Membrane contactors have been proposed for decades as a way to achieve intensified mass transfer processes. Post-combustion CO2capture by absorption into a chemical solvent is one of the currently most intensively investigated topics in this area. Numerous studies have already been reported, unfortunately almost systematically on small, laboratory scale, modules. Given the level of flue gas flow rates which have to be treated for carbon capture applications, a consistent scale-up methodology is obviously needed for a rigorous engineering design. In this study, the possibilities and limitations of scale-up strategies for membrane contactors have been explored and will be discussed. Experiments (CO2absorption from a gas mixture in a 30%wt MEA aqueous solution) have been performed both on mini-modules and at pilot-scale (10 m2 membrane contactor module) based on PTFE hollow fibers. The results have been modeled utilizing a resistance in series approach. The only adjustable parameter is in fitting the simulations to experimental data is the membrane mass transfer coefficient (km), which logically plays a key role. The difficulties and uncertainties associated with scaleup computations from lab scale to pilot-scale modules, with a particular emphasis on the km value, are presented and critically discussed.@en

Motivation: Protein function prediction is a difficult bioinformatics problem. Many recent methods use deep neural networks to learn complex sequence representations and predict function from these. Deep supervised models require a lot of labeled training data which are not available for this task. However, a very large amount of protein sequences without functional labels is available. Results: We applied an existing deep sequence model that had been pretrained in an unsupervised setting on the supervised task of protein molecular function prediction. We found that this complex feature representation is effective for this task, outperforming hand-crafted features such as one-hot encoding of amino acids, k-mer counts, secondary structure and backbone angles. Also, it partly negates the need for complex prediction models, as a two-layer perceptron was enough to achieve competitive performance in the third Critical Assessment of Functional Annotation benchmark. We also show that combining this sequence representation with protein 3D structure information does not lead to performance improvement, hinting that 3D structure is also potentially learned during the unsupervised pretraining.@en