Download as pdf or txt
Download as pdf or txt
You are on page 1of 40

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257246810

Progress in an Industrial Application of Fluidized Beds: Advances in the Sand


Core Making Process

Chapter · January 2002


DOI: 10.13140/RG.2.1.4908.3760

CITATIONS READS

0 289

3 authors, including:

Sayavur I. Bakhtiyarov Dennis Siginer


New Mexico Institute of Mining and Technology University of Santiago, Chile
158 PUBLICATIONS 635 CITATIONS 320 PUBLICATIONS 1,297 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Existence and structure of steady solutions of the Bénard problem with shear-rate dependent viscosity fluids in a two dimensional quadrangular cavity View project

Exact Solution for Forced Convection Gaseous Slip Flow in Corrugated Microtubes View project

All content following this page was uploaded by Dennis Siginer on 21 May 2016.

The user has requested enhancement of the downloaded file.


TRANSPORT PROCESSES
/IN BUBBLES, DROPS,
AND PARTICLES

SECOND EDITION

D.De Kee
Tulane University
Department of Chemical Engineering
New Orleans, LA

R.P. Chhabra
Indian Institute of Technology
Department of Chemical Engineering
Kanpur, India

TAYLOR & FRANCIS


New York • London
Denise T. Schanck, Vice President
Robe1t H. Bedford, Editor
Brandy Mui, Production Editor
Catherine M. Caputo, Assistant Editor
Thomas Hastings, Marketing Manager
Mariluz Segana, Marketing Assistant

Published in 2002 by
Taylor & Francis
29 West 35th Street
New York, NY l 000 l

Published in Great Britain by


Taylor & Francis
11 New Fetter Lane
London EC4P 4EE

Copyright© 2002 by Taylor & Francis

Printed in the United States of America on acid-free paper.

All rights reserved. No part of this book may be reprinted or reproduced or utilized in
any form or by any electronic, mechanical, or other means, now known or hereafter
invented, including photocopying and recording, or in any information storage or retrieval
system, without permission in writing from the publisher.

Library of Congress Cataloging-in-Publication Data


Transport processes in bubbles, drops, and particles/[edited by)
by D. De Kee & R.P. Chhabra.-2nd ed.
p. cm.
Includes index.
ISBN l-56032-906-8 (alk. paper)
I. Fluid dynamics.
2. Transport theory. I. De Kee, D. (Daniel) II. Chhabra, R.P.

QA9 l I .T68 2002


530. I '8-dc2 l 2001027964
Contents vii

Interphase Heat Transfer 178


Illustrative Example 179
Nomenclature 180
References 181

8. Progress in an Industrial Application of Fluidized Beds:


Advances in the Sand Coremaking Process 187
S.I. Bakhtiyarov, R.A. Ovetfelt and D.A. Siginer
Introduction 187
CFD Modeling of Sand Core-Shooting Process 188
Experimental Apparatus and Procedure 189
Results and Discussion 196
Conclusion 221
Acknowledgments 221
References 222
Further Readings 222

9. Sedimentation and Fluidization of Solid Particles


in Liquids 223
R. Di Felice
Introduction 223
Fluidization and Sedimentation of Single-Species Solids 224
Fluidization and Sedimentation of Binary-Species Solids 236
Acknowledgment 242
Notation 242
References 243
Further Readings 245

10. Determination of the Constitutive Relationship for Filter


Cakes in Cake Filtration Using the Analogy between
Filtration and Diffusion 246
B. V. Ramarao, C. Tien, and C.N. Satyadev
Introduction 246
Generalized Diffusivity and its Definition 248
Initial and Boundary Conditions 252
Filtration Viewed as a Diffusion Process 253
Prediction of Constant Pressure Filtration Characteristics 256
Inversion of Filtrate Flux Data 265
Conclusion and Discussion 270
Symbols 271
References 272
Further Readings 273
CHAPTER 8
Progress in an Industrial
Application of Fluidized Beds:
Advances in the Sand
Coremaking Process
S.I. Bakhtiyarov, 1 R.A. Overfelt,1 and D.A. Siginer2
1
Auburn University, Department of Mechanical Engineering, Auburn, AL.
2
Wichita State University, College of Engineering, Wichita, KS.

Introduction

The flow of liquid droplets and solid particles in gas is important to numerous mate-
rial processing and manufacturing techniques. The solid-gas flow is a very complex
fluid mechanics problem. There are many experimental and theoretical approaches
for modeling gas-solid fluidized beds. The work related to achieving an understand-
ing of the hydrodynamics of aggregative fluidization (gas-solid two-phase flow),
with special emphasis on the rheological aspects of the problem, was considered re-
cently in our review paper (Bakhtiyarov and Overfelt 1999). The metal casting in-
dustry is one of the areas where product quality directly relates to the fluidization of
solid particles and liquid droplets. Spray casting, mixing, and tempering of foundry
molding sand, filling of a molding box with sand in lost foam and vacuum castings,
blast cleaning, and sand coremaking in a cold box process are just a few examples
where proper fluidization of solid particles is a very important factor in process
performance.
In the effort to optimize the structure and properties of cast metals, a variety of
casting techniques have evolved over the past few decades. This particular process-
ing technique involves curing at room temperature of a resin-bonded sand core ac-
celerated by a gas catalyst passed through the sand-binder mixture. In principle, the
process has the typical relatively fast sand-blowing, gassing, purging, and stripping
cycles, which are mainly affected by sand properties (density, grain size and distrib-
ution, shape, moisture content, close-packed volume fraction, and so on), type and
amount of binders (ratios for binders systems of two or more parts), tolling design

187
188 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

(blow tubes, sand magazine, input and exhaust piping and manifolds, vents, exhaust
ports, core box, and so on), and gassing and purging performances (gas type, pres-
sure, time, temperature, flow rate, and so on).
Recently, we investigated experimentally the flow dynamics of a phenolic-
urethane-amine process in a core box of an inverted U configuration, using a com-
mercial LAEMPE LI Core Shooter (Bakhtiyarov and Overfelt, 1998a). The effects
of reduction in the area of the vents were evaluated by manual reduction of the active
vent areas. In addition, experimental data on the pressure decrease and friction factor
for the vents were obtained as a function of the amounts of sand deposited in the
vent area. The experimental data were compared with the predictions of the Blake-
Kozeny and Ergun porous flow equations, and good agreement was found.
An experimental study and a numerical analysis of high-pressure injection of
sand air into sand core molds were performed by Snider and co-authors (1999). The
high-speed photography showed complex filling patterns which are functions of the
core shape, the blow tube, and vent locations. An Eulerian-Lagrangian numerical
model was used for theoretical analysis. Particle collision frequency was resolved by
mapping particle properties to the Eulerian grid, and then mapping the particle stress
gradient back to the individual particles. Results from the calculations compare well
with the experimental data.
In this article, we present the results of numerical simulations and experimental
measurements of process parameters during the PUA sand coremaking process.

CFD Modeling of Sand Core-Shooting Process

CFD modeling of the sand core-shooting process was based on a two-fluid flow
approach. The following governing equations were used in our computations.

