A computer-aided approach to compare the production economics of fed-batch and perfusion culture under uncertainty

Fed-batch and perfusion culture dominate mammalian cell culture production processes. In this paper, a decision-support tool was employed to evaluate the economic feasibility of both culture modes via a case study based upon the large-scale production of monoclonal antibodies. The trade-offs between the relative simplicity but higher start-up costs of fed-batch processes and the high productivity but higher chances of equipment failure of perfusion processes were analysed. Deterministic analysis showed that whilst there was an insignificant difference (3%) between the cost of goods per gram (COG/g) values, the perfusion option benefited from a 42% reduction in capital investment and a 12% higher projected net present value (NPV). When Monte Carlo simulations were used to account for uncertainties in titre and yield, as well as the risks of contamination and filter fouling, the frequency distributions for the output metrics revealed that neither process route offered the best of both NPV or product output. A product output criterion was formulated and the options that met the criterion were compared based on their reward/risk ratio. The perfusion option was no longer feasible as it failed to meet the product output criterion and the fed-batch option had a 100% higher reward/risk ratio. The tool indicated that in this particular case, the probabilities of contamination and fouling in the perfusion option need to be reduced from 10% to 3% for this option to have the higher reward/risk ratio. The case study highlighted the limitations of relying on deterministic analysis alone.     

Growth monitoring and control through computer-aided on-line mass balancing in a fed-batch penicillin fermentation

A computer-aided methodology is developed for on-line monitoring and control of cell growth in fed-batch penicillin fermentation using a semidefined medium containing low corn steep liquor concentration (5.7 g/L). Cell growth is monitored and controlled with the use of experimental correlation and carbon-balancing equatiions on a real-time basis throughout the fermentation. Through a combination of feed-forward and feedback control of sugar addition, residual glucose concentration in the broth was maintained below 1 g/L and cell-growth rate was kept at constant at preset vaiues. The accuracy and reproducibility of this technique are demonstrated. The use of real-time control of cell growth is expected to aid future investigations of this antibiotic fermentation.     

Two-dimensional fluorescence spectroscopy: a novel approach for controlling fed-batch cultivations

In industrial fed-batch cultivations it is often necessary to control substrate concentrations at a low level to prevent the production of overflow metabolites and thus optimize the biomass yield. A new method for on-line monitoring and fed-batch control based on fluorescence measurements has been developed. Via instantaneous in situ measurements and multivariate data analysis a chemometric model has been established, which enables the rapid detection of ethanol production at aerobic Saccharomyces cerevisiae fed-batch cultivations. The glucose feed rate is controlled by predicting the metabolic state directly from the fluorescence intensities. Thus, ethanol production could be avoided completely while increasing the biomass yield accordingly. The robust instrumentation is suitable for industrial applications.     

Growth monitoring and control in complex medium: a case study employing fed-batch penicillin fermentation and computer-aided on-line mass balancing

To broaden the practicality of on-line growth monitoring and control, its application in fedbatch penicillin fermentation using high corn steep liquor (CSL) concentration (53 g/L) is demonstrated. By employing a calculation method that considers the vagaries of CSL consumption, overall and instantaneous carbon-balancing equations are successfully used to calculate, on-line, the cell concentration and instantaneous specific growth rate in the penicillin production phase. As a consequence, these equations, together with a feedback control strategy, enable the computer control of glucose feed and maintenance of the preselected production-phase growth rate with error less than 0.002 h(-1).     

Biological treatment of high strength wastewater by fed-batch operation

Biological treatment systems for high strength wastewaters are usually operated in continuous mode such as activated sludge systems. When operated at steady-state, continuous systems result in constant effluent standards. However, in the presence of shock loadings and/or toxic compounds in feed wastewater, system performance drops quite significantly as a result of partial loss of microbial activity. In fed-batch operation, wastewater is fed to the aeration tank with a flow rate determined by effluent standards. In this type of operation, wastewater can be fed to biological oxidation unit intermittently or continuously with a low flow rate without any effluent removal. Feed flow rate is adjusted by measuring COD concentration in the effluent. As a result of intermittent addition of wastewater high COD concentrations and toxic compounds are diluted in large volume of aeration tank and inhibition effects of those compounds are reduced. As a result, biological oxidation of these compounds take place at a much higher rate. In order to show the aforementioned advantage of fed-batch operation, a high strength synthetic wastewater consisting of diluted molasses, urea, KH₂PO₄ and MgSO₄ was treated in an biological aeration tank by fed-batch operation. Organisms used were an active and dominant culture of Zooglea ramigera commonly encountered in activated sludge operations. COD removal kinetics was found to be first order and the rate constant was determined.     

