The Research Programme

FMP is an industrial research consortium dedicated to fluid mixing, undertaking projects decided by its members at steering group meetings. FMP develops reliable and industrially relevant design and scale up (or scale down) rules for mixing processes in stirred or jet mixed vessels, and bridges the gap between academic research and industrial needs.

Typical causes for gaps in current mixing knowledge from open literature include:

• Insufficient information for reliable scale-up
• Lack of published raw data
• Differences in experimental methods.

Subject coverage of FMP research

• Comparative mixing time testing of agitators for a range of reactor scales and process materials
• Gas/liquid mixing and mass transfer
• Solid suspension and solid/liquid mixing
• Studies of fluid flow using Laser Doppler Velocimetry and Computational Fluid Dynamics
• Fluid flow and mass transfer in bioreactors and fermentation vessels - gas/liquid mixing
• Droplet dispersion in immiscible liquid systems
• Liquid jet mixing and jet solids suspension
• Software and Design Guides

Project Areas
The project is organised into 7 areas:

• Liquid Blending
• Software and Design Guides
• Solid-Liquid Mixing
• Gas-Liquid Mixing
• Jet Mixing
• Liquid-Liquid Dispersion
• Flow Modelling: CFD and LDA

All of these areas draw on our extensive range of test facilities.

Area 1: Liquid Blending
Blending of miscible liquids. Measurements of mixing time and power consumption.

Mixing time is often used as a criterion of mixer performance. A knowledge of comparative mixing times is needed when selecting a mixer to be used for a blending process, or in situations where the process result is dependent on the mixing rate (e.g. mixing-limited chemical reactions, addition of catalysts, inhibitors or initiators).

In published literature, reliable information based on a consistent body of experimental data, particularly from large scale, is lacking. Reliable assessment of the performance of different agitators, internals (e.g. coils, non-standard baffle arrangement), the effects of viscosity and rheology on mixing time is therefore not possible.

FMP has tested 12 different impellers and 13 internal combinations in tanks from 0.3m to 2.7m diameter using Newtonian liquids over a Reynolds number range from 100 to 2,000,000. Further work has been undertaken on the blending of shear thinning liquids, where practically no reliable information exists in the open literature. Experimental work has already been carried out for four single impellers in tanks from 0.3 to 1.83m diameter for a number of different fluids. As a result of this research programme, reliable correlations to predict mixing time for Newtonian and shear thinning liquids over the turbulent and transitional have been developed by FMP for use by industrial members.

Current activities in this Area of FMP have focused on the blending of liquids with different viscosities. Experiments have been undertaken in vessels (in up to 1.8m diameter vessels) much larger than those used by previous researchers. Three types of impellers were studied over the turbulent regime and studies in the transitional regime are underway. Operational conditions over which correlations for similar property liquids cannot be used have been identified; recommendations have been made to improve blending under these conditions.

Future work will cover CSTRs in collaboration with Area 7, Flow Modelling/CFD.

Area 2: Software and Design Guides
The purpose of this area is to disseminate FMP mixing knowledge. Two formats are used: paper Design Guides and software equivalents.

Design Guides on paper or in software are available for the following applications:

Stirred Vessels

• Shaft design
• Impeller power numbers
• Blending miscible liquids (Newtonian and shear thinning)
• Solid-liquid Mixing
• Gas-liquid Mixing

Jet Mixing

• Blending of miscible liquids

Area 3: Solid-Liquid Mixing
Suspension, distribution and draw down of solids in stirred tanks. Measurements of the minimum speed for solid suspension (N_JS) and draw down (N_JD), local solids concentration and power consumption.

Most chemical products involve solid-liquid processing requiring either off-bottom suspension or draw down from the surface, some also require achieving a certain degree of homogeneity. However, there is little published information on design criteria, solids homogeneity, draw down of solids, scale-up, effects of tank internals and influence of the physical properties of solid and liquid phase.

FMP has developed instrumentation for the measurement of N_JS and solids concentration in large-scale tanks. The range of operational conditions in terms of scale (tanks from 0.3 to 2.7m diameter), impeller type (including hydrofoils and multiple impellers) and clearance (T/3 to T/8), particle properties and liquid phase viscosity is much more comprehensive than the work reported in the literature and reflects common industrial practice for solid-liquid mixing.

Many correlations are available in the open literature for predicting 'just suspension speed' of the impeller. The corresponding power consumption obtained from these can differ by orders of magnitude. Based on the extensive research undertaken, FMP has identified a reliable correlation and its limitations to predict just suspension speed in the turbulent regime. Present work on solid suspension investigates the effect of liquid phase viscosity.

Current work on the draw down of solids investigates the effect of impeller type and pumping mode, submergence/off-bottom clearance, number of baffles, liquid height. Studies on the effect of particle properties are starting.