Continuity equations:

(1)

08s
-
at + V · (BsP.J = 0. (2)

Momentum equations:

Pg[ 2+
0
Pg Vvg] = V · Tg - VP - : (vg - r,,.) + Pgg, (3)

(3
= V · Ts - VP., + -;- (Pg - P..) + Bs (Ps - Pg), (4)
g
Experimental Apparatus and Procedure 189

where

- + ( A; - 2 µ; ) tr (D;)/,
T; = 2 µ; D; - - (5)
3
-
D; = 1 ( 'vv; + (V 11;) T) . (6)
2
If the momentum equations are considered in terms of the resulting coefficients Cu
and nodal velocities uj, then the pressure gradients may be determined from
ap
C;;U; = -2.Cuu; + s; - Jn;-dv, (7)
ax
, 1 ap , ap
U· = u - - Jn--dv = u - K- (8)
' C;; ' ax ' ax'
-2,Cij+s; Jn;dv
Ct=----- K;=--. (9)
C;; C;;

Heres is the volume fraction; v, velocity; p, density; T, viscous stress tensor; P, pres-
sure; (3, drag coefficient; D, strain tensor rate; µ, viscosity; K, conductivity matrix;
n, interpolating function; and subscripts g and s refer to the gas and solids phases,
respectively.
Earlier we investigated the rheological properties of a resin-bonded sand-air mix-
ture using both capillary and rotational viscometers (Bakhtiyarov and Overfelt
1998b ). It was demonstrated that the resin-bonded sand-air mixture exhibits strong
non-Newtonian, shear-thinning behavior over a wide range of shear rates. Therefore,
in our simulations, we used the Ostwald model as a rheological equation to describe
resin-bonded sand-airflow:
·n-1
YJ = K"f (10)

where K is a measure of the consistency of the system and n is a measure of the de-
gree of non-Newtonian behavior.

Experimental Apparatus and Procedure

Silica sand supplied by the Nugent Sand Co. (Muskegon, MI) was used in all our ex-
periments. The size distribution of sand particles is presented in Figure 8.1. With
today's trend toward lower binder levels, the physical characteristics of sand, such as
sand grain shape and size, become very important. In his studies, Hoyt ( 1987) uses
four categories to describe the sand grain shapes: angular, subangular, compound,
and rounded. The surface areas of all these grain shapes are greater than that of per-
fectly rounded grains. The angular grains have a high smface area and require more
binders. The subangular grains require fewer binders and are compacted to a higher
density than angular grains. The rounded grain structure requires the fewest bonding
190 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

40
i I
: .11-,,
35 I
f "I '-'
'
'
I 1- -0- - Manufacturer's data
· · · O· - · Current data I
30 I /j t)
\I
I

I
!

' I I I
Ii I
I I I I
25 I I
I
§:'. I ,O•,
I \
I
'
I
! I
c 20 I I

.ii '\
I I .
"2 I
I
\
I
"
Cl.
15
I
\
I I !
y o, I

! ,..
I
I
\ \
: :

10
,r
I,'
p
I ,o
'. \
'o_ ",
I
'

5 ' ',
I
·,' --~.-:--~----
0 -~~ I - ---
--
0 100 200 300 400 500 600 700 800 900
Particle Size. µ,m

Figure 8.1 Size distribution of noncoated sand used in experiments_

agents and compacts to a higher density (8 percent to 10 percent higher than angular-
shaped sand). The compound grains are agglomerates of small grains bonded to-
gether. This bonding is not enough to maintain the integrity of grains during foundry
processes (Hoyt 1987).
In engineering calculations and nmnerical predictions, an appropriate shape (or
sphericity) factor must be applied to the particle size of the nonspherical particles.
The shape factor <j) is defined as the ratio of the surface area of a spherical particle
having the same volume as the particle of arbitrary shape A., to the surface area of an
arbitrarily shaped particle A:
P = cxV0111 - (3. (11)

Because a spherical particle provides a given mass with the least surface area, a
sphericity factor will be less than unity. Leva ( 1959) proposed the following relation
to calculate a sphericity factor:
vv2
<p = 0.205A' 0 2)
where Vis the volume of the particle.
Two methods have been used to determine sphericity factors of sand particles of
different sizes:
I. Microscopic method. Sand grain size and shape have been determined by scan-
ning electron microscopy (SEM) and the Olympus SZH Photo Microscope (PM).
Experimental Apparatus and Procedure 191

Air Compressor

Differential
Manometer

Filter & Dryer

CN 8511-Fl
Controller

Vent Sand PVC Tube

Air Flowmeter ECPV Power


Supply
Ul2YIOO

Figure 8.2 Experimental apparatus for studying the pressure decreases and rates of air flow-
ing through the sand bed. PVC, polyvinyl chloride; ECPV, electronically controlled proportion-
ing valve.

SEM achieved a magnification of up to I 000 times. Using PM, magnification of


up to 128 times was achieved with Olympus DF Plan 2X lenses. A side mirror
was used to obtain three-dimensional images of the particles. The photographs
were taken with Polaroid 667 Coatless black and white film with ISO 3000/36 de-
grees. LiteScope fiber-optic illumination was used during visual observations and
photography. The average volume of the particle was determined as

M
V=~ (13)
Np/
where Mis the mass of the N particles and Pp the molecular density.
2. Pressure-decrease measurements. A special experimental setup was designed to
measure the pressure decrease and rate of air flowing through the sand bed of
a certain particle diameter (Figure 8.2). Compressed, dried, and prefiltered air
under pressure of up to 700 kPa was provided to the test tube with a packed sand
bed of certain particle size. The airflow rate was controlled through the electron-
ically controlled proportioning valve (ECPV) and controller CN 8511-F I. The
Matheson 604 (E 700) rotameter was used to measure the airflow rate. The test
sample was placed inside a round tube made of polyvinyl chloride (PVC) with an
inside diameter of 15 mm and a length of 570 mm. The pressure decrease was
measured with a differential manometer. The specific surface of the sand bed was
192 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

Table 8.1 Sphericity factors of sand particles of different sizes

Sphericity Factor Sphericity Factor


by Microscopic by Pressure-Decrease
Particle Size (µm) Method Measurements

<150 0.724 0.730


150-212 0.756 0.742
212-250 0.756 0.764
250-300 0.780 0.784
300-355 0.805 0.786
355-425 0.810 0.820
425-500 0.835 0.826
500-600 0.840 0.830
600-710 0.860 0.848
710-850 0.855 0.840
850-1000 0.804 0.818

estimated from pressure-decrease measurements by rearranging the Carman-


Kozeny equation:

AP / g!J.Pe3
(14)
~' = \) 5(1 - e)2µ1 UL'
where e is the average void fraction and was determined by the method described
in previous work (Bakhtiyarov and Overfelt 1998a), µ1 is the viscosity of fluid
(air), U is the average velocity of air, and Lis the sand bed length. The results of
the sphericity analyses by both methods are presented in Table 8.1 and Figure 8.3.
The molecular density of the sand was determined by Archimedes' principle
using the Mettler Toledo densitometer. The density of the sand portion was deter-
mined with the aid of an auxiliary liquid having known density. We used pure alco-
hol. A portion of the sand was weighed in air and then in the auxiliary liquid. The
molecular density of the sand was calculated from the two weights as follows:

P = · Pa.I.• (15)
Wair - Wa.l.

where Pa.I. is the density of the auxiliary liquid at a given temperature and Wair and
Wa.l. are the weights of the test specimen in air and in the auxiliary liquid, respec-
tively. Measurement accuracy was O.OOlpercent. The results of the density mea-
surements versus particle size are presented in Figure 8.4. As seen, the average
molecular density of sand particles of size 200 to 600 µmis about 2.60 g/cm 3 . Due
to nonsilica inclusions, the density of particles below 200 µm and above 600 µm is
lower.
The density and volume fraction of noncoated and coated sand beds were deter-
mined by the Gay-Lussac principle. The results of the measurements are shown in
Table 8.2.
Experimental Apparatus and Procedure 193

0.86

0.82

0
:g 0.78
1l0..
(I)

0 Microscopic Method

D Pressure Drop Method


0.74

0.7
0 200 400 600 800 I 000
Particle Size, fLill

Figure 8.3 Variation of sphericity versus sand particle size.

