Machine learning applied for metabolic flux-based control of micro-aerated fermentations in bioreactors.

Various bio-based processes depend on controlled micro-aerobic conditions to achieve a satisfactory product yield. However, the limiting oxygen concentration varies according to the micro-organism employed, while for industrial applications, there is no cost-effective way of measuring it at low levels. This study proposes a machine learning procedure within a metabolic flux-based control strategy (SUPERSYS_MCU) to address this issue. The control strategy used simulations of a genome-scale metabolic model to generate a surrogate model in the form of an artificial neural network, to be used in a micro-aerobic fermentation strategy (MF-ANN). The meta-model provided setpoints to the controller, allowing adjustment of the inlet air flow to control the oxygen uptake rate. The strategy was evaluated in micro-aerobic batch cultures employing industrial Saccharomyces cerevisiae yeast, with defined medium and glucose as the carbon source, as a case study. The performance of the proposed control scheme was compared with a conventional fermentation and with three previously reported micro-aeration strategies, including respiratory quotient-based control and constant air flow rate. Due to maintenance of the oxidative balance at the anaerobiosis threshold, the MF-ANN provided volumetric ethanol productivity of 4.16 g·L-1 ·h-1 and a yield of 0.48 gethanol .gsubstrate-1 , which were higher than the values achieved for the other conditions studied (maximum of 3.4 g·L-1 ·h-1 and 0.35-0.40 gethanol ·gsubstrate-1 , respectively). Due to its modular character, the MF-ANN strategy could be adapted to other micro-aerated bioprocesses.

Metabolic fluxes-oriented control of bioreactors: a novel approach to tune micro-aeration and substrate feeding in fermentations.

BACKGROUND: Fine-tuning the aeration for cultivations when oxygen-limited conditions are demanded (such as the production of vaccines, isobutanol, 2-3 butanediol, acetone, and bioethanol) is still a challenge in the area of bioreactor automation and advanced control. In this work, an innovative control strategy based on metabolic fluxes was implemented and evaluated in a case study: micro-aerated ethanol fermentation. RESULTS: The experiments were carried out in fed-batch mode, using commercial Saccharomyces cerevisiae, defined medium, and glucose as carbon source. Simulations of a genome-scale metabolic model for Saccharomyces cerevisiae were used to identify the range of oxygen and substrate fluxes that would maximize ethanol fluxes. Oxygen supply and feed flow rate were manipulated to control oxygen and substrate fluxes, as well as the respiratory quotient (RQ). The performance of the controlled cultivation was compared to two other fermentation strategies: a conventional "Brazilian fuel-ethanol plant" fermentation and a strictly anaerobic fermentation (with ultra-pure nitrogen used as the inlet gas). The cultivation carried out under the proposed control strategy showed the best average volumetric ethanol productivity (7.0 g L-1 h-1), with a final ethanol concentration of 87 g L-1 and yield of 0.46 gethanol g substrate -1 . The other fermentation strategies showed lower yields (close to 0.40 gethanol g substrate -1 ) and ethanol productivity around 4.0 g L-1 h-1. CONCLUSION: The control system based on fluxes was successfully implemented. The proposed approach could also be adapted to control several bioprocesses that require restrict aeration.

Anaerobic digestion of wastewater rich in sulfate and sulfide: effects of metallic waste addition and micro-aeration on process performance and methane production.

This work explores the effect of two metallic wastes (mining wastes, MW; fly ashes, FA) and micro-aeration (MA) on the anaerobic digestion of wastewater which is rich in sulfate and sulfide. Two initial COD concentrations (5,000 and 10,000 mg/L) were studied under both conditions in batch systems at 35 °C, with a fixed COD/SO42- ratio = 10, with 100 mg/L of S2-. It was observed that the use of MW and FA in the assays with an initial COD concentration of 10,000 mg/L resulted in a simultaneous increase in COD removal, sulfate removal, sulfide removal and methane generation, while MA only improved the COD and sulfide removals in comparison with the control system. On the contrary, the use of MW, FA or MA in systems with initial COD concentrations equal to or lower than 5,000 mg/L did not show any improvement with respect to the control system in terms of COD removal, sulfate removal or methane generation, with only sulfide removal being positively affected by MW and FA.

Recovery of elemental sulphur from anaerobic effluents through the biological oxidation of sulphides.

The aim of the present study was to evaluate the biological oxidation of sulphide in two different UASB reactors by assessing the occurrence of oxidized forms of sulphur in the effluents and the amount of S0 that could be recovered in the process. The bioreactors employed were an anaerobic hybrid (AH) reactor employing porous polyurethane foam as support media and a micro-aerated UASB reactor equipped with an aeration device above the digestion zone. The AH reactor produced a final effluent containing low concentrations of S2- (3.87% of total sulphur load). It was achieved due to a complete oxidation of 56.1% of total sulphur. The partial biological oxidation that occurred in the AH reactor allowed the recovery of 30% of the sulphur load as S0. The effluent from the micro-aerated UASB reactor contained 5% of the sulphur load in the form of S2-, while 20.9% was present as dissolved SO42- and 46% was precipitated as S0. It is concluded that the AH reactor or micro-aeration carried out above the digestion zone of the UASB reactor favoured the biological oxidation of S2- and the release of odourless effluents. Both technologies represent feasible and low-cost alternatives for the anaerobic treatment of domestic sewage.

Performance evaluation of micro-aerobic hydrolysis of mixed sludge: Optimum aeration and effect on its biochemical methane potential.

This study evaluated the performance of a micro-aerobic hydrolysis of mixed sludge and its influence as a pretreatment of this waste for its subsequent anaerobic digestion. Three experimental series were carried out to evaluate the optimum micro-aeration levels in the range from 0.1 to 0.5 air volume/min.reactor volume (vvm) and operation times within the range of 24-60 h. The maximum methane yield [35 mL CH4/g volatile suspended solids (VSS) added] was obtained for an aeration level of 0.35 vvm. This methane yield value increased 114% with respect to that obtained with the non-aerated sludge. In the micro-aeration process carried out at an aeration level of 0.35 vvm, increases in soluble proteins and total sugars concentrations of 185% and 192% with respect to their initial values were found, respectively, after 48 h of aeration. At the above micro-aerobic conditions, soluble chemical oxygen demand (CODS) augmented 150%, whereas VSS content decreased until 40% of their initial respective values. Higher COD increases and VSS decreases were found at 60 h of micro-aeration, but the above parameters did not vary significantly with respect to the values found at 48 h.