5DWPaper2 PDF
5DWPaper2 PDF
5DWPaper2 PDF
T
s
= temperature of the part [K]
T
l
= temperature of the quenching liquid [K]
A = surface area of the part being quenched [m
2
]
V = Volume of Part [m
3
]
58
It can be seen that the lines of best fit for the four alloys all lie directly on each
other whereas the curves for pure aluminum deviate from the alloys at higher
temperatures. For this reason, it was assumed that the specific heat of 6061 aluminum
will follow a similar trend as the other alloys presented within the figure. In addition, the
specific heat of 6061 aluminum at room temperature is equal to 0.896 J/Kg K, which falls
directly on the line of best fit above. For these reasons, it will be assumed that 6061 will
follow similar trends at higher temperatures and this curve, with an equation of
Cp [J/KgK] = 0.007T + 0.644, will be used in our heat transfer coefficient calculations.
In addition, the density of the material can be determined as a function of
temperature. The density of the material is not constant because the volume of the part
will increase with increasing temperature. The linear coefficient of thermal expansion, ,
for 6061 aluminum is equal to 2.26 x 10
-6
/K [4]. Equation 2 illustrates the formula used
in calculating this variation at elevated temperatures.
The variation in density over the temperature range of interest is quite small. The
values range from 1.875 g/cm
3
to 1.94 g/cm
3
, which is a variation of only 1.7% about the
mean [4]. For this reason, it will be assumed that the density is constant and an average
value equal to 1.905 g/cm
3
will be used in our calculations.
m = mass of probe [g]
= Coefficient of thermal expansion [cm/cmK]
l
o
= Initial length of probe at room temp [cm]
d
o
= Initial diameter of probe at room temp [cm]
T = Change in temperature (T
part
- 273) [K]
( )( )
2
T d d T l l
m
o o o o
+ +
=
(2)
59
Quench Factor Analysis (QFA)
Quench factor analysis (QFA) provides a single value that quantitatively classifies
quench severity for a specific alloy [9-16]. QFA is an analysis of the cooling curve
associated with a particular quench coupled with a Time Temperature Property (TTP)
curve defined by Eq. 4 below [9-16]. The analysis begins with calculation of a variable
called the incremental quench factor (), which is performed for each time step in the
cooling process.
The C
T
function is defined below in Eq. 4 as well as the variables that help create
it [6, 9-11, 13, 14, 16, 17]
Where:
C
T
= critical time required to form a constant amount of a new phase or reduce the
hardness be a specific amount.
K
1
= constant which equals the natural logarithm of the fraction untransformed during
quenching (typically 99.5%: (ln (0.995)) = -0.00501
K
2
= constant related to the reciprocal of the number of nucleation sites
K
3
= constant related to the energy required to form a nucleus
K
4
= constant related to the solvus temperature
K
5
= constant related to the activation energy for diffusion
R = 8.3143 J/K mole
T = absolute temperature (K)
= incremental quench factor
t = time step used in cooling curve data acquisition
T
C
t
=
(3)
=
RT
K
T K RT
K K
K K C
T
5
2
4
2
4 3
2 1
exp
) (
exp
(4)
60
The values of the constants used in this research were found in literature and shown
below in Table 1.
K1 K2 K3 K4 K5
-0.00501 2.20E-19 5190 850 1.80E+05
Table 1: K-Constant Values for Quench Factor Analysis[12]
The incremental quench factor () above, represents the ratio of the amount of time
an alloy was held at a particular temperature divided by the time required for 1%
transformation at that given temperature. This value is then summed over the entire
transformation range to produce the cumulative quench factor (Q). The summation
equation is seen below.
In Equation 5, the values of T
1
and T
2
are taken as the maximum and minimum
temperature values, respectively, off of any Time-Temperature-Property (TTP) Curve. A
typical TTP Curve is seen below and the location and values of T
2
and T
1
can be readily
seen as well as how the C
T
function fits into the analysis [1, 13]. For the purpose of this
analysis, T
1
and T
2
are equal to 150C and 425C, respectively [10].
= =
2
1
T
T T
C
t
Q
(5)
61
Figure 3: Cooling Curve and TTP Curve Analysis[12]
Above, the dashed line represents the temperature of the alloy as a function of
time and the corresponding TTP Curve is seen to its right. The t value is taken as the
time interval in which data points are collected. From these curves, the quench factor
(Q) can be calculated as the summation of t/C
T
. The quench factor is proportional to
the heat removal characteristics of the quenchant as depicted in the cooling curve for the
quenching process [12]. The Q value can classify the severity of the quench for a
particular alloy. The cooling rate is inversely proportional to QFA, that is to say, the
greater the cooling rate, the smaller the quench factor, Q.