2.7

2.6

(<)

E 2.5
~
.~
"
'2" 2.4
3"'u
0"
:;S
2.3

2.2
0 200 400 600 800 1000

Particle Size, fLill

Figure 8.4 Variations in molecular density versus sand particle size.


194 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

Table 8.2 The density and volume fraction of noncoated and coated sand beds

Sand Beds Density (g/cm 3 ) Volume Fraction

Noncoated and noncompacted 1.51641 0.58323


Noncoated and compacted 1.67136 0.64283
Coated and noncompacted 1.20188 0.46226
Coated and compacted 1.59155 0.61213

Microscopic photography of coated sand particles of different size ranges al-


lowed us to estimate the size distribution of resin-coated sand. The observations
show that the binders are retained on the sand grains and coat their surfaces almost
uniformly. Based on this information, we can predict the size of coated sand parti-
cles by the following expression:

(16)

where De and D 11 are the diameters of resin-coated and noncoated sand particles, re-
spectively; C1, is the binder's level based on sand weight (in our experiments, it is
0.015); V, and S, are the volume and surface area of sand grains, respectively; and p,
and P1, are the densities of the sand material and binders (2.60 and 1.10 g/cm 3 in our
experiments), respectively. The results of predictions and microscopic measurements
are presented in Figure 8.5, along with data for noncoated sand size distribution. As
seen from Figure 8.5, the addition of binders does not change the shape of the size
distribution curve, but shifts it to higher values of the grain sizes due to the coating
layer. Also, good agreement exists between predictions and observations. Micro-
scopic photography of the fluidized sand-binders system showed that during flu-
idization (blowing) the sand particles create agglomerates of 500 to 2000 µ,m in size.
As a cold box binder system we used the ISOCURE LF-305/904 G system pro-
duced by Ashland Chemical Company (Dublin, OH). This system consists of an
ISOCURE Part I LF-305 phenol-formaldehyde binder and an ISOCURE Part II 52-
904 GR binder containing a polymeric isocyanate with solvents and additives. The
reaction between Parts I and II results in the formation of phenolic urethane polymer.
According to the existing procedure of the sand coremaking process in the foundry
industry and following the manufacturer's instructions, in our experiments we used
the composition at a 55-to-45 ratio of ISOCURE Part I binder to ISOCURE Part II
binder components, Ashland Chemical ( 1994). Increasing the amount of Part I
decreases the humidity resistance of the core and the bench life of the mixed sand. In-
creasing the amount of Part II increases the cost and shakeout time of the core prod-
uct. Recently, we investigated experimentally the rheological and thermal properties
of the phenolic resin and polymeric isocyanate (ISOCURE and TECHNIKURE),
and their blends (Bakhtiyarov and Overfelt 1997a, 1997b, 1998a). We determined that
although both binders are Newtonian liquids, their blends exhibit non-Newtonian
shear-thinning fluid flow behavior and elasticity. The dynamic viscosity of the blends
increases with both time and increasing the ISOCURE Part I LF-305 content, and
Experimental Apparatus and Procedure 195

25
-¢-Non-coated sand (sieve
measurements)
- - Coated sand
20 (calculations)
o Coated sand (microscopic
measurements)

15

0
100 200 300 400 500 600 700
Particle Size, µ.m

Figure 8.5 Size distribution of noncoated and coated sand used in experiments.

may reach comparatively large values at large values of either parameter. The in-
creasing viscosity of blends is explained as the result of the rubbery nature of the
phenolic urethane polymer, which was produced as a product of the reaction be-
tween Parts I and II. Due to the degradation effect at high shear rates, the viscosity of
blends decreases. At fixed shear rates we observed a temperature increase during the
mixing of binders. Following the manufacturer's instructions, the total binder level
was 1.5 percent based on sand weight. Before it was blown into the pattern, the sand
was coated with the binders. The bulk density of the coated sand was 1.426 g/cm 3 •
For gassing we used liquid Triethylamine (Catalyst 700), which was vaporized dur-
ing curing.
A special pressure measurement system was designed and built with the safety
and portability requirements of the foundry environment. All sensor electronics were
housed in a special briefcase and were connected to a computer-controlled data
acquisition system through PC-LPM-l 6/DAQCard-700. Up to 14 piezoresistive ab-
solute pressure sensors PX72-1 OOV (Omega Engineering, Inc.) suitable for mount-
ing on a personal computer board (linearity: ±0.5 percent FS, repeatability:
±0.3 percent FS, operating temperature range: -15°C to 85°C) were used in a sin-
gle experiment. These silicon diaphragm sensors are stable, high-performance, mi-
cromachined sensors without temperature compensation for operating ranges of O to
690 kPa. The sensors with a response time of 0.1 sec were operated from a constant
current excitation supply. To prevent of sand and dust particles from penetrating the
196 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

I
Pressure, sensor# 14

ro 150 '---~~~~~~~~~~-'-~~~~____j~=-~~~-1---~~~~--1
:!;i

<> Without snubber


50 - D With snubber

0 l._.c..['~....__....-1...~~~~...._..J_....__...._._~---1~~...._....._....__..L_._~~~.,__J

200 300 400 500 600 700


Output Voltage, mV

Figure 8.6 Calibration of pressure sensor (#14) with and without the snubber.

measurement system, porous metallic-element brass pressure snubbers (Chemiquip


Products Co., Inc.) were mounted on each pressure tap. The snubbers also protect
transducers and other delicate line installations from shock damage caused by pres-
sure pulsation, surges, and fluctuations. The snubbers contain a Type-316 stainless
steel porous disc (porosity designation D and a rating of 40 to 45 µm). After each ex-
periment all snubbers were brushed and washed in acetone. Special experiments
were run to determine the effect of the snubbers on measured pressure values. The
test results for pressure transducer #14 are shown in Figure 8.6, and as can be seen,
the influence of the snubber on measured pressure values is negligible ( < 1 percent).
The calibration results for all pressure sensors used are presented in Figure 8.7.
As seen, the output voltage V0111 of pressure transducers increases linearly with in-
creasing pressure. This relationship can be expressed as
P = aVuut - (3, (17)
where a and (3 are calibration coefficients of the pressure sensors. These coefficients
were determined experimentally for each pressure sensor, and the results are pre-
sented in Table 8.3.

