent ranking. (C) 2019 Published by Elsevier B.V. on behalf of Institution of Chemical Engineers.