Download CFD Autoclave Circuit Design A PDF

TitleCFD Autoclave Circuit Design A
TagsComputational Fluid Dynamics Fluid Dynamics Momentum Suspension (Chemistry)
File Size756.0 KB
Total Pages15
Document Text Contents
Page 1

2003 International Symposium on Hydrometallurgy Edited by C. Young
TMS (The Minerals, Metals & Materials Society), 2003


Lanre Oshinowo, Lowy Gunnewiek and Kevin Fraser

Hatch Associates Ltd.
2800 Speakman Drive

Mississauga, Ontario, CANADA L5K 2R7

The trend in process engineering is to design compact, more efficient processes, there is the
paramount requirement to get the job done right the first time going from the drawing board to
full-scale commercial operation. To accomplish this goal, there is also the need for tools
beyond the traditional engineering toolkit to evaluate designs through virtual prototyping,
thereby reducing the risks associated with making design decisions. One of the most important
tools that has recently come to the forefront of process design and development is
Computational Fluid Dynamics (CFD). CFD has been used at HATCH to address the key
process parameters that drive the design of hydrometallurgical unit operations. This paper will
detail the role of CFD at HATCH in achieving a superior level of confidence in the process
design of autoclave circuits. Specifically, the optimum application of multiphase modeling
including hydrodynamic, heat and mass transfer to hydrometallurgy operations, the impact of
non-Newtonian slurry rheology on autoclave performance, and the challenges of optimizing the
mixing of key reactants into slurries in autoclave reactors, is discussed.

Modern metallurgical operations require process intensification and higher efficiencies while
striving to protect capital investment, market position and return on investment capital. With
process enhancements and the application of new technology that pushes the envelope on
materials of construction, new plants become a very expensive capital and risky investment.
Hence, the efficiency and optimization of process design is paramount to achieving targets for
the unit operations. And requires an in-depth evaluation and understanding of the process. The
traditional approach of employing a combination of process experience, simplified analytical
and empirical models with trial and error are no longer acceptable. A multidisciplinary group at
Hatch is experienced in the design, construction and operation of modern hydrometallurgical
facilities and autoclave circuit design. CFD is one of the indispensable tools utilized by Hatch
and has been integrated into the autoclave circuit design practice and is used to manage the
technological risk of unit operation design. CFD is the science of predicting fluid flow, heat
transfer, mass transfer, chemical reactions, phase change, multiphase flow, and related
phenomena by solving the mathematical equations that govern these processes. The results of
CFD analyses are relevant engineering data used in conceptual studies of new designs, detailed
process development, troubleshooting and redesign. Using CFD, many variations in design can
be made, virtually, before deciding on an optimal configuration and committing to building a
physical prototype. The commercial CFD software, FLUENT (1), is used at Hatch and for the
work presented in this paper.

Hydrometallurgical process operations involve the transport and conversion of an ore through
various unit operations with water as the primary phase. When autoclaves are used in
hydrometallurgical processes, the unit operations typically operate at temperature and pressure
with corresponding chemical reactions. Unit operations in an autoclave circuit include mixers,
settlers, clarifiers, hydrocyclones, off-gas systems, heat exchangers, flash vessels, and the
autoclave(s). It is typically difficult to establish similarity between the commercial and lab scale

Page 2

for the important dimensionless groups when testing on lab or pilot scale equipment. This
makes CFD one of the only means to evaluate the performance of industrial-scale designs.

The fluid flow phenomena in the autoclave circuit and in hydrometallurgical applications are
complex due to the presence of multiple phases. The complexity of modeling multiphase
systems is one of the reasons that CFD is not as widely deployed in the chemical and
metallurgical process industry as it is in the aerospace and automotive industries. However,
faster computers, better numerical models of more complex physics and more user-friendly
commercial CFD codes has now made the challenging applications in the hydrometallurgical
process more accessible than in the past (2). A survey of metallurgy literature on the subject of
CFD reflects the rapidly growing trend with an order of magnitude increase over the past 10
years (See Figure 1). Two databases were used as a basis for the survey: The Engineering
Compendex®, a comprehensive interdisciplinary engineering database referencing 5,000
engineering journals and conference materials and; the Metadex® index is a source for
publications in the field of metallurgy and material sciences. The growth in usage is indicative
of the utility of CFD as a tool for enabling engineers to design and troubleshoot systems when
traditional correlations and rules-of-thumb are not.

