CFD simulation of mixing for high-solids anaerobic digestion

A computational fluid dynamics (CFD) model that simulates mechanical mixing for high-solids anaerobic digestion was developed. Numerical simulations of mixing manure slurry which exhibits non-Newtonian pseudo-plastic fluid behavior were performed for six designs: (i) one helical ribbon impeller; (ii) one anchor impeller; (iii) one curtain-type impeller; (iv) three counterflow (CF-2) impellers; (v) two modified high solidity (MHS 3/39°) impellers; and (vi) two pitched blade turbine impellers. The CFD model was validated against measurements for mixing a Herschel-Bulkley fluid by ribbon and anchor impellers. Based on mixing time with respect to mixing energy level, three impeller types (ribbon, CF-2, and MHS 3/39°) stand out when agitating highly viscous fluids, of these mixing with two MHS 3/39° impellers requires the lowest power input to homogenize the manure slurry. A comparison of digestion material demonstrates that the mixing energy varies with manure type and total solids concentration to obtain a given mixing time. Moreover, an in-depth discussion about the CFD strategy, the influences of flow regime and impeller type on mixing characteristics, and the intrinsic relation between mixing and flow field is included.     

Torque measurements reveal large process differences between materials during high solid enzymatic hydrolysis of pretreated lignocellulose

ABSTRACT: BACKGROUND: A common trend in the research on 2nd generation bioethanol is the focus on intensifying the process and increasing the concentration of water insoluble solids (WIS) throughout the process. However, increasing the WIS content is not without problems. For example, the viscosity of pretreated lignocellulosic materials is known to increase drastically with increasing WIS content. Further, at elevated viscosities, problems arise related to poor mixing of the material, such as poor distribution of the enzymes and/or difficulties with temperature and pH control, which results in possible yield reduction. Achieving good mixing is unfortunately not without cost, since the power requirements needed to operate the impeller at high viscosities can be substantial. This highly important scale-up problem can easily be overlooked. RESULTS: In this work, we monitor the impeller torque (and hence power input) in a stirred tank reactor throughout high solid enzymatic hydrolysis (< 20% WIS) of steam-pretreated Arundo donax and spruce. Two different process modes were evaluated, where either the impeller speed or the impeller power input was kept constant. Results from hydrolysis experiments at a fixed impeller speed of 10 rpm show that a very rapid decrease in impeller torque is experienced during hydrolysis of pretreated arundo (i.e. it loses its fiber network strength), whereas the fiber strength is retained for a longer time within the spruce material. This translates into a relatively low, rather WIS independent, energy input for arundo whereas the stirring power demand for spruce is substantially larger and quite WIS dependent. By operating the impeller at a constant power input (instead of a constant impeller speed) it is shown that power input greatly affects the glucose yield of pretreated spruce whereas the hydrolysis of arundo seems unaffected. CONCLUSIONS: The results clearly highlight the large differences between the arundo and spruce materials, both in terms of needed energy input, and glucose yields. The impact of power input on glucose yield is furthermore shown to vary significantly between the materials, with spruce being very affected while arundo is not. These findings emphasize the need for substrate specific process solutions, where a short pre-hydrolysis (or viscosity reduction) might be favorable for arundo whereas fed-batch might be a better solution for spruce. RESULTS: In this work, we monitor the impeller torque (and hence power input) in a stirred tank reactor throughout high solid enzymatic hydrolysis (< 20% WIS) of steam-pretreated Arundo donax and spruce. Results from hydrolysis experiments at a stirrer speed of 10 rpm show that a very rapid decrease in impeller torque is experienced during hydrolysis of pretreated arundo (i.e. it loses its fiber network strength), whereas the fiber strength is retained for a longer time within the spruce material. This translates into a relatively low, rather WIS independent, energy input for arundo whereas the stirring power demand for spruce is substantially larger and quite WIS dependent. By operating the impeller at a constant power input (instead of a constant impeller speed) it is shown that power input greatly affects the hydrolysis yield of pretreated spruce whereas the hydrolysis of arundo seems unaffected. CONCLUSIONS: The results clearly highlight the large differences between the two materials, both in terms of needed energy input, and hydrolysis yields. The impact of power input is furthermore shown to vary significantly between the materials, with spruce being very affected while arundo is not. These findings emphasize the need for substrate specific process solutions, where a short pre-hydrolysis (or viscosity reduction) might be favorable for arundo whereas fed-batch might be a better solution for spruce.     