Results and Discussion

One would assume that during the mold-filling process, the coated sand depositions
in the vent areas would cause a vent permeability decrease. Therefore, it will be an
important issue for mathematical modeling and optimization of the core sand-filling
Results and Discussion 197

250
<>l o2 t.3 o4 +5 a61..7 e8 <>9 oIO t.11 012 ~13 llll4

200

150
"'
~
i
~ 100
~
0...

50

0
100 200 300 400 500 600 700

Output Voltage, mV

Figure 8.7 Calibration curves of pressure sensors.

process. In our previous studies the variation of the pressure losses with the amount
of sand-binders depositions in the vent area was measured experimentally (Bakhti-
yarov and Overfelt 1998a). In this study we modified the test section to obtain more
precise and consistent data for the test sand-binders system. An experimental setup
shown in Figure 8.8, consists of a round tube of PVC with an internal diameter of

Table 8.3 Calibration coefficients for pressure sensors.

Sensor Number a, kPa/mV (:J, kPa

0.7415 88.822
2 0.6756 112.54
3 0.5639 124.33
4 0.5665 158.37
5 0.6704 153.31
6 0.6810 127.79
7 0.5070 97.275
8 0.5047 130.87
9 0.6638 125.93
10 0.6627 155.14
11 0.6723 164.98
12 0.6581 146.89
13 0.5073 120.21
14 0.5264 125.07
198 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

Com pressed Air


PC

Filter & Dryer

CN 8511-Fl
Controller

Air Flowmeter ECPV Power


Supply
Ul2YIOO

Figure 8.8 Experimental technique for studying pressure losses in a core box due to sand
deposition in the vent area during the mold-filling process. PC, personal computer; PVC,
polyvinyl chloride; ECPV, electronically controlled proportioning valve.

19 mm and a length of 428 mm. A test vent was installed at the end of the tube. Com-
pressed, dried, and prefiltered air under pressure of up to 500 kPa was provided to
the tube filled by the test material. The airflow rate was controlled through the
(ECPV) and controller CN 8511-FI. Matheson (604, 605, GE 606) rotameters were
used to measure the airflow rate. The locations of the pressure sensors along the test
tube are presented in Figure 8.9.
The total pressure decrease 11P due to air flow in the cylindrical tube filled with
sand (bonded or unbonded) with the vent at the outlet can be represented as
(18)

where .1P1, 11P,,, and 11P., are the pressure losses in the tube, vent and sand, respec-
tively. We ran special experiments to estimate the pressure losses due to the tube

25.4 25.4 25. 25. 25.4 25.4 25. 25. 72.4

#14 #13 #12 #11 #IO #9 #8 #7 #6 #5 #4 #3 #2 #I

Figure 8.9 Positions of pressure sensors along the test tube.


Results and Discussion 199

120

100

80

60

40

0 100 200 300 400 500 600 700 800


3
Air Flow rate, cm /s

Figure 8.10 Pressure variations with the airflow rate in an empty conduit with a vent.

walls and vent. Figure 8.10 represents the variation of the pressure against the air-
flow rate in an empty tube with the vent. The pressure was measured at the inlet
(sensor #1) and outlet (sensor #14) of the tube. As seen in Figure 8.10, pressure
losses due to the tube and vent are small (up to 2 percent). Nevertheless, in analyses
of pressure losses in a sand bed (bonded and unbonded sand) these contributions
were taken into consideration.
The pressure variations with the airflow rate in a cylindrical tube filled with un-
bonded and resin-bonded sand are presented in Figures 8.11 and 8.12, respectively.
To imitate the real sand core-blowing process, sand was not preliminary compacted
in the test tube. Therefore, a nonlinear relationship between the pressure and airflow
rate was established. Due to sand compaction, the pressure increased when the air-
flow rate was increased, for both unbonded and resin-bonded sands.
Figures 8.13 and 8.14, respectively, represent the variation of the pressure de-
crease with the airflow rate for unbonded and resin-bonded sands. As seen from
Figures 8.13 and 8.14, the nonlinearity is more pronounced for longer sand beds in
the tube. This again confirms the relationship between the nonlinearity of the pres-
sure decrease behavior and sand compaction in the conduit.
The variations of the pressure decrease with the amount of unbonded and resin-
bonded sand deposited in the vent area are shown in Figures 8.15 and 8.16, respec-
tively. A linear relationship was found between the pressure decreases and the
unbonded sand bed length at a wide range of airflow rate values. However, pressure
200 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

300
-¢--#I
--0- -#2
- - -t,- - - #3
250
-·X--#4
--)K-- #5

----o--#6
200
- -+- -#7
. - ..... - #8
~
"'
0..
-----#9
,,; 150 __ ..__ #10
'""'
~ ---11---#ll
"'...
0..
----&---#12
100
---•--· #13
-·D·-

50

0
0 100 200 300 400 500 600 700
3
Air Flowrate, cm /s

Figure 8.11 Pressure variations with the airflow rate through noncoated sand.

decrease increases nonlinearly with the amount of deposited resin-bonded sand in


the tube. Moreover, at certain values of the airflow rnte (-250 cm3/sec) a significant
increase in pressure decrease was observed for resin-bonded sand. We suggest that
this behavior is related to the complex rheological behavior of the resin-bonded
sand. Earlier it was demonstrated that the resin binders provide certain plasticity to
the sand (Bakhtiyarov and Overfelt 1998a). The plasticity of the sand-binders mix-
ture is characterized by a yield shear stress, which provides the system with some re-
sistance against compaction. However, further investigations and analyses are
needed to explain this kind of behavior in pressure decrease.
The flow dynamics of sand and the sand-binder system in a core box was run
in the system shown schematically in Figure 8.17. The test core specimens were
produced using the LAEMPE Test Specimen Curing Machine LI. This machine con-
sists of a column, a shoot unit, a table, fixed and moveable side clamps, a gas carriage,
a cabin, and a control unit. The machine also houses the control box, valve manifold,
and gas generator. The machine table moves pneumatically under the shoot unit for
shooting a sand-binder mixture. After the table moves down, the gassing plate moves
in pneumatically and is pressed between the shoot head and core box when the table
moves up again. Gassing of the core is then begun. The gas generator heats a liquid
catalyst until it reaches a gaseous form. The liquid is dosed (0.2 cm 3/stroke) into a
heater where the catalyst evaporates into a gas, which is fed to the sand core through
a system of gassing plates. The gas together with heated air are blown through the
Results and Discussion 201

300
---0--#l
--0- -#2
250 ---!:,--- #3
--x--#4
- -:I:-- #5
200 ---o--#6
- -+- -#7
-----. #8
:;;1"' 150 -----#9
i --+--- #10
~
~ . -11--#l l
0.. 100
--Jr--#12
···•··· #13
50

0
0 JOO 200 300 400 500 600 700 800
3
Air Flowrate, cm /s

Figure 8.12 Pressure variations with the airflow rate through resin-coated sand.

