Simulated cell trajectories in a stratified gas–liquid flow tubular photobioreactor
- 336 Downloads
The fluid dynamic environment within a photobioreactor is critical for performance as it controls mass transfer of photosynthetic gases (CO2 and O2) and the mixing environment of the algal culture. At a cellular level, light fluctuation will occur when cells move between the “light”, well-illuminated volume of the culture near the light source and the “dark”, self-shaded zone of the culture. Controlled light/dark frequency may increase the light to biomass yield and prevent photoinhibition. Knowledge of cell trajectories within the reactor is therefore important to optimize culture performance. This study examines the cell trajectories and light/dark frequencies in a stratified gas–liquid flow tubular photobioreactor. Commercially available computational fluid dynamics software, ANSYS Fluent, was used to investigate cell trajectories within the half-full solar receivers at different liquid velocities and reactor tube diameters. In the standard configuration 96-mm solar receiver tube, the light/dark cycle frequencies ranged from 0.104 to 0.612 Hz over the liquid velocity range of 0.1 to 1 m s−1. In comparison, the smaller diameter 48- and 24-mm tubes exhibit higher light/dark frequencies, 0.219 to 1.30 Hz and 0.486 to 2.67 Hz, respectively.
KeywordsComputational fluid dynamics (CFD) Design Hydrodynamics Mixing Particle tracking
Computational fluid dynamics
Microalgal biomass can be used for a wide range of products and have several advantages over more conventional crops. Production can take place in regions and climates not suitable for farmland and forests, with no requirement for freshwater (Borowitzka and Moheimani 2011). Compared to land grown plants, they also exhibit faster growth and shorter harvesting cycles. In addition, the production of algal biomass can be synergistic with environmental applications such as carbon dioxide sequestration and wastewater remediation (Schenk et al. 2008).
Due to the high cost of production, microalgae products are currently only used within high value sectors such as pharmaceuticals and health foods (Pulz and Gross 2004). However, should large-scale microalgal biomass production be made economically viable, it has the potential to be a part of future sustainable communities as a source of protein for food and feed, lipids for renewable fuel, and other energy or carbon sequestration products. For microalgal biomass to be economically competitive for such purposes, reliable, high-performance, large-scale production systems will be needed. Closed culturing systems (photobioreactors) provide intrinsic advantages over open systems such as raceway ponds in terms of process control and productivity, but the challenge is to realize these benefits at low capital costs (Pulz 2001; Carvalho et al. 2006).
It is also critical to have a comprehensive understanding of the variables involved in microalgae cultivation and their impact on biomass production and composition. One important criterion for optimized microalgal production is sufficient mass transfer of photosynthetic gases so that the culture growth is limited only by the light availability and a suitable light regime on a cellular level.
Mass transfer within the design takes place in both sections and has previously been investigated (Moberg et al. 2009; Moberg et al. 2010). These studies analyzed the relationships between productivity, liquid velocity, gas composition, and reactor length to support high productivity operations. The focus of the current study is radial mixing within the liquid phase, its effect on the light regime, and how that varies with liquid velocity and the diameter of the solar receivers.