This paper will address the application of CFD to unit operations in the autoclave circuit. This
paper is not fully comprehensive since many applications are yet to be addressed. However, it
does provide an assessment of the application of CFD in the field of hydrometallurgy, in
particular, the autoclave circuit. First, a description of the modeling methodology as pertaining
to the multiphase flows will be discussed followed by the presentation and discussion of
different applications.
























Engineering Compendex (
Metadex (

Figure 1. Number of publications on CFD on the Engineering Compendex and

Metadex scientific publication databases.

Modeling Multiphase Flow using CFD
Modelling multiphase flows using CFD is complicated by the physics of the phenomena that
makes the solution of said problems to be typically computationally expensive. Most
multiphase models use empirical or semi-empirical descriptions of phase momentum, heat and
mass interactions. Therefore, validation and verification of the models is required for each

There are a number of models that are used to model multiphase flow: Lagrangian-Eulerian
model, drift flux or slip mixture models, Eulerian and Eulerian Granular models. The
applicability, and therefore, the choice of these models, is typically dependent on a knowledge
of the volume concentration of the secondary solid phase in the primary fluid phase with the
exception of the Eulerian models. The Lagrangian Eulerian model solves the equation of
motion for the discrete particle trajectories. The coupling between the phases through drag

Page 7

suggested that the scale-up exponent is also affected by scale itself, although to a much lesser
degree than it is affected by particle settling velocity. Furthermore, it was observed in
experiments that the scale-up exponents for just-suspended speed and to obtain the same
relative cloud height were actually different. Therefore, the scale-up methods used to predict
the conditions for suspension are not necessarily suited for predicting the solids distribution
uniformity in the vessel, which is of principal importance for process performance and is the
main parameter predicted by CFD. Another issue that destabilizes the foundation of the
Zwietering-type correlations used by engineers, including agitator vendors, is that the
correlations are based on solid material types, such as glass or sand, with much lower specific
gravities of the materials typically found in the hydrometallurgical process industry.

The present case is an investigation into an existing design and the planned modifications for
the agitation system in an autoclave reactor. The autoclave had experienced operational
difficulties due to the possibility of incomplete suspension. The d80 of the solids was 150 µm
and was 30wt% or 8.5vol% of the reactor contents. The particle size distribution in the feed
slurry, with a solids specific gravity of approximately 5, was higher after process start-up than
the original design values and the large size fractions were not being suspended with the
existing agitation system (impellers, shaft and drive motor). However, complete suspension was
a design requirement for the autoclave, and process performance would be compromised with
poor suspension. A lab study by the agitator vendor was performed to evaluate the performance
of the impellers. The lab tests involved making qualitative visual observations of the solids on
the bottom of the vessel. CFD was used to model the scale-down and scale-up configurations
using the Eulerian Granular model in FLUENT 6 and grid refinement in the impeller region and
at the walls. The grid sizes are in the range of 4,000 to 8,000 quadrilateral cells depending on
the level of refinement in the regions of interest.

Existing Autoclave – Commercial Scale
The existing configuration has an impeller/tank-diameter (D/T) ratio of 0.4, and off-bottom
clearance/tank-diameter (C/T) ratio of 0.36. Figure 6(a) shows the predicted liquid flow
distribution in the existing commercial-scale reactor illustrating the axial pumping
characteristics of the dual down-pumping hydrofoil impellers. Adding solids to the calculations
severely disrupts the single flow loop on either side of the impellers as shown in Figure 6 (b).
Two distinct flow loops are observed; one for each impeller. Figure 6(c) shows the distribution
of the solids and the complete lack of suspension at the bottom of the autoclave verifying the
operational performance of the autoclave.

(a) Velocity Vectors (Liquid only;
No solids in the autoclave)

(b) Velocity Vectors (Liquid and
solids in the autoclave)

(c) Distribution of solids volume

Figure 6. The flow and solids distribution in the commercial-scale autoclave reactor – design

Page 8

Lab Scale – Existing Configuration
The lab test rig was a 1/8th scale model of the commercial scale autoclave (diameter, liquid
height and impeller diameters). The shaft speed was scaled based on power per unit volume.
Figure 7(a) shows the flow distribution with the liquid and solids in the autoclave. The double
flow loop is observed at the lab scale as well. Figure 7(b) shows the distribution of the solids in
the lab-scale autoclave. The suspension of the solids is slightly better than in the commercial
scale though there is not off-bottom suspension. The lab tests verified that the solids were not
off-bottom. This demonstrates that the level of agitation is insufficient.