160
o 13-14
D
t;
12-14
11-14
.J
.I-

.,II
f

120 X 10-14 •
::i:9--14 •-·
;
;

o 8-14
+ 7-14 -:"'
:;;1"'
·6-14 --....
., •.. .·
2
ci.
80 -5-14
...
Q
~
~
en
+4--14
11 3--14
-· ,, _:1--7

£ .+. 2-14
40 e 1--14

0 0.0002 0.0004 0.0006 0.0008


3
Air Flowrate, m /s

Figure 8.13 Pressure decrease variations with the airflow rate through noncoated sand.
202 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

100
o 13--14
D 12--14
,:,, 11--14
80 X 10--14
:( 9--14
o 8-14
60 +7--14
-6--14
-5-14
+4-14
40
• 3--14
•2--14

20

0~~~~~~~~2:::~~~~
0 0.0002 0.0004 0.0006 0.0008
1
Air Flowrate, m· /s

Figure 8.14 Pressure decrease variations with the airflow rate through resin-coated sand.

core to purge it. When gassing is complete, the table moves down again; the gassing
plate moves out; and when the side clamps and cabin doors open, the core can be re-
moved from the machine. All functions of the machine are pneumatically moved and
controlled. An exhaust system with an exhaust capacity of about 500 m3/h was pro-
vided to the cabin. The shapes and sizes (in millimeters) of the core boxes (A and B)
used are shown in Figures 8.18 and 8. I 9. The core boxes have an aluminum pattern
with a clear Plexiglas plate mounted on the front wall to perform visual observations
of the flow. The sealing of the core box is achieved when the table is lifted and
pressed, creating very strong closing forces on the tooling. Core box A has a hole
with a diameter of 12. 7 mm in the middle of the H-shaped insertion.
In a cold box process type, number, size, and placement of vents are important
factors for core quality, core density, and catalyst consumption. They affect both the
ability to blow a dense uniform core and the uniform permeation of the catalyst gas
throughout the core. Recommended total vent area in the cope side is 356 X 1o-s to
569 X 10- 5 cm 2/g of core (Carey and Sturtz 1977). In our experiments we used four
slotted steel vents of 12.7-mm diameter and two slotted steel vents of 10-mm
diameter, with 0.5-mm slots. To evaluate the vent area available to the flow, we
introduced the mean hydraulic radius R,,, which is the ratio of the cross-sectional
area of the vent available to the flow to the wetted perimeter. Regarding the vent as
an opening with a complicated cross-section, we ran experiments to determine the
mean hydraulic radius R1, of the vent used. The experimental apparatus used to
Results and Discussion 203

140
-----0- 0.000 I 56m"3/s

- -o- - 0.000222ni'3/s
120 - - -/',- - - 0.000247ni'3/s
- -x -- 0.000277ni'3/s

-
- - )I(- - 0.000308m"3/s
JOO
- o - 0.000334m"3/s
- -+- - 0.000366m"3/s
" 80
""
-"
~
- - . - - - 0.000399m"3/s
- - - - - 0.000432ni'3/s
2
0 - . +- . 0.000465m"3/s
Q)
I-<
60
i;; - 0.000522ni'3/s
"'~ - -*" - 0.000551 rd'3/s
"" 40 - • -<>· - - 0.000581 m"3/s
- · o -- 0.000641 ni'3/s
20

0 •
0 0.04 0.08 0.1 2 0.16 0.2 0.24

Length of Sand Bed, m

Figure 8.15 Pressure decrease variations with the amount of noncoated sand deposited in
the vent area.

determine the hydraulic radius of the vent is sketched in Figure 8.20. The liquid
(distilled water) is supplied from a pressurized feed tank to the test section with the
vent. The flow rate Q is controlled through the valve located at the exit of the test
tube. At the exit, the test liquid is collected in the graduated cylinder and the mean
volumetric flow rate is measured by the weight method. A PVC tube with an internal
diameter of 15 mm and a length of 570 mm was used in the experiments. A slotted
steel vent was attached to the exit of the flow cell. Average velocity in the tube was
defined by the Hagen-Poiseuille equation:

(v) = fj,,.PR2. (19)


4µ,L
The mean hydraulic radius of the vent was defined in terms of the average flow rate
and the average velocity in the tube as
Q )05 (20)
R1, = ( 7f (v)

The experimental data suggest that the hydraulic radii of the vent tested are 4.26
and 3.35 mm, respectively, for vents of 12.7 and 10 mm in diameter. Hence, the hy-
draulic cross-sectional areas of the vents are 0.57 and 0.35 cm 2 , respectively. In Fig-
ures 8.18 and 8.19, we present the number and locations of the vents in the core box.
204 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

100
--o--0.000005ni'3/s
- --o- - 0.0001 I6ni'3/s
·. ·6·. · 0.000164ni'3/s

80 - - · X --
- - ;i::- -
0.000213ni'3£,
0.000252ni'3/s
~
I

- o - - 0.000303ni'3/s
- -+. - 0.000350ni'3/s
60 - . - - - - 0.000415ni'3/s
~"' - - - . - 0.000460ni'3/s
Cl.
0>.; - . +- - 0.000505ni'3/s
Cl
~
::,
:2
"'
>.;
p..
40
-a--0.000545ni'3/s
- ....,.... - 0.000581 ni'3l,
- . -•·. · 0.000631ni'3/s
- . o . - 0.000668ni'3l,
-
<'
: ~-
20 -

0
0 0.04 0.08 0.12 0.16 0.2

Length of Sand Bed, m

Figure 8.16 Pressure decrease variations with the amount of resin-coated sand deposited in
the vent area.

Following the recommendations of Carey and Sturtz ( 1977), we placed vents in low
pockets and near sharp corners to direct the sand and the gas catalyst vapors toward
those points. During the experiments, the total vent areas were 3.86 X 10- 3 and 3.88 X
10- 3 cm 2/g of core boxes A and B, respectively. The total area of box exhaust vents
was 365 percent of the total vent input area. The size and shape of the coupling be-
tween the hopper and core box (input vent) are presented in Figure 8.21. No special
blow tubes were used in these experiments.
Visual observations of the dynamics of the sand-blowing process were made by
using a video-camera recorder (VCR) Sony CCD-TR93 (Figure 8.17). The sand-
filling patterns and interface displacements were computerized by means of TBC
(Time Base Corrector), VT (Video Toaster), and PC "AMIGA." Fourteen piezore-
sistive absolute pressure sensors PX72-IOOV (Omega Engineering, Inc.) were used
to measure the pressure variations during the sand-blowing and gassing-purging
processes in the hopper, blow tube area, and core box. The locations of the pressure
taps were chosen based on the visual observations of the sand-binders and airflow
dynamics (Figure 8.22).
Figures 8.23-8.26 show a variation of pressure versus time for airflow through
the hopper, blow tube, and empty core boxes A and B. As seen, the measured pres-
sure values in the hopper are quite consistent for both core boxes A and B. However,
Results and Discussion 205

LAEMPE® Curing Machine


Sand

Corebox
Camcorder - Sensor

[J=1 - Electronics

D D
Data Acquisition Board
VCR
DAQ Card 700
Pressure & Flow

D
Sensors

D
Time Base PC
Corrector Video Toaster

Figure 8.17 Experimental apparatus used to study the dynamics of the core shooting
process. VCR, video-cassette recorder; PC, personal computer.