Polysaccharide production: Experimental comparison of the performance of four mixing devices

This investigation was undertaken with the objective to compare experimentally the performance of four different mixing devices for the production of the polysaccharide pullulan with Aureobasidium pullulans (2552). Fermentations were performed using identical bioreactors with respectively an assembly of three Rushton turbines (RTB), a helical ribbon impeller (HR) and two different reciprocating plates (RPB1, RPB2). Each mixing vessel had identical geometry and working volume (18 L). These four fermentations were performed with an equal level of power input per unit volume (1000?W/m³) and gas flow rate (0.5 vvm, 9 L/min). For each system, the evolution of biomass, polysaccharide concentration, dissolved oxygen and agitation speed or frequency were recorded as a function of time along with the rheological properties of the culture broths. The type of mixing device used had a significant impact on the rate of biomass production and on polysaccharide physical properties. However, the rate of polysaccharide production appears to be less sensitive to the bioreactor design. The overall productivity, which represents the ability of micro-organisms to convert rapidly substrate into biomass and polysaccharide, was maximised using the RPB2 system. The quality of the synthesised polysaccharide, in terms of viscosity producing power, was highest using the HR system but the overall productivity was very low. Thus, the best compromise between productivity and product quality was achieved with the RPB2.     

Dependence of mycelial morphology on impeller type and agitation intensity

The influence of the agitation conditions on the morphology of Penicillium chrysogenum (freely dispersed and aggregated forms) was examined using radial (Rushton turbines and paddles), axial (pitched blades, propeller, and Prochem Maxflow T), and counterflow impellers (Intermig). Culture broth was taken from a continuous fermentation at steady state and was agitated for 30 min in an ungassed vessel of 1.4-L working volume. The power inputs per unit volume of liquid in the tank, P/V(L), ranged from 0.6 to 6 kW/m(3). Image analysis was used to measure mycelial morphology. To characterize the intensity of the damage caused by different impellers, the mean total hyphal length (freely dispersed form) and the mean projected area (all dispersed types, i.e., also including aggregates) were used. [In this study, breakage of aggregates was taken into account quantitatively for the first time.]At 1.4-L scale and a given P/V(L), changes in the morphology depended significantly on the impeller geometry. However, the morphological data (obtained with different geometries and various P/V(L)) could be correlated on the basis of equal tip speed and two other, less simple, mixing parameters. One is based on the specific energy dissipation rate in the impeller region, which is simply related to P/V(L) and particular impeller geometrical parameters. The other which is developed in this study is based on a combination of the specific energy dissipation rate in the impeller swept volume and the frequency of mycelial circulation through that volume. For convenience, the function arising from this concept is called the "energy dissipation/circulation" function.To test the broader validity of these correlations, scale-up experiments were carried out in mixing tanks of 1.4, 20, and 180 L using a Rushton turbine and broth from a fed-batch fermentation. The energy dissipation/circulation function was a reasonable correlating parameter for hyphal damage over this range of scales, whereas tip speed, P/V(L), and specific energy dissipation rate in the impeller region were poor. Two forms of the energy dissipation/circulation function were considered, one of which additionally allowed for the numbers of vortices behind the blades of each impeller type. Although both forms were successful at correlating the data for the standard impeller designs considered here, there was preliminary evidence that allowing for the vortices would be valuable. (c) 1996 John Wiley & Sons, Inc.     

Power input measurements in a production scale bioreactor

Summary Power input measurements are carried out in a production bioreactor with a liquid volume up to 25 m³. The results show that the cavity formation principle is applicable to reactors at this scale. It can also be observed that empirical correlations are not useful to predict gassed power input accurately. It is found that at gas flow rates for normal production conditions (N_Q =0.1), the gassed power input is about 30–40 % of the non gassed power input.Nomenclature C_p specific heat J/kgK - D impeller diameter m - D_b1 impeller blade diameter m - d baffle diameter m - Fr Froude number - - g gravitation m/s² - h impeller clearance m - H liquid height m - N stirrer speed s^-1 - N_p power number - - N_Q gas flow (aeration) number - - N_Q ^* critical gasflow number for 3 cavity formation - - P_o ungassed power consumption W - P_g gassed power consumption W - Q gas flow rate (273 K, 10⁵ N/m²) m³/s - Re Reynolds number - - T tankdiameter m temperature K - t time s - V liquid volume m³ - ^Vtip impeller tip speed m/s - V_s impeller correlated superficial gas flow rate m/s - W impeller blade width m - density kg/m³     