Reactor mixing determines the light exposure pattern of individual cells in the culture. Microalgal cultures are self-shading, i.e., mutual shading occurs among cells. The available light decreases exponentially from the surface of the reactor that is closest to the irradiation source. The difference between light and dark parts of the culture increases with increased culture density or longer light paths (Merchuk et al. 1998; Molina Grima et al. 1999). Due to this mutual shading cells are not experiencing continuous illumination, but a pattern of light and darkness as they move through the light and dark volumes of the reactor. These illumination cycles are called light/dark (L/D) cycles. The frequency of these cycles and the ratio of L/D periods are important, as these parameters can be optimized to use high irradiances more effectively. Increased productivity is a result of the culture capturing more photons while avoiding light inhibition, which occurs when a cell is exposed to too much light (Richmond 1986). Light/dark cycling is effectively a way of diluting light by giving each cell a smaller dose of light during a given time span. In cultures of high cell density, a higher degree of mixing, and hence, L/D cycles of higher frequency, are needed to effectively utilize light. Ideally, the L/D cycle should be equal or close to the photosynthesis reaction time which is approximately 1–15 ms (Richmond et al. 2003; Carvalho et al. 2011). Such conditions are however difficult to achieve in practical fluid dynamic conditions, and would require short light paths and extremely good mixing conditions. Even if the maximum effect of L/D cycles cannot be reached, reactor design should aim for conditions with L/D frequencies within an order of magnitude of 1 Hz, where increases in culture efficiency have been shown to occur (Grobbelaar et al. 1996; Eriksen 2008; Posten 2009), and a photic volume of ca. 10% to yield 1:10 light to dark ratio L/D cycles (Richmond 1996; Degen et al. 2001).
Light/dark cycle frequencies within photobioreactors are seldom known, as they depend both on the cell trajectories and the radiative field within the reactor, both of which can be difficult to determine. A few studies on the characterization of the light regimes within photobioreactors have however been published, where the main approach used is to first determine cell trajectories and then combine the results with a light distribution profile (Eriksen 2008; Pruvost et al. 2008). Cell trajectories in photobioreactors have previously been determined by either schematic representation of the flow (bubble column, internal loop airlift, and thin layer cascade) (Wu and Merchuk 2002; Wu and Merchuk 2004; Masojídek et al. 2011), experimental measurement (bubble column, draft tube, and split airlift columns) (Luo et al. 2003; Luo and Al-Dahhan 2004; Merchuk et al. 2007) or simulations (annular cell, tubular with static mixers, spiral tubular, and draft tube airlift) (Pruvost et al. 2002; Perner-Nochta and Posten 2007; Wu et al. 2010; Luo and Al-Dahhan 2011).
This study examines cell trajectories within half-full solar receivers as a first step to characterize the light regime within the studied photobioreactor. Commercially available computational fluid dynamics (CFD) software, ANSYS Fluent, is used to investigate three half-full tubes of different diameters over a liquid velocity range. A simplified light distribution profile is then used to compare light regimes at different liquid velocities in the standard case 96-mm solar receiver tube. The results are then compared to two tubes of smaller diameter. The aim of the study is to investigate if any conclusions for reactor design and operating purposes may be drawn from the results, and to develop a light regime model that later can be used to calculate optimal combinations of culture density, solar receiver tube diameter, and liquid velocity to achieve suitable L/D cycle frequency and light to dark interval ratios for a chosen species and location.
Geometry and mesh generation
The examined photobioreactor consist of two sections, one airlift section and one tubular (solar receiver) section (Fig. 1). The airlift section is mainly dark and cells will experience negligible light exposure within this section. It was therefore not included in the following simulations. The solar receiver section consists of partially filled tubes with an inner diameter of 96 mm in the standard configuration. In this study, three different tube diameters were investigated: 96, 48, and 24 mm. The half-full tube (half-cylinders) geometries were created by extruding half-circles by 10,000 mm in the z-direction in ANSYS DesignModeler 12.1. Geometries were thereafter imported into ANSYS Meshing 12.1 where they were meshed by the sweep method. Face sizing and edge inflation were used to control the mesh size. The resulting 96-, 48-, and 24-mm diameter meshes consisted of approximately 990,000, 520,000, and 410,000 cells for the half-cylinder geometries.