an absence of the opening in the middle of the H-shaped insertion (core box B)
causes about a 9 percent pressure increase inside the core box.
Pressure variations (in the hopper, blow tube, and core box) versus time during
sand-binders blowing into core boxes A and B are presented in Figures 8.27-8.32.
As seen, actual sand-shooting takes only 0.3 sec. Again, there is not a significant dif-
ference in pressure values measured in the hopper ( ~4 percent) and blow tube area
( ~8 percent) for core boxes A and B. After the mold-filling, the pressure at the blow
tube area for core box B sustains about 17 percent more from its peak value than the
same pressure in core box A. A significant pressure increase (up to 60 percent) was
observed in core box A compared with core box B. We assume that the elimination
of the cylindrical opening in the middle of the insertion generates additional energy
losses in the core box.
A significant pressure decrease in the blow tube is due to the large difference in cross-
sectional areas of the hopper and blow tube (the cross-sectional area of the hopper is
10 times greater than that of the blow tube). However, pressure decrease is also re-
lated to the density and sand-to-air ratio of the resin-bonded sand bed. The density of
the sand bed in the hopper was 1.426 g/cm 3 , and the void factor was</> = 0.42. Vig-
orous fluidization occurs at the exit from the blow tube. Our previous experiments
showed a dramatic decrease in density of the sand bed and increase in the void fac-
tor. As we know, fluidized systems require much less energy for throughflow gas
206 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

152

2 Vents
D=l2.7
116
152

Figure 8.18 Dimensions of core box A.

than that required for porous medium flow. As further fluidization takes place inside
the core box, additional pressure decrease is observed. The Sand-air system flow ex-
periences a 20-fold increase in cross-sectional area, and one would expect a pressure
increase in the core box. However, further fluidization of the sand bed occurs in the
core box, which results in a decrease in the density and sand-to-air ratio of the sys-
tem. As seen from Figure 8.10, the pressure inside the blow tube is twice of that in
the core box. The pressure data obtained in these experiments are very informative.
From these data one could estimate the vaiiation of the density and the void factor
inside the core box during the sand-molding process.
The data obtained by barometric measurements were used to create contour maps
of the pressure distribution in core boxes A and B. The isobars were simulated using
the cell-valued discretization method. This method was described by Bakhtiyarov
and co-authors ( 1978), and it was used earlier to study the dynamics of the displace-
ment of fluids from channels with cavities. Figures 8.33 and 8.34 compare pressure
distributions (isobars) at the instant of sand-shooting inside core boxes A and B, re-
spectively. As seen from Figures 8.33 and 8.34, for both core boxes, a high-pressure
Results and Discussion 207

152

2 Vents
0=12.7
116
152

Figure 8.19 Dimensions of the core box B.

area was found adjacent to the blow tube. Low pressures were found at the bottom of
the core box and under the H-shaped insertion of core box B. Unfortunately, a lim-
ited number of the pressure transducers did not allow for creating isobars in small in-
crements of pressure, and especially did not allow for revealing stagnation zones
inside the core box. Further experimentation is needed to obtain more information
on pressure variations during the sand mold-filling process.
Variation of volume flow rate of the air blown into core box A versus the time dur-
ing the sand-filling cycle (measurements are made at point #1) are shown in Figure
8.35. As seen, variations in the airflow rate are consistent with the data obtained for
the pressure variations inside the core box. Actual blowing time is about 0.3 sec, dur-
ing which the airflow rate reaches its peak value (fluidized bed), then decreases al-
most 70 percent, and for about 2 sec remains constant (flow through porous medium).
Finally, the flow rate decreases to zero. This flow curve and pressure curves clearly
represent the flow dynamics of the sand-air mixture inside the core box.
In Figure 8.36, we present the results of both simulations and experiments for
pressure variations along the vertical symmetric line of core box B. Here y is the
208 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

Air

Feed Tank

Cylinder

Figure 8.20 Experimental apparatus to determine the hydraulic radii of vents. PC, personal
computer.

....- .... '


.;
'
Hopper
I
/
;
25.4
' \
I \
\
7 \
I
\
I
8 \
I
I \
18.5
I I
I 25.4
I
Core Box

Figure 8.21 Coupling between the hopper and core box.


Results and Discussion 209

18.5
9.5
45
45
40

45

45
20

9.5

Figure 8.22 Positions of pressure transducers over core boxes A and B.

vertical axis with origin at pressure tap #3, and H is the distance between pressure
taps #3 and #13. As seen, there is good agreement between the numerically simu-
lated and measured values of pressure. A significant pressure decrease occurs be-
tween the blow tube and upper surface of the horizontal section of the insertion.
Under the lower surface of this section, we can see a parabolic variation of pressure.
Zero values of pressure refer to the solid smfaces of the core box and insertion.
It is primarily of practical interest to identify a pressure distribution inside the
core box during the gassing and purging cycles. During sand mold-filling, we had a
highly fluidized sand-air system in which gassing and purging cycles flowed air
through the fixed sand bed. The pressure insider the core box was mainly affected
by the properties of the porous medium. Figures 8.37-8.42 show pressure varia-
tions (in the hopper, blow tube, and core box) versus time during airflow through
the sand-binders bed in core boxes A and B. Again, a significant pressure decrease
can be observed in the blow tube and core boxes. However, pressure segregation
across the core box is more pronounced. High pressures ( ~25 kPa) are found in the
upper portion of the core box; low pressures ( ~s kPa) are measured in its lower
sections. In contrast to the sand mold-filling cycle, because no further significant
21 0 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

450 ----0-- l
--D--2
400
- - -le,- •• 3
!
I --0--4
350 - L

--+-- 5
300 -------6
--A-- 7

:;l"' 250 ···•··· 8


i --X --9
;::l
200 ---- 10
:2
J: 150
-:K-11
--+--12
---0·-· 13
100

50

0 !:JI
3.7 3.8 3.9 4 4. l 4.2 4.3 4.4
Time, sec

Figure 8.23 Pressure variations versus time for airflow through the hopper (curves 1 and 2),
the blow tube (curve 3), and empty core box A (curves 4-14).

175
- - ·il· - - 3

173 --o--4
' -·+-· 5
171 -------6
--A---7
169
···•··· 8
--x--9
:;l"' 167 -+
---· 10
i -:K-11
~
<U
165
.... --+--12
0...
163 ---0--- 13
--<>--14
161

159

157
3 .88 3.9 3 .92 3 .94 3.96 3 .98 4
Time, sec

Figure 8.24 Pressure variations versus time for airflow through empty core box A (curve 3
pertains to the blow tube, curves 4-14 to the core box).
Results and Discussion 211

400 ----o---- I
--D--2
350 3
---6---
--o--4
300 --+-- 5
--6
----.ir--7
250
---•--- 8
:;i"' --x--9
a[ 200 ---- 10
~ -:l(-11
~ 150 --+--12
0..
---0--- 13

100

50

6.4 6.5 6.6 6.7 6.8 6.9 7 7 .I

Time, sec

Figure 8.25 Pressure variations versus time for airflow through the hopper (curves 1 and 2),
the blow tube (curve 3), and empty core box B (curves 4-14).

