Bioreactor design is an important concern and element for the biotechnology and biochemical engineering industry. There are issues with all bioreactors that have to be confronted – the competing elements are between adequate mixing, controlling reactions and various transfer phenomena which determines the degree of detrimental heterogeneity. It was realised a long time ago that the potential for a mismanaged bioreactor gets worse as the whole system becomes larger (Enfors et al., 2001).
Bioreactor Models
There have been methods available over the years to understand how a reactor can be designed and scale-up. Here we look at some of the critical methods in scale-up that have been employed over many years.
Having the knowledge and being able to predict mixing and mass transfer in bioreactors especially ones that need agitation is fundamental in both their development and scaling-up.
The main process parameters are mixing times and the volumetric mass transfer coefficient.
Mixing in a bioreactor is one of the factors that truly exercises fermentation (bioreactor) modellers. We’ve discussed elsewhere about mixing as an engineering science in scale-up of reactors. Having knowledge about a reactor’s mixing capabilities is an absolute requirement so that analyses over the time scale of a process can be made when assessing limitations in mixing and whether they can be predicted or expected.
There is a special emphasis on computation fluid dynamics because of its importance in recent times to design but it is by no means the only method. We will see how CFD on its own and in combination with other techniques helps to provide that knowledge.
Computational Fluid Design As A Method
Computational Fluid Dynamics (CFD) plays a crucial role in the design and optimization of bioreactors. Bioreactors are vessels used in biotechnology and pharmaceutical industries for the cultivation of cells, microorganisms, or biological molecules. The performance of a bioreactor greatly depends on the distribution of key parameters such as temperature, nutrients, dissolved gases, and shear stress within the culture medium. CFD simulations enable engineers and researchers to understand and optimize these parameters by modeling the fluid flow, mass transfer, and heat transfer within the bioreactor. It is a method that has found wide application in the food industry too.
Fluid Flow Analysis
CFD helps in predicting the flow patterns within the bioreactor, including velocity profiles, turbulence, and recirculation zones. Understanding flow dynamics is critical for ensuring homogeneous mixing of nutrients, gases, and cells throughout the culture medium. Proper mixing enhances mass transfer rates, prevents concentration gradients, and improves the overall efficiency of the bioprocess.
CFD simulations of bioreactor hydrodynamics can be performed with commercial CFD code such as Fluent (ANSYS, Inc, Canonsburg, PA 15317, USA), Cadence Fidelity CFD Software of which there are many versions. A single phase flow is considered where the density and viscosity are defined for water i.e. density ρ =998.2 kg/m3 and viscosity μ =0.001003 PaS. If the rotation speed of an impellar is known then the Reynolds number can be calculated. A turbulence model is used which is the standard k-epsilon model usually found in CFD packages such as Fluent (Delafosse et al., 2014).
Mass Transfer Modeling
CFD allows engineers to simulate the transport of nutrients, metabolites, and gases within the bioreactor. By analyzing mass transfer phenomena, such as diffusion and convection, researchers can optimize the design parameters to achieve optimal nutrient delivery to cells and efficient removal of waste products. This helps in maximizing cell growth and product yield while minimizing substrate limitations and metabolic stress.
Heat Transfer Analysis
Maintaining the temperature within the optimal range is crucial for the viability and productivity of biological systems in bioreactors. CFD simulations help in predicting temperature distributions, heat transfer rates, and thermal gradients within the culture medium. Engineers can optimize the design of heating or cooling systems to achieve precise temperature control, thereby ensuring the stability of biological processes and preventing thermal stress on cells.
Shear Stress Prediction
Shear stress exerted by fluid flow on cells can affect cell viability, morphology, and productivity. CFD allows for the prediction of shear stress distribution within the bioreactor, enabling engineers to design systems that minimize shear-induced damage to cells while ensuring efficient mixing and mass transfer. By optimizing impeller configurations, agitation speed, and vessel geometry, researchers can mitigate the adverse effects of shear stress on cell cultures.
Scale-Up and Optimization
CFD can aid in the scale-up of bioreactor systems from laboratory-scale to industrial-scale operations. By simulating the performance of different reactor geometries and operating conditions, engineers can optimize the design parameters to maximize productivity, minimize energy consumption, and ensure consistent product quality across different scales.