Dispersion-induced behavior in sub-critical operation of a recombinant fed-batch fermentation with run-away plasmids

Fermentations with recombinant bacteria containing run-away plasmids are typically operated alternately above and below a critical temperature. To minimize the risks of run away reactions, it is preferable to keep the high temperature periods as short as possible. In this study the possibility of sustained low temperature (sub-critical) operation in a suitably non-homogeneous broth is analyzed. Fluid dispersion is used as a measure of non-homogeneity. The fed-batch production of β-galactosidase by Escherichia coli containing the plasmid pOU140 and operated below 37 ^°C is analysed as a model system. To characterize non-homogeneity, an earlier model visualizing the broth as a set of two reactors with internal recycle has been modified for fed-batch fermentation. Three dilution rates, two internal and one external, quantify fluid dispersion. While plasmid replication and fermentation become quenched in sub-critical operation in a well-mixed reactor, with finite dispersion there may be an increase in the concentration of plasmid-containing cells and the recombinant protein. The concentration profiles many also have one or more peaks in the time domain. Thus, sustained fermentation with run-away plasmids appears feasible in a bioreactor with controlled non-homogeneity.     

Constant-volume fed-batch operation for high density cultivation of hyperthermophilic aerobes

When hyperthermophilic archaebacteria are cultivated under aerobic conditions, the loss of the culture volume becomes significant due to vaporization of water with aeration. To eliminate the problems caused by evaporation of the medium, we have devised a constant-volume fed-batch system, where the loss of water due to evaporation is compensated by feeding additional water into the fermentor. The feeding strategy for maintaining a constant volume during fed-batch cultivation was developed from the water balance and the cell growth and substrate consumption kinetics, and the developed method was applied to high density cultivation of Sulfolobus solfataricus DSM 1617. The maximum cell density obtained was 22.6 g/L, which is the highest value reported for hyperthermophiles in fed-batch operations.     

Simple nonsingular control approach to fed-batch fermentation optimization

Determination of the optimal feed rate for fed-batch fermentation is normally a problem in singular control with a state inequality constraint and as such is, in general, difficult to solve, especially for those described by a large number of dynamic mass balance equations. In this article we use a new set of state variables and the culture volume as the control variable. In this way the problem is converted to one of nonsingular control with the magnitude and rate constraints on the manipulated variable and can be numerically solved by a gradient-based technique, thus avoiding the difficulty associated with singular control problems. Examples are given to illustrate the method.     

Lactic acid production directly from starch in a starch-controlled fed-batch operation using Lactobacillus amylophilus

Based on the batch results, we constructed a simplified simultaneous saccharification and fermentation (SSF) model for the simulation of lactic acid production directly from unhydrolyzed potato starch using Lactobacillus amylophilus. The results of batch operation at different initial starch concentrations (20, 40 and 60 g/l) indicated that a higher initial starch concentration would lead to a slightly lower productivity, but would largely decrease the yield. Among that, the batch with 20 g/l of initial starch had the maximum productivity and the maximum yield, which would be 0.31 g/(l h) and 98% (g/g), respectively. In view of increasing the productivity and the final lactic acid concentration, a starch-controlled fed-batch operation with 20 g/l of initial starch was performed. It showed the fed-batch operation with starch controlled at 8 ± 1 g/l by adjusting the starch-feeding rate led to the maximum productivity of 0.75 g/(l h) and the yield of 69%.     