Fluid flow solution
Determination of cell trajectories
Cell (particle) trajectories were calculated within ANSYS Fluent 12.1.4 using the stochastic Discrete Random Walk model. To emulate the properties of microalgae cells, particle reflection was included in the wall boundary conditions while “interaction with liquid phase” (used when particles exchange mass, momentum, and/or energy with the continuous phase) and “Saffman Liftforce” (used to model lift due to shear for sub-micron particles) options were neglected. Particles were defined as spherical particles with the diameter of 0.01 mm and a density of 1,000 kg m−3. Surface injection was chosen, meaning one particle per inlet face was released (983, 407, and 171 particles respectively for the 96-, 48-, and 24-mm geometries). Five surface injections were performed for each scenario. Recorded variables were the particle ID and the spatial (x, y, z) position of each particle over time.
Post processing and light regime calculations
For the purpose of this study, a simplified light/dark environment was modeled where the light source was assumed to be positioned vertically above the solar receiver. Refraction effects and diffuse light were neglected. The result is an upper volume of the culture assumed to be “light” and a bottom volume assumed “dark” (Fig. 2b). The volume fraction in each zone is determined by the light/dark cut-off level or light penetration depth. The light penetration depth is dependent on reactor geometry, light intensity, cell density, and the properties of individual algal species. In this study, the light penetration depth in each geometry was chosen so that the illuminated volume constituted 10%, a light to dark culture volume ratio of 1:9. The resulting cut-off levels were located 3.8, 1.9, and 0.94 mm below the surface of the 96-, 48-, and 24-mm half-cylinder geometries, respectively.
To calculate the light regimes of each particle the particle tracking variables recorded in ANSYS Fluent were imported and processed in MATLAB 7.7.0 (Mathworks). The particle tracks were separated by their particle ID and L/D cycles were calculated from the y–z (distance from surface and distance from inlet) position at each time point. The mean light and dark intervals in each scenario were calculated from two vectors containing all light and dark intervals recorded (based on the L/D cut-off value) for all particles, and the mean L/D cycle as a sum of these two means. The frequency was calculated as the reciprocal of the length of a L/D cycle. The vectors containing light and dark intervals were also used for the calculation of interval residence time distributions. The light and dark intervals of individual particle tracks were visualized by calculating the binary light/dark pattern. When the position of a particle was above the light/dark cut-off level the particle was in the light zone and given the binary value of 1. When the particle was in the dark zone it was given a value of 0.
Cell trajectories and binary pattern
Light regimes in the 96-mm tube
For each tested liquid velocity in the three geometries, the mean light and dark intervals of all particle tracks in all injections were calculated. The mean L/D cycle length and frequency was derived from this data. The results show an overall trend where a higher liquid velocity increases the rate of L/D cycling, as does smaller tube diameters.
Light regimes in the 48- and 24-mm tubes
This study has been based on particle trajectories calculated by the commercially available CFD code ANSYS Fluent 12.1.4. The particle tracking model is relying on the fluid flow simulation and it should be noted that the calculated solution is only an approximation of real flow. As the investigated solar receivers consist of basic geometries and fluid flows that are straightforward to model with CFD this approach was deemed sufficient for comparing different tube diameters and operating velocities for design purposes.
The unstructured cell trajectories (Fig. 4) and varying length of light and dark intervals (Fig. 5) show that the vertical/radial mixing of particles within the examined photobioreactor is not of a regular, but of a random nature. Particles will be subjected to L/D cycles of different length as they move through the solar receiver. Individual algae cells may therefore at times experience photoinhibition caused by too much light exposure, others may not be productive because of long dark periods. However, light and dark interval distribution analysis of the example case of a half-full 96-mm solar receiver operating at 0.5 m s−1 showed that a large fraction of light intervals, 19.9%, were shorter than or equal to 100 ms, 86.4% equal to or shorter than 1 s. Sixty percent of dark intervals were shorter than or equal to 1 s, and 94.3% were shorter than or equal to 10 s.