Characterization and application of a miniature 10 mL stirred-tank bioreactor, showing scale-down equivalence with a conventional 7 L reactor

The aim of this study was to characterize the engineering environment of an instrumented 10 mL miniature stirred-tank bioreactor and evaluate its potential as a scale-down device for microbial fermentation processes. Miniature bioreactors such as the one detailed in this work have been developed by several research groups and companies and seek to address the current bottleneck at the screening stage of bioprocess development. The miniature bioreactor was characterized in terms of overall volumetric oxygen transfer coefficient and mixing time over a wide range of impeller speeds. Power input to the miniature bioreactor was directly measured, and from this the power number of each impeller was calculated and specific power input estimated, allowing the performance of the miniature bioreactor to be directly compared with that of a conventional 7 L bioreactor. The capability of the miniature bioreactor to carry out microbial fermentations was also investigated. Replicate batch fermentations of Escherichia coli DH5alpha producing plasmid DNA were performed at equal specific power input, under fully aerobic and oxygen-limiting conditions. The results showed a high degree of equivalence between the two scales with regard to growth and product kinetics. This was underlined by the equal maximum specific growth rate and equal specific DNA product yield on biomass obtained at the two scales of operation, demonstrating the feasibility of scaling down to 10 mL on the basis of equivalent specific power input.     

Carbon monoxide mass transfer for syngas fermentation in a stirred tank reactor with dual impeller configurations

This study compares the power demand and gas-liquid volumetric mass transfer coefficient, kLa, in a stirred tank reactor (STR) (T = 0.211 m) using different impeller designs and schemes in a carbon monoxide-water system, which is applicable to synthesis gas (syngas) fermentation. Eleven different impeller schemes were tested over a range of operating conditions typically associated with the "after large cavity" region (ALC) of a Rushton-type turbine (D/T = 0.35). It is found that the dual Rushton-type impeller scheme exhibits the highest volumetric mass transfer rates for all operating conditions; however, it also displays the lowest mass transfer performance (defined as the volumetric mass transfer coefficient per unit power input) for all conditions due to its high power consumption. Dual impeller schemes with an axial flow impeller as the top impeller show improved mass transfer rates without dramatic increases in power draw. At high gas flow rates, dual impeller schemes with a lower concave impeller have kLa values similar to those of the Rushton-type dual impeller schemes but show improved mass transfer performance. It is believed that the mass transfer performance can be further enhanced for the bottom concave impeller schemes by operating at conditions beyond the ALC region defined for Rushton-type impellers because the concave impeller can handle higher gas flow rates prior to flooding.     

Mixing by impeller agitation in continuous flow systems containing polysaccharide solutions

System response data for step changes in input tracer concentration have been obtained for two different impeller agitated continuous flow mixing systems containing aqueous polysaccharide solutions. The vessel volumes were 1.6 and 10.9 liters. Polysaccharide concentration, dilution rate, and impeller speed were varied according to a plan devised using dimensional analysis and assuming that bulk motion is the predominant mass transport mechanism in the system. The data show that this is not true and that serious errors may occur if scale-up calculations are based on assuming that bulk motion predominates. Under the operating conditions used, perfect mixing was not observed.     

Selective suppression of bacterial contaminants by process conditions during lignocellulose based yeast fermentations

Background

When scaling up lignocellulose-based ethanol production, the desire to increase the final ethanol titer after fermentation can introduce problems. A high concentration of water-insoluble solids (WIS) is needed in the enzymatic hydrolysis step, resulting in increased viscosity, which can cause mass and heat transfer problems because of poor mixing of the material. In the present study, the effects of mixing on the enzymatic hydrolysis of steam-pretreated spruce were investigated using a stirred tank reactor operated with different impeller speeds and enzyme loadings. In addition, the results were related to the power input needed to operate the impeller at different speeds, taking into account the changes in rheology throughout the process.

Results

A marked difference in hydrolysis rate at different impeller speeds was found. For example, the conversion was twice as high after 48 hours at 500 rpm compared with 25 rpm. This difference remained throughout the 96 hours of hydrolysis. Substantial amounts of energy were required to achieve only minor increases in conversion during the later stages of the process.