Effects of lactate concentration on hybridoma culture in lactate-controlled fed-batch operation

To investigate the effects of lactate on cell growth and antibody production, a new method of maintaining the lactate concentration constant in a fed-batch culture is described. When the pH was initially adjusted by sodium hydroxide, the specific growth rate decreased and specific death rate increased with an increase of lactate concentration. To investigate whether the inhibition was due to the lactate concentration itself or to the osmotic pressure, the effect of the osmotic pressure adjusted by sodium chloride was compared with that of sodium lactate. When the osmotic pressure was adjusted to same condition as that of sodium lactate using sodium chloride, the specific growth data showed the same degree of growth inhibition. It was thus evident that the inhibition to cell growth was mainly due to osmotic pressure while lactate production from glucose was found to be inhibited by the lactate itself compared with sodium chloride. The specific antibody production rate had a maximum value within a certain range of lactate concentration. Moreover, specific antibody production rate had a unified relationship with the kinetic parameter mu, in spite of the different causes of inhibition by lithium lactate and sodium lactate. A certain "trade-off" relationship between growth and antibody production existed at higher growth rates.     

Adsorbent supplemented biological treatment of pre-treated landfill leachate by fed-batch operation

Biological treatment of landfill leachate usually results in low COD removals because of high chemical oxygen demand (COD), high ammonium-N content and presence of toxic compounds. Coagulation-flocculation with lime addition and air stripping of ammonia were used as pre-treatment in this study in order to improve biological treatability of the leachate. Pre-treated leachate was subjected to adsorbent supplemented biological treatment in an aeration tank operated in fed-batch mode. COD and NH(4)-N removal performances of powdered activated carbon (PAC) and powdered zeolite (PZ) were compared during biological treatment. Adsorbent concentrations varied between 0 and 5 gl(-1). Percent COD and ammonium-N removals increased with increasing adsorbent concentrations. Percent COD removals with PAC addition were significantly higher than those obtained with the zeolite. However, zeolite performed better than the PAC in ammonium-N removal from the leachate. Nearly 87% and 77% COD removals were achieved with PAC and zeolite concentrations of 2 gl(-1), respectively. Ammonium-N removals were 30% and 40% with PAC and zeolite concentrations of 5 gl(-1), respectively at the end of 30 h of fed-batch operation.     

A heuristic approach to fed-batch optimisation of streptokinase fermentation

Previous studies have shown that the rate of formation of streptokinase, a secondary metabolite, in batch fermentation is proportional to the specific growth rate of the biomass, which in turn is inhibited by its substrate and the primary product (lactic acid). These kinetics suggest the suitability of fed-batch operation to increase the yield of streptokinase. A near-optimal feed policy has been calculated by the chemotaxis algorithm, and it shows a substrate feed rate decreasing nonlinearly and vanishing after 11 hours. This is followed by batch fermentation for a further 8 hours, at the end of which 12% more streptokinase is generated than by purely batch fermentation. Further improvements in productivity are possible.List of Symbols k _dh^–1 decay constant for active cells - k _ph^–1 decay constant for streptokinase - K _Igl^–1 inhibition constant for lactic acid - K_S gl^–1 inhibition constant for substrate - M gl^–1 lactic acid concentration - P gl^–1 streptokinase concentration - Q 1h^–1 substrate feed rate - S gl^–1 substrate concentration - S _ingl^–1 inlet concentration of substrate - t h time - t _bh end-point of batch fermentation - t _fh end-point of fed-batch fermentation - V l volume of broth in fermenter - V ₀ l initial value of V (at t=0) - V _ml maximum value of V - X gl^–1 total biomass concentration - X _agl^–1 concentration of active biomass - Y _MX yield coefficient for lactic acid from biomass - Y _PX yield coefficient for streptokinase from biomass - Y _XS yield coefficient for biomass from substrate Greek Letters h^–1 specific growth rate of biomass - _mh^–1 maximum specific growth rate     

Microbial population dynamics during fed-batch operation of commercially available garbage composters

Microbial populations in terms of quantity, quality, and activity were monitored during 2 months of start-up operation of commercially available composters for fed-batch treatment of household biowaste. All the reactors, operated at a waste-loading rate of 0.7 kg day^–1 (wet wt), showed a mass reduction efficiency of 88–93%. The core temperature in the reactors fluctuated between 31°C and 58°C due to self-heating. The pH declined during the early stage of operation and steadied at pH 7.4–9.3 during the fully acclimated stage. The moisture content was 48–63% early in the process and 30–40% at the steady state. Both direct total counts and plate counts of bacteria increased via two phases (designated phases I, II) and reached an order of magnitude of 10¹¹ cells g^–1 (dry wt) at the steady state. Microbial community changes during the start-up period were studied by culture-independent quinone profiling and denatured gradient gel electrophoresis (DGGE) of PCR-amplified 16S rDNA. In all the reactors, ubiquinones predominated during phase I, whereas partially saturated menaquinones became predominant during phase II. This suggested that there was a drastic population shift from ubiquinone-containing Proteobacteria to Actinobacteria during the start-up period. The DGGE analysis of the bacterial community in one of the reactors also demonstrated a drastic population shift during phase I and the predominance of members of the phyla Proteobacteria and Bacteroidetes during the overall period. But this molecular analysis failed to detect actinobacterial clones from the reactor at any stage.     