10
-2
10
-1
10
0
10
1
10
2
10
3
log (t) (sec)
TTP Curve
T
2
= 425
T
1
= 150
T ( C)
Cooling Curve
t
C
T
62
Experimental Plan
To determine the effect of bath temperature and agitation level on heat transfer
coefficients and QFA, the following test matrix was created. Table 2 lays out the
conditions to be tested. Each bath temperature will be tested with each corresponding
agitation level, so a total of 15 conditions will be explored. Five samples were quenched
at each test condition to determine the repeatability of the quench.
Table 2: Distilled Water Test Matrix
The agitation rate will be quantified by the rotations per minute of the immersed
impeller. The fluid velocity at the location of the probe could not be measured; therefore,
the Reynolds number could not be calculated to quantify the fluid flow. At 880 rpm, the
flow in the quench tank appears to be lamellar, whereas at 1850 rpm, the flow is highly
agitated and appears to be turbulent.
Water Temperature ( C)
Agitation
Rate
(rpm)
5 25 40 80 100
0 X X X X X
880 X X X X X
1850 X X X X X
63
Experimental Procedure
The test piece, a CHTE probe and coupling machined out of 6061 aluminum is
shown below in Figure 4 [18-21].
Figure 4: Probe-Coupling Dimensions
An ungrounded K-type thermocouple was embedded in the center of the probe, in
order to take temperature data at its geometric center. In addition, graphite powder was
placed in the center of the probe to maintain an excellent electrical contact between the
test piece and thermocouple. Initially, a grounded K-type thermocouple was used, but
created several data acquisition problems due to an electrochemical reaction between the
sheath of the thermocouple and the center of the aluminum probe. The switch to the
ungrounded thermocouple alleviated these problems and allowed for the collection of
smooth, continuous data.
In order to eliminate the possibility of water leaking into the probe center,
Resbond 989, High-Purity Alumina Ceramic, purchased from Cotronics Corporation, was
applied to the interface between the probe and coupling. Once applied, it was air
hardened for 4 hours then baked at 200C for 4 hours to insure proper curing and
decrease the risk of cracking.
1.50
2.50
3/8
Probe
Coupling
Thermocouple
24 threads/in
1/4 L 5/16 D
20 threads/in
1/4 L 1/4 D
64
This probe was used in conjunction with a CHTE probe quench system in order to
conduct the experiments. The CHTE quenching system is composed of a modified box
furnace, constant temperature bath tank and variable agitation unit [18-23]. Figure 5 is a
schematic of the system illustrating the location of each component and the nature of the
agitated fluid.
Figure 5: Schematic of CHTE Quench System
The probe is connected to a coupling and connecting rod, both machined out of
the same material, 6061 aluminum. The connecting rod is controlled by a pneumatic
piston, which lowers the probe from inside the furnace to the quenching bath. The probe
is solutionized to 530C before quenching in the distilled water quenching tank. The bath
Variable Agitator
Probe
Furnace
Baffle
Agitating
Rod
530 C
Connecting
Rod
Quench Tank
Quench Bath
Heating Elements
K- Type Thermocouple
65
is 12 liters in volume and its temperature is controlled by the variable heating elements
located on the tank floor.
The bath is constantly mixed by the use of the variable agitation rod, which forces
flow from the right side of the tank to the left side with the help of a baffle located in the
center. The baffle is an H-shaped bracket that acts as an exit gate for water to be forced
from the right side of the tank to the left and then a re-entry gate for water to return to the
agitation side. The opening of the baffle is aligned with the agitation propeller to allow
for a maximum degree of agitation flow.
Once the probe is lowered from inside the furnace, a switch triggers a data
acquisition to begin collecting time-temperature data from the embedded thermocouple at
a rate of 1000 data points per second. From start to finish, 40,000 data points are
collected. The data is then reduced to 1000 data points by a smoothing operation by
taking a running average of the data. Furthermore, the derivative (dT/dt) of the data is
taken in order to analyze the cooling rates associated with each quench condition. The
derivative is calculated by use of a four-point numerical method [24]. Equation 6 shows
the equation that was used focused around data point 3 in the time-temperature data
collected during the quench.