Conclusions

Impeller speed strongly affected both the hydrolysis rate of the pretreated spruce and needed power input. Similar conversions could be obtained at different energy input by altering the mixing (that is, energy input), enzyme load and residence time, an important issue to consider when designing large-scale plants.     

Simulation of the effect of mixing,scale-up and pH-value regulation during glutamic acid fermentation

A mixing model is coupled with fermentation kinetics in order to simulate a fermentation as a function of mixing conditions and scale-up. The mixing model for a batch stirred tank with three stirrers consists of three regions, each of them characterized by an ideally mixed compartment around the stirrer and two macromixers, i.e. cascades of tank-in-series, describing the recirculation flow. The model contains four parameters — radial and axial circulation time, volume of the ideally mixed stirrer compartment and the number of tanks in each cascade. These values, determined by Mayr et al. in function of the operational conditions and scale-up, were choosen to simulate the fermentation of glutamic acid to show the pH-fluctuation at different control and scale conditions. By choosing optimal regulation properties, such as input flow rate and/or concentration of the base, regulation span, position of the pH-electrode and base input location, etc., fluctuations of the pH-value in the bio-reactor can be minimized. However, the negative effect of insufficient mixing conditions can be reduced only by an increasing number of the base input places. In large scale fermentors, the axial circulation time is rather high, about 5–10 times larger than the radial one. This might result in a large amplitude of the pH-fluctuation. As it is shown, using an input place for base in each stirrer region, the negative impact of the insufficient axial mixing on the fermentation can be diminished perfectly. In this case ammonia should be fed into the reactor as an aqueous solution.     

A mathematical model for agitation-induced fragmentation of Penicillium chrysogenum

The morphology of filamentous organisms in submerged cultures varies between the pelleted and the dispersed forms depending on the strain of organism and the culture conditions. The dispersed form consists of branched and unbranched hyphae (freely dispersed form) and clumps (filamentous material in aggregates). In agitated systems, the choice of impeller geometry as well as the total power input determines the mechanical forces that might affect the morphology of filamentous species (e.g. by fragmentation) with simultaneous effects on their growth and productivity. To find out more about fragmentation of Penicillium chrysogenum caused by mechanical forces of different impeller types and agitation intensities, a population balance model has been developed. The projected area measured by image analysis was used to characterise the morphology (size) of the mycelia. In the model, the kinetics of mycelial fragmentation were expressed by a breakage rate constant K, which was assumed to be only dependent on the agitation conditions. The fragmentation rate was considered to follow a first order process in size (area) which was based on assumptions made for the mechanism of mycelial break-up, and work reported in the literature. Previously published mean and distributional data from off-line fragmentation experiments in ungassed vessels of sizes from 1.4 to 180?l were used to validate the model. For the first time a model has been found that is capable of fitting changes in mycelial morphology caused by mechanical forces generated by different impellers at various power inputs and scales. Besides the mean projected areas of the mycelia, the model allowed simulations of the projected area distributions, and changes in those distributions because of the agitation. At the small scale (1.4?l), the breakage rate constant K could be correlated well with either impeller tip speed or the “energy dissipation/circulation function”, which is based on mycelial circulation through the impeller region. The simpler but commonly used power input per unit tank volume did not correlate K adequately. The scale up data showed that only the “energy dissipation/circulation function” correlated mycelial fragmentation well. The dependence of K on biomass concentration, and its detailed dependence (if any) on the fermentation conditions at sampling, which might indicate likely breakage mechanisms, remain to be elucidated.     

Scale-up analysis of geometrically dissimilar single-use bioreactors

Cell culture scale-up is a challenging task due to the simultaneous change of multiple hydrodynamic process characteristics and their different dependencies on the bioreactor size as well as variation in the requirements of individual cell lines. Conventionally, the volumetric power input is the most common parameter to select the impeller speed for scale-up, however, it is well reported that this approach fails when there are huge differences in bioreactor scales. In this study, different scale-up criteria are evaluated. At first, different hydrodynamic characteristics are assessed using computational fluid dynamics data for four single-use bioreactors, the Mobius® CellReady 3 L, the Xcellerex™ XDR-10, the Xcellerex™ XDR-200, and the Xcellerex™ XDR-2000. On the basis of this numerical data, several potential scale-up criteria such as volumetric power input, impeller tip speed, mixing time, maximum hydrodynamic stress, and average strain rate in the impeller zone are evaluated. Out of all these criteria, the latter is found to be most appropriate, and the successful scale-up from 3 to 10 L bioreactor and to 200 L bioreactor is confirmed with cell culture experiments using Chinese Hamster Ovary cell cultivation.     