A reliable scale up rule for solids distribution has been developed for given particle properties. This takes into account the effect of impeller type and tank-to-impeller diameter ratio for a given solid type. Currently, the effect of particle properties is investigated. Experimental studies on solid distribution have successfully been coupled with the activities of Area 7, Flow Modelling/CFD.

FMP continues to expand its database on performance characteristics of impellers in this Area.

Area 4: Gas-Liquid Mixing
Gas dispersion in stirred tanks, measurements of gas hold-up, gassed and ungassed power consumption, gas-liquid mass transfer coefficients, liquid phase mixing times, determination of the hydrodynamic state.

Gas-liquid mixing is involved in many high-value applications which are critical to the effective running of process plant: hydrogenation, chlorination, organic oxidation, fermentation. Work in the open literature is often fragmented, of limited scope and performed with limited resources.

FMP has carried out experimental work in standard tanks (liquid depth equal to tank diameter, single impellers) and tall tanks (liquid depth approximately three times the tank diameter, multiple impellers). In standard tanks FMP has tested Rushton turbines, pitched blade turbines, concave bladed turbines and hydrofoils at scales up to 2.7m tank diameter. In tall tanks, three impeller combinations have been tested in 0.6m and 0.95m tanks. This work is aimed at investigating segregation in multiple impeller systems, hence improving process performance. Experiments have also been undertaken using viscous Newtonian and shear thinning liquids.

Gas-liquid hydrodynamics have been investigated using a number of impellers. Reliable correlations to predict gas hold up and k_La have been developed. These encompass a large range of conditions: single and multiple impellers in standard and tall vessels, Newtonian and shear thinning liquids. Measurements of liquid phase mixing time with single and multiple impellers have demonstrated the validity of the mixing time correlation for gassed systems.

The above programme is currently being expanded to cover a larger range of conditions, very high gas flow rates, other impeller types, etc. An experimental technique is also being developed to make measurements of the gas phase residence time distribution in future work.

Area 5: Jet Mixing
Blending liquids and solids suspension using a liquid jet. Measurements of mixing time and minimum jet velocity for solids suspension (V_JS) respectively.

Liquid jets offer an economic means of agitating the contents of a vessel where the installation of a mechanical agitator is impractical, e.g. limited headroom or vessel geometry constraint, or due to the product properties (highly corrosive acids). Capital costs are lower, particularly if a pump is already installed. Maintenance costs can also be lower since there are no moving parts in the vessel due to the pump being remote from the tank.

In designing a jet blending system, a knowledge of mixing times is needed together with minimum design criteria for blending. FMP has measured mixing times in tanks from 0.6 to 3.8m diameter. A large number of variations in geometry have been tested, including vertical and angled jets, multiple jets and vertical and horizontal cylindrical tanks. Current work continues on the measurement of mixing times in viscous liquids.

To obtain reliable design procedures of a jet mixed system to suspend solids off the vessel base, FMP has carried out tests to determine just suspension velocity. V_JS in 0.3, 0.6 and 1.1m tanks with a flat or a dished base. The effects of tank and jet geometry and solids physical properties have been investigated. A reliable scale-up rule has been developed for solid suspension with jets and the preparation of Design Guides are underway.

Knowledge Gaps in Open Literature

Insufficient information for reliable scale-up
Academic research is conducted at small scale, often in tanks of about 0.3m in diameter. Extrapolation of results to industrial scale processes can therefore be unreliable and lead to process failure. Research at large scales is expensive and time consuming. Many of the experimental techniques in common use at small scales cannot be applied to large scale tanks. FMP has large scale tanks available for research (1.8 and 2.7m diameter) and is able to test correlations and mixer performance. Experimental techniques have been developed suitable for use at large scales. Industrial collaboration allows FMP to carry out work which would be prohibitively expensive for a single company to support.

Correlations in the open literature often agree at the small scale, but can predict different results at the large scale for the same process. Disagreement at the large scale arises because different correlations are based on different theoretical or semi-theoretical models and the dominant mechanism can change as the scale of operation is increased. FMP is able to test different correlations in its range of large scale tanks, and confirm which give the most reliable scale-up.

Lack of Published Raw Data
Researchers usually publish correlations, not their raw data, and the data on your process may be different, affecting the concluding correlation. To address this issue, FMP has always provided its members with the raw data as well as the correlations. This has allowed the member companies to incorporate their results in the data set and/or make comparisons.

Differences in Experimental Methods
Different researchers often use different experimental methods, so that the results from the different methods often cannot be compared directly. Problems can arise when information from several sources is needed for a process design, or where different workers' results appear to disagree with each other. FMP has always used consistent experimental methods at all scales, for all conditions and reported to the members the details of these methods.

Droplet dispersion experiments

The Project Areas continued

Area 6: Liquid-Liquid Dispersion
Drop deformation and break up in stirred tanks. Measurements of droplet size and drop size distribution, minimum speed for dispersion and power consumption.