(a) Velocity Vectors (liquid and solids) (b) Distribution of solids volume fraction (-)

Figure 7. The flow and solids distribution in the lab scale autoclave reactor – design

Lab Scale – Proposed Configuration
Based on the suspension achieved with the level of agitation in the existing configuration, the
proposed improvement to the autoclave agitation was to increase the power per unit volume to
the autoclave. The bottom impeller was changed to a high-solidity hydrofoil and the impeller
diameter was increased to a D/T of 0.5. Figure 8(a) shows the flow distribution with the liquid
and solids with the proposed mixer configuration in the lab-scale autoclave. The double flow
loop is not completely eliminated. Figure 8 (b) shows an improved distribution of the solids in
the lab-scale autoclave. However, there is not off-bottom suspension. The lab tests verified that
the solids were not off-bottom. Further lab tests were run to show that lowering the impellers
and increasing the shaft speed could achieve complete off-bottom suspension. However, these
additional modifications would require a much larger motor (4 times the size of the existing
motor) and a larger shaft seal to handle the larger shaft diameter and length.

Commercial Scale-Up
The lab-scale proposed configuration was scaled-up at constant power per unit volume. The
actual process conditions are modelled so that the liquid phase properties of density and
viscosity at operating temperature and pressure are used in the analysis. Figure 9(a) shows the
flow pattern in the commercial-scale autoclave with the proposed mixer configuration. The
proposed mixer configuration is under-sized and unable to establish the single- loop flow
pattern needed to distribute the solids throughout the autoclave. Figure 9 (b) shows the resulting
distribution of solids. The suspension in the commercial-scale is worse than the lab-scale
indicating that the vendor-proposed changes will not work in the field and the scale-up
methodology used is flawed. Further modification to the mixer configuration, including
lowering the off-bottom clearance and increasing the shaft speed, should be considered to
improve the distribution of the solids suspension in the vertical autoclave.

Page 14

Future work will predict splash tower performance using the Eulerian multiphase model by
directly integrating the interfacial area available for heat and mass transfer. The results of a 2D,
“cold” flow, simulation is shown in Figure 16. The counter-current flow of steam and slurry is
shown. The variation in the slurry velocity in the film on the splash plate leads to a variation in
the trajectory of the slurry in the film impacting the subsequent plate over a wide area. Overall,
the CFD modeling of the splash towers has produced better designed units for contact between
the slurry and the steam and have reduced the risk by verifying the splash tower hydrodynamic
capacity and heat transfer performance.

Figure 16. Slurry volume fraction and steam pathlines in a splash tower.

The application of CFD to modelling unit operations in the autoclave circuit to improve and
enhance process design is a reality. Recent advances in the capabilities of commercial CFD
software, in particular FLUENT, has enabled engineers at HATCH to understand the
performance of the design and perform pre-construction optimization based on the results of
CFD analysis.

1 Fluent Inc., Lebanon, NH, USA. Available at

2 A. Bakker, A. Haidari and L. Oshinowo, “Realize Greater Benefits from CFD”, Chemical
Eng. Progress, March 2001, pg.45-53, 2001,

3. J.C. Godfrey, Z.M. Zhu, “Measurement of particle-liquid profiles in agitated tanks”, AIChE
Symposium Series. 299, (1994), 181-185.

Page 15

4 Maruyama, T. and Sato, H., “Liquid fluidization in conical vessels”, The Chemical
Engineering Journal, 46, (1991) 15-21

5 T.N. Zwietering, “Suspending solid particles in liquid by agitators,” Chemical Engineering, 8
(1958), 244-253.

6 R. Corpstein, J.B. Fasano and K.J. Myers, “The high-efficiency road to liquid-solid
agitation,” Chemical Engineering, 10 (1994), 138-144.

7 S. Sideman and D. Moalem-Maron, “Direct Contact Condensation,” Advances in Heat
Transfer, Volume 15, Academic Press, (1982), 227-281.

8 A.P. Solodov, “Calculation Models of Heat Transfer with Contact-Type Condensation,”
Teploenergetika, v. 37, n. 10, (1990) 12-16.

9 F. Kreith and R.F. Boehm (Eds.), “Direct Contact Heat Transfer.” (1988).

10 J.R., Fair, “Designing Direct-Contact Coolers/Condensers,” Chemical Engineering, June 12,
(1972), 91- 100.

Similer Documents