A comprehensive comparison of mixing, mass transfer, Chinese hamster ovary cell growth, and antibody production using Rushton turbine and marine impellers

Large scale production of monoclonal antibodies has been accomplished using bioreactors with different length to diameter ratios, and diverse impeller and sparger designs. The differences in these physical attributes often result in dissimilar mass transfer, mechanical stresses due to turbulence and mixing inside the bioreactor that may lead to disparities in cell growth and antibody production. A rational analysis of impeller design parameters on cell growth, protein expression levels and subsequent antibody production is needed to understand such differences. The purpose of this study was to examine the impact of Rushton turbine and marine impeller designs on Chinese hamster ovary (CHO) cell growth and metabolism, and antibody production and quality. Experiments to evaluate mass transfer and mixing characteristics were conducted to determine if the nutrient requirements of the culture would be met. The analysis of mixing times indicated significant differences between marine and Rushton turbine impellers at the same power input per unit volume of liquid (P/V). However, no significant differences were observed between the two impellers at constant P/V with respect to oxygen and carbon dioxide mass transfer properties. Experiments were conducted with CHO cells to determine the impact of different flow patterns arising from the use of different impellers on cell growth, metabolism and antibody production. The analysis of cell culture data did not indicate any significant differences in any of the measured or calculated variables between marine and Rushton turbine impellers. More importantly, this study was able to demonstrate that the quality of the antibody was not altered with a change in the impeller geometry.     

Mass transfer and mixing rates in fermentation vessels

The effects on mass-transfer and overall mixing rates of varying impeller geometry and operating speed have been studied for flat-bladed turbines in laboratory fermentors, in aerated aqueous solutions, and in unaerated and aerated suspensions (1.6% w/v) of paper pulp. In the absence of suspended solid, oxygen absorption rates could be correlated directly with power input. In the pulp suspension, oxygen absorption at a given power input was influenced by impeller geometry and operating speed. The data for the three-phase system can be correlated by a dimensionless equation relating oxygen-transfer rates and mixing times to the geometrical and operating parameters of the impellers.     

Flow Pattern and Mixing in the Multi‐Turbine Agitated Bioreactor Filled with Poly(γ‐Glutamic Acid) Solution

We investigated the flow pattern and mixing behavior of a poly(γ‐glutamic acid) (γ‐PGA) solution in a bioreactor equipped with two Rushton turbines by simulation and experiment. Computational fluid dynamics (CFD) is used to solve the three‐dimensional hydrodynamics in the bioreactor and to obtain the flow patterns and tracer concentration at every point. The flow circulation patterns by inter‐impeller clearance and viscosity and their effects on overall mixing time were studied. Based on the results we can conclude that the impeller clearance should not be larger than 0.2 D for the efficient mixing under non‐aerated condition when the liquid viscosity is above 20 cp, which corresponds to concentrations of 20 g/L or above for γ‐PGA.     

Effect of Impeller Clearance on Flow Structure and Mixing in Bioreactor with Two‐Stage Impellers

The effect of impeller clearance on flow structure and mixing time was simulated using a commercial software package CFX 4.3 and was measured experimentally. The mixing time calculated by simulation exhibited good agreement with the experimental data. The trend of forming independent flow compartments by each impeller became stronger as the clearance between two impellers increased. The homogeneity in the bioreactor was affected mainly by flow exchange between the compartments by each impeller. The most efficient mixing occurred when the impeller clearance was in the range of 0.2–0.4 vessel diameter.     

A novel centrifugal impeller bioreactor. I. Fluid circulation, mixing, and liquid velocity profiles

A novel centrifugal impeller bioreactor for shear-sensitive biological systems was designed by installing a centrifugal-pumplike impeller in a stirred vessel. The fluid circulation, mixing, and liquid velocity profiles in the new bioreactor (5-L) were assessed as functions of the principal impeller designing and bioreactor operating parameters. The performances of the centrifugal impeller bioreactor were compared with those of a widely used cell-lift bioreactor. The newly developed bioreactor showed higher liquid lift capacity and shorter mixing time than the cell lift with comparable dimensions. Furthermore, the experiments of the liquid velocity profiles around an impeller region indicated that the centrifugal impeller bioreactor produced lower shear stress than the cell lift. This conclusion was also supported by evaluating the changes in size distributions of granulated agar particles that were sheared with those two types of impeller.     