195
!I ---6--- 3
--o--4
I
190
_.--r --+-- 5
--6
----.ir--7
185
-··•··· 8
--x--9
:;i"'
180 ---- IO
-:l(-11
a[
--+--12
",,,,,,~ 175
\.j ---0---13
0..
· --<>--14
170

165

160
6.57 6.59 6.61 6.63 6. 65 6.67 6.69
Time, sec

Figure 8.26 Pressure variations versus time for airflow through the empty core box B (curve 3
pertains to the blow tube, curves 4-14 to core box B).
212 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

500
-¢-I
--o--2
···A··· 3
400 I
I --0--4
I
I -·+-· 5
I
o -6
ro 300 I
--+-- 7
A-.
.-'<: I
···•··· 8
ii -·X --9
~ 200
<I) - - · - · IO
I-<
A-.
-:K-11
--+--12
100 ···D··· 13
-·<>--14

6.5 7 7.5 8 8.5


Time.sec

Figure 8.27 Pressure variations versus time during sand-binders blowing into core box A
(curves 1 and 2 pertain to the hopper, curve 3 to the blow tube, and curves 4-14 to core box A).

100
···A··· 3
--0·-4
-·+-· 5
80 -
-6
--+--7
.. ·•·-· 8
60 --x--9
·-A-·.t,··-A.
- · - · IO

i 40 -
-:K-11
--+--12
\I
<I)

~ ···D··· 13
-·<>--14

20

6.5 7 7.5 8 8.5


Time, sec

Figure 8.28 Pressure variations versus time during sand-binders blowing into core box A
(curve 3 pertains to the blow tube, curves 4-14 to core box A).
Results and Discussion 213

30
--+--5 -a-6 --+--7 ---•---8 --x--9
-;i:-11 --+--12 ---0---13 --<>--14

25

20

:;:J"'
OJ 15
'"'
~
<!.l
'"'
0...
10

0
7 7.2 7.4 7 .6 7.8 8
Time, sec

Figure 8.29 Pressure variations versus time during sand-binders blowing into core box A
(curves 4-14 pertain to core box A).

500
---0-l
--o--2
3 · ··A· - -
400 1-------+1---f-------1---------j,--\~---o--4
: --+-- 5
I -------6
I -4--7
300 ~-----J'-11_ - l - - - - - - - - + - - - - - - I - - - I - - . - -•- .. 8

3l"' --x--9
2£ ---- 10
5;
Ul
200 +------•----+--------4-------,----;i:-11
J: 1 --+--12
- - ·Cl· - - 13
--<>--14
100
'l!, .• t:,--t:,--A·· --A--A··A--A-- ··n.

0
8.5 9 9.5 10 10.5
Time, sec

Figure 8.30 Pressure variations versus time during sand-binders blowing into core box B
(curves 1 and 2 pertain to the hopper, curve 3 to the blow tube and curves 4-14 to core box B).
214 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

90
···A··· 3
--o--4
80
--+-- 5
70 ii .·. t -A·. -t,.·. -/',-. -/',··
.
4!'\
-11--6
J:....-A··-A··-/',·· •• --lk--7
60 ii 4 ···•··· 8
--x --9
.,."'
P... 50 ---- IO
,j
'-< -:i:-11
~ 40 --+--12
~
P... - - ·D· · · 13
30
--~--14
20

10

0
8. 5 9 9.5 10 10.5
Time, sec

Figure 8.31 Pressure variations versus time during sand-binders blowing into core box B
(curve 3 pertains to the blow tube, curves 4-14 to core box B).

25
--o--4 -·+-· 5 - 6 --lk--7 ···•··· 8 --X --9
---- 10-:te-l I --+--12---0--- 13--~--14

20

.... ·
.•......•... .. -·· ··•· """ ""." ·•··"" ··•·""" ..•. "

15
'.,1\
·1
'.\

9.2 9 .4 9.6 9.8 10 10.2

Time, sec

Figure 8.32 Pressure variations versus time during sand-binders blowing into core box B
(curves 4-14 pertain to core box B).
Results and Discussion 215

486

477

Figure 8.33 Pressure distribution inside core box A.


216 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

486

477

Figure 8.34 Pressure distribution inside core box B.


Results and Discussion 217

0.02

~
0.015

'~
"'-"' 8
0.01
~
'0is"'
G::

0.005

0
0
l
<

0.5 1.5
\ 2
Time, s

Figure 8.35 Airflow rate variations versus time during sand-binders blowing into core box A
(measurements are made at point #1).

160

0 Experiments
120 - - Simulations

"'
~
80
i
~
1:
0...

40

a
0
0 0.2 0.4 0.6 0.8
y/H

Figure 8.36 Pressure variations along the vertical symmetric line of the core box during
mold filling. H, distance between pressure taps #3 and #13; y, vertical axis.
218 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

600
--0--2 ···/1--· 3 --0--4 --+--- 5 - - 6 ----4----7
--x--9 ---- IO-:i:-11--+--12---o---13--<>--14
500

400
cj

~
i[ 300
~
"'
....
p..
200

100

10.25 10.75 11.25 11.75


Time, sec

Figure 8.37 Pressure variations versus time during airflow through sand-binders in core
box A (curves 1 and 2 pertain to the hopper, curve 3 to the blow tube, and curves 4-14 to core
box A).

. . -!1- .• 3

--o--4
-·+--· 5
--6
----4----7
···•--. 8
--X--9
---- 10
-:i:-11
--+--12
---0--- 13
--o---14

-l:S"..:'.l:!' =~:.:..::lf' ,l;S.


CJ =6 ~Cl ...--:o =Cl -

10.75 11.25 11.75

Time, sec

Figure 8.38 Pressure variations versus time during airflow through sand-binders in core
box A (curve 3 pertains to the blow tube, curves 4-14 to core box A).
Results and Discussion 219

30

_..o.,
cf..,.---~
25
f---·--•-··•·
,;· -·0--4
·I
20 1 - - - + - - - - - - 1 - - - - - - - - - - + - - - - ~ ~ - - , __ .,__ 5
',\ -6

::l"
, ·i --A.--7
15 f----,_;1-----+-------1----~'tt ....,.:.-i .. ·•· .. 8
"'~"' --x--9
~ -·-- 10
0...
10 ~.-H--l -:i:-11
--+--12
13
• ·.CJ..•

5 -'<-,..~-,\\----l -·<>--14

0
10.25 10.75 11.25 11.75
Time, sec

Figure 8.39 Pressure variations versus time during airflow through sand-binders in core
box A (curves 4-14 pertain to core box A).

600
---0-l --o--2 .. ·D.· .. 3 -·0·-4 -·+-· 5
-6 --A.--7 ···•··· 8 --x--9 - · - · 10
-:i:-11 --+--12 · - ·Cl· - - 13 --<>--14
500

400

"'
0...
-"' 300

"~
I-<

0)

'"'
0... 200

100 I
, • ···D.···D.· ·D.···D.···t,.···D.··-D.·;.r,. ...r,. ... ···D..

0 D~t~t~~B~~g;.t~~~.i.t.:..'tc§a,:,"tcia.-:
17.75 I8 18.25 18.5 18.75 19 19.25 19.5

Time, sec

FigureB.40 Pressure variations versus time during airflow through sand-binders in core box B
(curves 1 and 2 pertain to the hopper, curve 3 to the blow tube, and curves 4-14 to core box B).
220 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

80
- - -!:.- - - 3
70 --o--4
--+-- 5
60 -6
--A---7
50 ---•--- 8
--X --9
~"
40 . - ---- 10
~
-:i:-11
~
~ 30 --+--12
p...
13
- - -Cl- - -

20 --<>--14

10

0
17 .7 5 18. 25 18 .7 5 19.25 19.7 5
Time, sec

Figure 8.41 Pressure variations versus time during airflow through sand-binders in core
box B (curve 3 pertain to the blow tube, curves 4-14 to core box B).

