Dynamic feed blending of lignocellulosic feedstocks subject to raw material variability to maintain steady-state cultivation conditions: Data set and Code

Fermentation data and process control code for dynamic bleeding of lignocellulosic feedstocks to enable steady-state processing conditions in the presence of batch-to-batch variability of renewable feedstocks. Ultra-filtered spent sulfite liquor (UF-SSL), a cheap and readily available side stream from the pulp and paper industry with a significant residual sugar concentration but subject to raw material variability, should be valorised as a primary feedstock in a continuous cultivation. Wood hydrolysate (WH) from enzymatic digestion of softwood cellulose fibres after sulfite pulping was used as a secondary, higher value feedstock with constant composition. A multiple-input multiple-output controller was used to compensate for changes in primary feedstock composition, measured in-line by FTIR spectroscopy, by dynamically adjusting UF-SSL and WH feeding, as well as bleeding and cell-free harvesting of the culture. Both feedstocks were obtained from Norway spruce (Picea abies) pulping (Borregaard AS, Sarpsborg, Norway) and stored at 4°C until use.
The laboratory set-up of the process consisted of two continuously stirred tank reactor (CSTR) bioreactors (Labfors 5, Infors, Germany), one referred to as the 'feed tank' used as a holding tank for the prepared UF-SSL solutions and the second referred to as the 'cultivation tank' used for the cultivation of C. glutamicum ATCC 13032 pVWEx1-manA pEKEx3-xylAB with improved mannose uptake and additional xylose uptake compared to the wild type. The MIMO control was active for the first 120 h of the process, after which the control was switched off and default settings for all feeds were used for reference. The UF-SSL batch used had a composition of 42.0 g/L glucose, 132.8 g/L mannose and 56.5 g/L xylose, as well as other sugars at lower concentrations that were not metabolised by the strain. Sugar concentrations were varied for the purpose of the experiment by dilution with deionised water (25-75% UF-SSL) and sugar composition was varied by dissolving solid glucose, mannose or xylose powder in the diluted UF-SSL. The WH batch used for this study had a composition of 468.5 g/L glucose, 17.9 g/L mannose and 15.0 g/L xylose and was diluted to 25% before use. A nutrient solution [112.5 g/L urea, 170 g/L KH2PO4] was used to provide bioavailable nitrogen and phosphate (NP). All solutions were supplemented before use [50 mg/L kanamycin sulphate (selection for presence of pEKEx3 plasmid), 100 mg/L spectinomycin dihydrochloride (selection for presence of pVWEx1 plasmid), 0.2 mg/L biotin (essential vitamin for C. glutamicum) and 0.5 mM isopropyl-beta-d-thiogalactopyranoside (plasmid induction)] and sterile filtered [0.22 µm]. 2.5 M H2SO4 and 2.5 M KOH were used for pH control.
The dataset consists of feedtank data including in-line FTIR results and cultivation data including off-line sample measurements.
Matlab 2024a including the System Identification Toolbox (version 24.1) was used for control implementation.