Characterization of stirrers for screening studies of enzymatic biomass hydrolyses on a milliliter scale

The evaluation of mixing quality is an important factor for improving the geometry of stirred-tank reactors and impellers used in bioprocess engineering applications, such as the enzymatic hydrolysis of plant materials. Homogeneity depends on different factors, including the stirrer type and the reactor type (e.g., ratio of diameter/height, ratio of impeller tip diameter/reactor diameter) with or without baffles. This study compares two impellers for enzymatic hydrolysis of suspensions of biomass particles on a milliliter scale. Both impellers were derived from industrially relevant geometries, such as blade and grid stirrers, although the geometry of the second stirrer was slightly modified to an asymmetric shape. The stirrers were investigated with different stirrer–reactor configurations. This was done experimentally and with the aid of computational fluid dynamics. The flow field, mixing numbers, power characteristics and initial conversion rates of sugars were considered to compare the two stirrers. The simulated mixing numbers and power characteristics in baffled and unbaffled milliliter-scale reactors were found to be in good agreement with the measured mixing times and power consumption. The mixing numbers required to reach homogeneity were much higher for the symmetric impeller and remained at least twice as high as the mixing numbers required when using the asymmetric impeller. The highest initial sugar releases from milled corn stover suspensions were achieved with the asymmetric impeller shape. Regardless of the differences in the flow fields or mixing times, diverging enzymatic sugar releases could be confirmed for Newtonian media only.     

Polysaccharide concentration and molecular weight effects on the oxygen mass transfer in a reciprocating plate bioreactor

In most polysaccharide fermentations, the nature of the fermentation broth changes drastically with time and, as a result, the overall oxygen mass transfer coefficient (K(L)a) can vary by orders of magnitude. To obtain a better understanding of this phenomenon, an experimental program was devised to study the respective influence of molecular weight and concentration of dextran solutions on K(L)a. Experiments were conducted in a reciprocating plate bioreactor. This bioreactor uses a stack of perforated plates that is reciprocated axially in the column and it is therefore well suited for mixing viscous liquid broths and providing uniform overall mass transfer coefficients. The variation of K(L)a with the power input per unit volume and the superficial gas velocity were obtained for three ranges of molecular weights and five concentrations of dextran. In every medium, two regimes of operation were observed as a function of the power input per unit volume: a first regime, at low power inputs per unit volume where K(L)a remains constant until a threshold of power input is attained; and a second regime, which is characterized by a steep increase of K(L)a as a function of the power input per unit volume. The presence of dissolved biological macromolecules, not only because of their effect on the rheology of the medium but also because their effect on the gas-liquid interface, has a significant impact on K(L)a. It was found that, generally, small concentrations of polysaccharide favor oxygen mass transfer despite the increase in medium viscosity. However, the respective influence of polysaccharide concentration and molecular weight was different for the two regimes of operation. (c) 1996 John Wiley & Sons, Inc.     

Study of hydrodynamics in microcarrier culture spinner vessels: A particle tracking velocimetry approach

Three-dimensional particle tracking velocimetry (3-D PTV), a modern, quantitative, visualization tool, has been applied to the characterization of the flow field in the impeller region of cell culture reactor vessels. The experimental system used here is a 250-mL microcarrier spinner vessel. The studies were conducted at three different agitation rates, 90, 150, and 210 rpm, corresponding to healthy, mildly damaging, and severely damaging shear intensities, respectively. The flow can be classified into three regions: a predominantly tangential (azimuthal) flow generated by the impeller; a trailing vortex region coming off the impeller tip; and a converging flow region close to the center of the vessel. The latter two are the regions of highest velocity gradients. Energy dissipation rates due to mean velocity gradients were also calculated to characterize the impeller stream. Local specific energy dissipation rates > 10,000 erg/(cm(3)sec) . have been measured. It is proposed that the critical regions for microcarrier culture damage due to impeller hydrodynamics are the trailing vortex region and the high energy converging flow region. Graphical representation of the mean velocity flow fields and the distribution of energy dissipation rates in the impeller region are also presented here. The merits of using the dissipation function (measure of specific energy dissipation rate) as a possible scale-up parameter are also discussed. (c) 1996 John Wiley & Sons, Inc.