Applications
CFD has been successfully used to model and solve fundamental mass and momentum balance equations using numerical techniques. In all of these cases it is needed to solve fluid-flow equations. These include the Reynolds averaged Navier–Stokes equations which are non-linear and cannot be solved analytically. To be able to solve then means linearising the equations which are then solved over many small control volumes known as the computational mesh. To be able to determine the flow field means these simulations require an input on geometry, boundary conditions and fluid properties of the bioreactor.
One of the notable areas of application is biological wastewater treatment (Essemiani et al., 2004), anaerobic digestion reactors and food processing design.
The downside of CFD is the amount of effort and time required in conducting these computations. It has also been difficult to implement solutions when exploring the many different reactions occurring in bioreactors based on their outputs. It becomes especially complex difficult when examining multiphase simulations such as gas-liquid, liquid-solid or gas-liquid-solid simulations. They are difficult to visualise using CFD codes.
The Conventional Methods
Empirical Correlations
Empirical correlations are simplified mathematical relationships derived from experimental data. These correlations relate key parameters such as mixing time, oxygen transfer rate, and shear stress to geometric and operating parameters of the bioreactor, such as impeller speed, vessel geometry, and fluid properties (Magelli et al., 2013). The correlations are simple to use and provide a good representation of the empirical data. These correlations are dependent on the measurement technique as well as the placing of the feed inputs and probes. Correlations do not account for these geographical aspects in the bioreactor. Many mixing time correlations have been developed for single-impellar vessels and no aeration. There has been some attempt to develop correlations for multi-impellar reactors but without aeration (Magelli et al., 2013).
While empirical correlations lack the detailed insights provided by CFD simulations, they are often used for quick estimations and initial design calculations.
Scale-Down Models
Scale-down models involve experimental techniques that mimic the flow behavior and mixing characteristics of large-scale bioreactors in smaller-scale setups. The scale-down step can also be thought of as the step from ‘regime analysis’ to an experimental simulation based on constant characteristic times. When considering process analysis it is only a variant of the scale-down that is based on dimensionless numbers (Asenjo, 1994).
These scale-down models, such as shake flasks, spinner flasks, and microfluidic devices, allow researchers to study bioprocess conditions at a smaller scale while retaining relevant hydrodynamic conditions. Scale-down models complement CFD simulations by providing experimental validation and insights into phenomena such as oxygen transfer and shear stress at the microscale.
Experimental Flow Visualization
Experimental flow visualization techniques, such as particle image velocimetry (PIV), laser Doppler anemometry (LDA), and dye injection methods, provide direct visualization of fluid flow patterns and velocity fields within the bioreactor. While not as quantitative as CFD, these techniques offer valuable qualitative insights into flow phenomena, mixing efficiency, and turbulence characteristics. Experimental flow visualization is often used to validate CFD predictions and optimize bioreactor designs.
Analytical Solutions
Analytical solutions involve mathematical modeling of simplified flow scenarios based on fundamental fluid dynamics principles. These analytical models, such as the ideal stirred tank reactor (ISTR) model and the axial dispersion model, provide closed-form solutions for flow, mixing, and mass transfer phenomena under idealized conditions. While less versatile than CFD, analytical solutions offer quick estimations and insights into key design parameters of bioreactors.
Dimensional Analysis and Scaling Laws
Dimensional analysis and scaling laws involve the systematic study of how physical quantities scale with system size, geometry, and operating conditions. By identifying relevant dimensionless numbers and scaling parameters, engineers can extrapolate performance data from small-scale laboratory experiments to predict the behavior of larger-scale bioreactors.
Generally, dimensionless numbers are used as a constant when scaling a process up. The Buckingham Pi theorem is then applied to determine the minimum number of dimensionless numbers. Dimensional analysis complements CFD simulations by providing insights into scale-up and scale-down principles.
Machine Learning and Data-Driven Modeling
Machine learning techniques, such as artificial neural networks (ANNs) and support vector machines (SVMs), can be used to develop data-driven models that predict bioreactor performance based on input-output relationships derived from experimental or simulation data. These models can capture complex nonlinear relationships and optimize bioprocess parameters for improved performance and efficiency.