(6)
Once all experiments were completed, heat transfer coefficients and Quench
Factor Analysis was performed for each quench condition.
t
T T T T
dt
dT
+
=
12
8 8
5 4 2 1
t = time difference between data points
T = temperature at a given data point
66
Experimental Results: Heat Transfer Coefficients
Heat transfer coefficients were calculated using Equation 1. The data from each
of the five quench experiments was averaged and plotted below in Figures 6 through 8.
These three figures illustrate the change in h as a function of initial bath temperatures
for each of the three agitation levels. Figure 6 demonstrates that as the bath temperature
increased, the effective heat transfer coefficient decreases. The curves for 5C and 25C
are relatively similar in magnitude starting at 530C until 460C where the curve at 5C
achieves higher h values. In addition, as the temperature approaches 80C and 100C, a
Liedenfrost point becomes clearly visible on each plot indicating that the probe is being
alternatively covered with a vapor blanket and liquid layer which causes oscillating
surface temperatures [23]. Furthermore, at 100C, the Liedenfrost point is prolonged
much further before the maximum h value is achieved. As a result, the temperature at
which the maximum h value occurs decreases from 360C to 225C. The low
temperature gradient between the quenching medium and the hot 6061 probe may
stabilize film boiling before leading into nucleate boiling, thus causing this decrease in
temperature [23].
67
0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 100 200 300 400 500 600
Temperature (C)
'
h
'
[
W
/
m
^
2
K
]
5C
25C
40C
80C
100C
Figure 6: Effective Heat Transfer Coefficients for Various Temperatures With No
Agitation
Figures 7 and 8 show similar trends that were observed in Figure 6, except the
presence of the Leidenfrost point does not appear in the curves for 80C. It is believed
that since agitation is now present, the fluid flow mechanically broke down the vapor
layer and promoted nucleate boiling. As a result, film boiling could not become stable
and, therefore, does not appear on either curve [25].
68
0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 100 200 300 400 500 600
Temperature (C)
'
h
'
[
W
/
m
^
2
K
]
5C
25C
40C
80C
100C
Figure 7: Effective Heat Transfer Coefficients for Various Temperatures With An
Agitation Level of 880 rpm
The final noteworthy aspect of these three figures is the maximum heat transfer
coefficient that is attainable for each condition. At an agitation rate of 0 rpm, values that
range from 1100 W/m
2
K to 2100 W/m
2
K can be achieved whereas at agitation rates of
880 and 1850 rpm, ranges of 1500 W/m
2
K to 2800 W/m
2
K and 2000 W/m
2
K to 3850
W/m
2
K can be achieved, respectively
69
0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 100 200 300 400 500 600
Temperature (C)
'
h
'
[
W
/
m
^
2
K
]
5C
25C
40C
80C
100C
Figure 8: Effective Heat Transfer Coefficients for Various Temperatures With An
Agitation Level of 1850 rpm
Experimental Results: Quench Factor Analysis (QFA)
Using Equation 5, Quench Factor Analysis (QFA) was performed. Table 3
presents the Quench Factor, Q, for each quench condition. Looking down the columns of
the table, it can be seen that Q decreases with increasing agitation level, indicating that
the quench rate increases with increasing agitation. In addition, Q increases with
increasing bath temperature as seen by the increasing values across the rows of the table.
Overall, the fastest quench rate was observed at 5C with an agitation level of 1850 rpm
and the slowest at 100C with no agitation. The Q values for these conditions are equal
to 3.78 and 40.6, respectively.
70
Table 3: Variation in Quench Factor as a Function of Agitation and Bath
Temperature
Figure 9 plots the data presented in Table 3 against the corresponding maximum
heat transfer coefficient seen in Figures 6 through 8. A hyperbolic curve was fit to the
data, but looking at the R
2
value, the relationship does not appears to be hyperbolic since
R
2
= 0.49. Overall, there is a general trend of higher maximum heat transfer coefficients
at lower values of Q. The greatest value of maximum heat transfer coefficient was
obtained at the same condition as the lowest Q value from Table 3. This data is
consistent with theory since it is understood that low temperature and high agitation
should yield high rates of heat transfer and low values in QFA [1].