Optimal fed-batch operation of recombinant cells subject to plasmid instability and death

Optimal feed rate strategy is studied for fed-batch culture of recombinant cells with plasmid instability and with different death rates for the plasmid-free cells (PFC) and plasmid-bearing cells (PBC). Most of the fed-batch fermentation is known to have first-order singularity and therefore a single singular arc. However, this study shows that a singular arc with second-order singularity and therefore two distinct singular arcs are possible for a recombinant cell process if PFC and PBC are subjected to death, and their specific growth rates are proportional to each other. Two types of singular arcs are elucidated and analyzed. The optimal policies over the singular arcs are theoretically explored as these findings reveal qualitative information on the singular arc, which is critically important in providing the optimal initial conditions in numerical computation of optimal feed rate profile.     

Approximation of continuous growth of Cephalotaxus harringtonia plant cell cultures using fed-batch operation

In Cephalotaxus harringtonia plant cell cultures, periods of batch growth that are limited by hexose uptake are too short to make an accurate estimate of the Monod saturation constant. Continuous cultures are infeasible on a laboratory scale, and semicontinuous cultures require too frequent sampling. Fed-batch operation, consisting of intermittent removal from a culture that is fed continuously, was investigated as a possible solution to these problems. For a constant feed rate, computer simulations showed that a steady state can be achieved which is useful for studying growth at different specific growth rates. In terms of the dilution rate it was confirmed that the operation is essentially equivalent to continuous culture when the samples represent a small fraction of the total culture volume. Experiments with glucose or fructose as the carbon source were carried out in shake flasks fed by a multichannel syringe pump. Results indicate that Monod kinetics based on medium glucose levels cannot adequately describe growth under these conditions. Monod's expression for specific growth rate using internal glucose concentration gives an improved correlation.     

Phosphate enrichment and fed-batch operation for prolonged beta-lactamase production by Bacillus licheniformis

AIMS: Investigation of the phosphate effect and feeding strategy, i.e. linear and exponential feeding, to improve beta-lactamase production by Bacillus licheniformis considering the viability of the cells. METHODS AND RESULTS: Effect of phosphate enrichment on beta-lactamase production was investigated and resulted in 1.2-fold increase in beta-lactamase activity. Thereafter, exponential and linear feed profiles were established, after an initial batch phase for t = 0-7.5 h. The highest beta-lactamase activity was obtained at fed-batch operation with exponential feeding (FBO1) condition as A = 106 U cm(-3), which is c. 1.7-fold higher than that of the phosphate-enriched batch operation (PE-BO). CONCLUSIONS: Biphasic variations in beta-lactamase production was enhanced to monophasic variation with the exponential feeding strategy where the activity was obtained as A = 106 U cm(-3) at t = 16 h. SIGNIFICANCE AND IMPACT OF THE STUDY: Phosphate enrichment decreases the intracellular ammonium concentration and organic acid excretion, but increrases beta-lactamase production. When batch operation (BO) and PE-BO are compared, it is seen that succinic acid formation decreased with the phosphate enrichment as a result of smooth operation of the tricarboxylic acid cycle. At FBO1 despite the increased lactic and acetic acid formation, beta-lactamase production increased 1.7-fold, and 92% of the cells were alive at the end of the fermentation.     

Dynamic reoptimization of a fed-batch fermentor

Traditionally, fed-batch biochemical process optimization and control use complicated models and off-line optimizers with no on-line model adaptation and re-optimization. This work demonstrates the applicability, effectiveness, and economic potential, of a simple phenomenological model for modeling, and a novel optimizer for on-line re-optimization and control of an aerobic fed-batch fermentor.