25
.fI "":.
1-6
--0- --0- - --0- ---0-.. - -0 "- 4
". .... o--..,
!1 ·-t-:-j:-&---.-•---:.,-.--,-,~--- -"'· I

--+-- 5
,3: 1· --+-- ---- •i-... -....., __:;\
20 .• • --A---7
{ - i- \\ "--•-"" 8

ro
p...
15 ,_-.f
1:
I 1
I
\I =:~~~
,\i -:i:-11
0

~
!l) o., . ~I --+--12
l
f. ,,
rL,.__~- \"' -
, .._
\.;
---Cl--- 13
~
10 u J ~ - .....
...-'f1~- ... ,_-Ir ,-.t.\ ','\'
- . <>. -
14
"\.;
p... !1
,,
1
I 11.- -
, ,
:
°t'
'II: I
1/X·. . ~ __..x,_ \... 1\:
!: 1 ,...'x-71o+ .,-1o+,.-x, ,..x ·X-,--io+. ·x 1 ],I
JI//,' .._..• ,,.,., "..___·_,,.,:,. ·--· ...... ",. \,;\
5

17.75 18.25 18.75 19.25


Time, sec

Figure 8.42 Pressure variations versus time during airflow through sand-binders in core
box B (curves 4-14 pertain to core box 8).
Acknowledgments 221

sand compaction occurred during the gassing and purging cycles, we did not ob-
serve any peaks on the pressure-time curves obtained in the blow tube or core box.

Conclusion

Special experiments were conducted to measure the variation of the pressure with the
length of the resin-bonded sand bed. A nonlinear relationship between the pressure
and airflow rate was established. Due to sand compaction, the pressure increased
with increasing the airflow rate for resin-bonded sands. It is shown that the pressure
decrease also increases nonlinearly with the amount of deposited resin-bonded sand
in the tube. Moreover, at certain values of the airflow rate (-250 cm3/sec) a signifi-
cant increase in pressure decrease was observed for resin-bonded sand, which con-
tributed to the complex rheological behavior of the test material. The plasticity of the
sand-binders mixture is characterized by a yield shear stress, which provides the sys-
tem with some resistance against compaction. However, further investigations and
analyses are needed to explain this kind of behavior in pressure decrease.
The results of experimental study and numerical simulations of resin-bonded
sand-airflow in two different core boxes are presented. A computer-controlled elec-
tronic system was designed and built to measure the pressure and airflow rate inside
the core boxes. A significant difference in pressure values was observed in the hop-
per, blow tube, and core boxes, which is related to the cross-sectional area changes
and transition of the system from the packed bed to the fluidized bed position. Con-
tour maps of the pressure distributions inside the core boxes are created based on
barometric measurements. High pressures are found in areas adjacent to the blow
tube. Low-pressures zones are created at the bottom of the core box and under the
H-shaped insertion of the core box. Good agreement was found between numerically
simulated and experimentally measured values of pressure along the vertical sym-
metric axis of the core boxes.
Pressure variations (in the hopper, blow tube, and core box) versus time during
airflow through sand-binders bed in the core boxes (gassing and purging cycles) was
studied experimentally. A significant pressure decrease was observed in the blow
tube and core box. However, pressure segregation across the core box is more pro-
nounced. High pressures ( -25 kPa) are found in the upper portion of the core box;
low pressures (-5 kPa) are measured in its lower sections. In contrast to the sand
mold-filling cycle, because no further significant sand compaction occurred during
the gassing and purging cycles, we did not observe any peaks on the pressure-time
curves obtained in the blow tube and core boxes.

Acknowledgments

The authors acknowledge the financial support of the Office of Life and Micrograv-
ity Sciences and Applications of the National Aeronautics and Space Administration
under Cooperative Agreement NCCS-128. We thank GM Powertrain for supplying
222 CHAPTER 8 Progress in an Industrial Application of Fluidized Beds

samples of Nugent sand and the ISOCURE LF-305/904G system. We gratefully rec-
ognize the technical assistance of Donald Sirois, Mike Crumpler, and John Marcell.
Finally, particularly helpful were discussions on the problems of the phenolic-
urethane-amine cold box process with Ken Williams of Flow Simulation Services,
Inc.; Dale Snider of Flow Analysis; David Goettsch, Sheila Palczewski, Mike
Walker, and John Siak of General Motor Corporation.

References

Ashland Chemical, Inc. 1994. "ISOCURE® LF-305/904 G," Foundry Products Division,
No. I, 1-3.
Bakhtiyarov S.I., Popov V .I., and Khabakhpasheva E.M. 1978. Fluid displacement from cav-
ity by spacer fluid. J. Oil Gas, 9, 17-20.
Bakhtiyav S.I., and Overfelt R.A. 1997a. Rheological and thermal properties of phenolic resin
and polymeric isocyanate and their blends. In: Proc. Rheology and Fluid Mechanics of
Nonlinear Materials. Eds. S.G. Advani and D.A. Siginer. FED-Vol. 243/MD-Vol. 78,
73-77. ASME International Mechanical Engineering Congress and Exposition. Dallas,
TX: United Engineering Center.
Bakhtiyarov S.I., and Overfelt R.A. 1997b. Study of Rheological Properties of ISOCURE®
LF-305/904G Binder System. J. Elastomers Plastics, 29(4), 314-325.
Bakhtiyarov S.I., and Overfelt R.A. 1998a. Rheological study of phenolic-urethane-amine
process. J. Elastomers Plastics, 30(1), 11-27.
Bakhtiyarov S.I., and Overfelt R.A. 1998b. Rheological and thermal characteristics of TECH-
NIKURE® Binder System used in core box process. J. Elastomers Plastics, 30(4),
328-339.
Bakhtiyarov S.I., and Overfelt R.A. 1999. Recent advances in the rheology of fluidized mate-
rials. In: Advances in the Flow and Rheology of Non-Newtonian Fluids, Part B, Rheology
Series 8, Eds. D.A. Siginer, D. De Kee and R.P. Chhabra et al., 1399-1433. Amsterdam:
Elsevier.
Bird R.B., Stewart W.E., and Lightfoot E.N. 1960. Transport Phenomena. New York: John
Wiley.
Carey P.R., and Sturtz G.P. 1977. Updating resin binder processes. Foundry Management
Tech., January.
Hoyt D. 1987. Back to the basics of silica sand. Modem Casting, September, 1-4.
Leva M. 1959. Fluidization, New York: McGraw-Hill.
Snider D.M., Goettsch D., Stahl K., and Couture DJ. 1999. An experimental and numerical
study of dense particle flow in sand core castings. Proc. 3rd ASME/JSME Joint Fluid En-
gineering Conference, San Francisco, CA.

Further Readings

Bakhtiyarov S.I., Overfelt R.A., and Reddy S. 1996. Study of the apparent viscosity of flu-
idized sand. In: Proc. Rheology and Fluid Mechanics of Nonlinear Materials-1996. Eds.
D.A. Siginer and S.G. Advani. AMO-Vol. 217, 243-249. ASME International Mechanical
Engineering Congress and Exposition. New York: United Engineering Center.
View publication stats

You might also like