The 1:9 volume fraction light regime results show that there is a reasonably good mixing rate within the solar receiver, with L/D cycles in the range of seconds. For comparison, the L/D cycles in raceway ponds that have a similar light path length are in the order of seconds to minutes (Grobbelaar et al. 1996). In the standard case 96-mm half-full tube, L/D frequencies over the liquid velocity range of 0.1 to 1 m s−1 are 0.104 and 0.612 Hz (Fig. 9), within an order of magnitude of 1 Hz. Previous studies have reported conflicting results within this medium frequency range of 0.1–1 Hz (Grobbelaar et al. 1996; Janssen et al. 2000), but it is plausible that some productivity increase could be seen in this range compared to cultures experiencing lower L/D frequencies.
The 48- and 24-mm cases showed that an increased L/D frequency can be expected in smaller diameter tubes. Frequencies above 1 Hz were seen at the tested liquid velocity of 1 m s−1 in the 48-mm tube and at 0.3 m s−1 and above in the 24-mm tube (Fig. 9). In this range the vertical mixing is very likely to contribute to increased productivity through L/D cycling (Grobbelaar et al. 1996; Eriksen 2008; Posten 2009). This indicates that smaller diameter tubes could be advantageous in terms of light regimes, but other design considerations also need to be considered. This includes the impact on mass transfer within the solar receiver and the increased cell density necessary in a smaller diameter tube to create a suitable light regime within a shorter light path culture environment.
In conclusion, this study has shown that the fluid flow-induced radial mixing within the investigated photobioreactor design can support favorable light regimes. Light/dark frequencies of the order of 1 Hz were seen in all the tested scenarios, indicating that the reactor can be operated in a range where productivity increases are plausible compared to cultures experiencing lower frequency L/D cycles. The CFD-based cell trajectory model used in this study will be an important tool for further optimization of light exposure within the investigated photobioreactor design, and will form the basis of future work looking at the operational management and performance optimization of the reactor in realistic light and cell density environments. In conjunction with the previously developed mass transfer models (Moberg et al. 2010), these models form the basic tools for overall system optimization of the photobioreactor process parameters and design.
Future work will investigate how optimal combinations of culture density, solar receiver tube diameter, and liquid velocity can be calculated to achieve suitable L/D cycle frequency and light to dark interval ratios for a chosen species and location. Other implementations will include using a light gradient instead of a L/D cut-off level that will allow for more detailed calculation of average cell illumination history, from which conclusions can be drawn on a culture level. In an outdoor environment the irradiance levels are highly variable throughout the day, and the light regime may also change due to changing cell density in the culture. By using a developed light regime model, operational parameters, such as liquid velocity, may be adjusted to prevent photoinhibition and minimize energy use, thereby maximizing culture performance in a varying light environment.
This work is part of a PhD project investigating the process engineering fundamentals of microalgae production. The project is supported by the industry partner, The Crucible Group Pty Ltd, The Tom Farrell Institute for the Environment, and The University of Newcastle, Australia.
- Borowitzka MA, Moheimani NR (2011) Sustainable biofuels from algae. Mitig Adapt Strateg Glob Change. doi: 10.1007/s11027-010-9271-9
- Moberg AK, Ellem G, Jameson GJ, Herbertson J (2009) Mass transfer of photosynthetic gases within a multiphase microalgal production system. In conference proceedings of Chemeca 2009: Engineering Our Future: Are We Up to the Challenge?, Perth, Australia. Engineers Australia, Barton, pp 425–436. ISBN: 9780858259225Google Scholar
- Moberg AK, Ellem G, Jameson GJ, Herbertson J. (2010) Fluid dynamical considerations in the design of a microalgae photobioreactor. In conference proceedings of Chemeca 2010: Engineering at the Edge, Adelaide, Australia. Engineers Australia, Barton, pp 421–431. ISBN: 9780858259713Google Scholar
- Molina Grima E, Acién Fernández FG, García Camacho F, Chisti Y (1999) Photobioreactors: light regime, mass transfer, and scaleup. J Biotechnol 70:231–47Google Scholar
- Richmond AE (1986) Microalgaculture. CRC Crit Rev Biotechnol 4:369–428Google Scholar