Industrial applications that involve dispersing immiscible liquids, either aim to achieve mass transfer (e.g. pharmaceuticals) or generate a liquid-liquid dispersion or emulsion (e.g. food industry, cosmetics). In both cases, it is crucially important to be able to predict the drop sizes that will be formed with different impellers at different scales of operation to achieve product consistency and save energy/time during the manufacturing process. Current knowledge does not allow design from first principles. There is limited information on the effect of scale, comparative performance of impellers, effect of physical properties of the liquids and process conditions. Lack of reliable data is another major limitation.

The design process requires extrapolating rather than interpolating. Mechanistic models are more reliable for this purpose. On the other hand, it is extremely complicated to investigate drop break-up and coalescence at the same time. Such an approach leads to empirical correlations rather than mechanistic ones. It is not possible to obtain physical insight from the experimental data, discriminate among different mechanisms and develop models. Therefore, the research strategy that has been adopted by FMP is to suppress coalescence and concentrate on identifying the mechanisms of drop breakup The ultimate aim is to develop industrially relevant design and scale-up rules for processes that involve drop breakup

FMP has implemented two techniques to obtain drop sizes and drop size distribution: macrophotography and more recently image capture by a video camera. Experimental work has been carried out in vessel sizes from 0.17 to 0.58m in diameter to identify the industrially relevant dominant breakup mechanism. Traditional and novel design impellers have been used: six bladed disc turbines and pitched blade turbines; low solidity ratio hydrofoils, such as Lightnin A310 and Chemineer HE-3. Additionally saw tooth impellers, which are commonly used in numerous dispersion applications in industry, have been studied. Little information exists on these impellers in open literature. Impeller-to-tank diameter ratio was between 1/3 to 1/2.

The research programme that has been undertaken by FMP for over 10 years, and the findings from this programme are therefore unique in terms of:

• the systematic approach taken to investigate the breakup phenomenon
• care taken to ensure that the results are reproducible
• the range of vessel sizes used: results in the open literature are mainly limited to 0.15-0.30m in diameter where the dominant breakup mechanism can be different from that in an industrial scale tank. FMP has used vessels of up to 0.58m diameter
• different types of impellers used to assess their relative performance and also to obtain a model that can account for the differences in impeller design and size. Published results from other researchers are mainly limited to Rushton turbines and, to some extent, pitched blade turbines
• care taken to ensure that there is good interaction between the findings of other research areas of FMP research programme (link with turbulent flow fields and impeller pumping capacity obtained from LDA results)

FMP has confirmed the effects of different drop breakage mechanisms. Based on this knowledge, an industrially relevant scale-up rule has been developed relating drop breakage to flow characteristics in the vessel.

Future work will investigate higher dispersed phase concentrations.

Area 7: Flow Modelling: CFD and LDA
Measurement of flows and turbulence in mixing tanks using the laser Doppler anemometry. Computational modelling of flows in single and multiphase systems.

A knowledge of flows and turbulence in mixing tanks is essential to develop optimised processes by exploiting the physics of mixing processes. Development of predictive CFD (Computational Fluid Dynamics) tools and validation with reliable experimental data are complementary activities.

Computational Modelling
Computational modelling can be used for process development based on an understanding of the fluid mechanics involved, allowing a move away from statistical correlations and a "black-box" approach. Computational modelling can be cheaper and more rapid than physical modelling or pilot plant work, can be used for hazardous processes and allows the investigation of a wide range of process parameters. However the computational methods must be tested and validated against experimental data before the predicted results can be used with confidence.

FMP has CFX, Fluent, FIDAP and Star-CD codes available for use. Comparison of CFD predictions with experimental data, obtained using laser Doppler anemometry (LDA) for single phase systems or results from Area 3 of FMP for solid-liquid systems, has added confidence in the results.

Simulations of single phase flows for jet mixed systems as well as stirred vessels have successfully been undertaken using the momentum source model and the sliding mesh. Different approaches (k-epsilon, RNG k-epsilon, differential stress model, algebraic stress model) have been investigated. Predictions of mixing time are underway.

Successful predictions of solid concentration distribution in a stirred vessel have also been obtained for a range of conditions (different impeller types, vessel sizes). Other test cases are considered to increase the reliability of the model. This tool can be used to make predictions of solid concentration for a given range of conditions which can otherwise be very time consuming.

LDA Measurements
LDA measurements were taken using six bladed disc turbines, pitched blade turbines and hydrofoil impellers in a 0.3m tank with either a flat or a dished base for different off-bottom clearances. Complementary measurements have also been made in a 1.8m tank to test scale-up criteria.

Apart from generating validation data for CFD, flow and turbulence measurements from the LDA have provided useful information to other areas of FMP (liquid-liquid dispersions, liquid blending and solid-liquid mixing). These have been used to interpret experimental data or to correlate them with turbulence or flow related parameters.

Experiments with viscous liquids are underway. Future work will also continue to provide data for validation of computational models and to support other FMP areas.