Compartment Models
One method for predicting hydrodynamics in mixing of a bioreactor is the use of compartment models also known as multi-zone or network-of-zone models. These models divide the domain into a limited number of interconnected volumes. In these volumes, the flow properties such as concentration, turbulence, etc. are assumed to be homogeneous. Unlike CFD simulations, compartment models are less demanding on computation time and it is easier to implement reactions even when complex reaction systems are involved (Vasconcelos et al., 1998; Vrábel et al., 2000).
The main issues with compartment models is that even if a large number of compartments is considered, the fluxes between them are often deduced from global quantities such as the flow number. These are macroscopic numbers and are not able to represent all of the flow complexity in a bioreactor because the compartment models usually fail to predict accurately such mixing behaviour.
One recent approach is to combine CFD with compartment modelling. It means solving the turbulent liquid flow with CFD and then developing a compartment model based on those results of the CFD application. The fluxes between the compartments are easily computed from the CFD velocity fields.
By combining these alternative methods with CFD simulations, engineers and researchers can gain comprehensive insights into the fluid dynamics, mixing characteristics, and mass transfer phenomena within bioreactors, leading to optimized designs and enhanced bioprocess performance.
Overall, CFD plays a vital role in the design, optimization, and scale-up of bioreactor systems, contributing to the development of efficient and cost-effective bioprocesses in various industries, including pharmaceuticals, biotechnology, and biofuels.
References
Asenjo, J.A. (1994) Bioreactor System Design. edt. J.A. Asenjo & JC Merchuk. Taylor & Francis/Marcel Dekker, Inc. NY-Basel-Hong Kong
Brannock, M., Wang, Y., & Leslie, G. (2010). Mixing characterisation of full-scale membrane bioreactors: CFD modelling with experimental validation. Water Research, 44(10), pp. 3181-3191.
Delafosse, A., Collignon, M. L., Calvo, S., Delvigne, F., Crine, M., Thonart, P., & Toye, D. (2014). CFD-based compartment model for description of mixing in bioreactors. Chemical Engineering Science, 106, pp. 76-85 (Article).
Enfors, S. O., Jahic, M., Rozkov, A., Xu, B., Hecker, M., Jürgen, B., … & Manelius, Å. (2001). Physiological responses to mixing in large scale bioreactors. Journal of Biotechnology, 85(2), pp. 175-185.
Essemiani, K., Vermande, S., Marsal, S., Phan, L. and Meinhold, J., (2004) Optimisation of WWTP units using CFD – A tool grown for real scale application, Prague.
Hutmacher, D. W., & Singh, H. (2008). Computational fluid dynamics for improved bioreactor design and 3D culture. Trends in Biotechnology, 26(4), pp. 166-172.
, , , & (2013). Mixing time in high aspect ratio vessels stirred with multiple impellers. Chemical Engineering Science, 101, pp. 712–720 (Article).
Reuss, M., Schmalzriedt, S., & Jenne, M. (2000). Application of computational fluiddynamics (CFD) to modeling stirred tank bioreactors. In Bioreaction Engineering: Modeling and Control (pp. 207-246). Berlin, Heidelberg: Springer Berlin Heidelberg.
Sengur, R., Deveci, G., Kaya, R., Turken, T., Guclu, S., Imer, D. Y., & Koyuncu, I. (2015). CFD modeling of submerged membrane bioreactors (sMBRs): a review. Desalination and Water Treatment, 55(7), pp. 1747-1761
, , , & (1998). Scale-up of mixing in gassed multi-turbine agitated vessels. The Canadian Journal of Chemical Engineering, 76, pp. 398–404 (Article).
, , , & (1999). Compartment model approach: Mixing in large scale aerated reactors with multiple impellers. Chemical Engineering Research and Design, 77, pp. 291–302 (Article).
, , , , & (2000). Mixing in large-scale vessels stirred with multiple radial or radial and axial up-pumping impellers: Modelling and measurements. Chemical Engineering Science, 55, pp. 5881–5896 (Article)
Leave a Reply