Figure 9: Hyperbolic Regression of QFA as a Function of h Max
Temperature C Agitation Level
(rpm)
5 25 40 80 100
AG0 0 6.82 8.31 8.26 12.5 40.6
AG1 880 5.27 6.45 5.86 7.51 32.97
AG2 1850 3.78 4.67 5.21 6.18 29.52
'h' max [W/m^2 K]
1000 2000 3000 4000
Q
F
A
(
Q
)
0
10
20
30
40
50
Max h
QFA
' ' 1 . 336
) 1 . 338 ( * ) 6 . 173 (
8 . 12
+
+ =
49 . 0
2
= R
71
Figure 10 presents a second method of analyzing the data in Figure 9. The figure
below presents the correlation between these two parameters in terms of a linear
regression. The data at the bottom of the figure can be fit with a linear regression curve
with a high correlation coefficient. The line of best fit is not representative of the data as
a whole because the labeled data points to the left are not included in the linear
relationship.
y =-0.002x +11.238
R
2
=0.9058
0
5
10
15
20
25
30
35
40
45
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Maximum Heat Transfer Coefficient [W/m^2 K]
Q
F
A
(
Q
)
No Liedenfrost Temp
Liedenfrost Temp Exhibited
Linear (No Liedenfrost Temp)
100 C AG0
100 C AG1
100 C AG2
80 C AG0
Figure 10: Quench Factor As a Function of Maximum Heat Transfer Coefficient
72
Upon further investigation, the eliminated data points were found to be the
conditions with a Liedenfrost point present in their heat transfer coefficient curves. It
appears that data with this point present cannot be compared to data without it. One
explanation is that since the Liedenfrost point drives the temperature at which maximum
heat transfer occurs to a lower temperature, the Q calculation in the Quench Factor
Analysis is not summing the values at h max, but rather at the temperature of the
Liedenfrost point. As a result, these values are quite high and will not correlate well with
the data collected from curves without a Liedenfrost point. Figure 11 illustrates the
regions of interest on the heat transfer coefficient curve.
Figure 11: Summation Region of Q as Seen on Heat Transfer Coefficient Curves
The dashed lines represent the summation region in the QFA calculation [9].
Looking at the curve for 5C, it is clear to see that the summation region is focused
mainly on the location of maximum heat transfer, whereas at 100C, the region is focused
around the location of the Liedenfrost point. As a result, the majority of the summation
values are coming from a region of very low heat transfer and not from the location of the
0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 100 200 300 400 500 600
Temperature (C)
'
h
'
[
W
/
m
^
2
K
]
5C
100C
Summation Region
in QFA
73
maximum heat transfer coefficient. In addition, if the stages of quenching are
superimposed onto the curves above, it can be seen that the summation region at 5C is
mainly nucleate boiling, whereas at 100C, the region encompasses film boiling and
nucleate boiling. For these reasons, it is concluded that conditions with a Liedenfrost
point cannot be directly compared to those without by means of QFA and h
relationship.
Conclusions
Based on the experimental data presented above, it can be concluded that:
An increase in agitation causes an increase in maximum heat transfer coefficient
and a decrease in Quench Factor, Q.
A decrease in bath temperature causes an increase in maximum heat transfer
coefficient and a decrease in Quench Factor, Q.
When a relationship between QFA and h max is desired, conditions with a
Liedenfrost point cannot be directly compared to those without due to the location
of the summation region used in Quench Factor Analysis.
References
1. ASM, Quenching and Distortion Control, ed. H.E. Boyer. 1998, Metals Park
Ohio: ASM International.
2. Mackenzie, D.S. and G.E. Totten, Aluminum Quenching Technology: A Review.
Materials Science Forum, 2000. Vol. 331-337: p. 589-594.
3. Hasson, J. Quenchants: Yesterday, Today, and Tomorrow. in Proceedings of the
First International Conference on Quenching and Control of Distortion. 1992
September. Chicago, Illinois.
4. ASM, ed. Aluminum and Aluminum Alloys: ASM Specialty Handbook., ed. J.R.
Davis. 1993, ASM International.
5. Mills, A.F., Heat Transfer. 2nd Ed ed. 1999, Upper Saddle River, NJ,: Prentice
Hall.
6. Bates, C.E., G.E. Totten, and N.A. Clinton, Handbook of Quenchants and
Quenching Technology. 1993, Materials Park, Ohio: ASM International.
7. Brimacombe, J.K., S.M. Gupta, and E.B. Hawbolt. Determination of Quench
Heat-Transfer Coefficients Using Inverse Methods. in Proceedings From the First
74
International Conference on Quenching and Control of Distortion. 1992
September. Chicago, Illinois.
8. University, P., Thermophysical Properties of Matter. Specific Heat: Metallic
Elements and Alloys, ed. Y.S. Touloukian. Vol. 4. 1970, New York: Purdue
Research Foundation - Thermophysical Properties Research Center (TPRC).
9. Bates, C.E., L.M. Jarvis, and G.E. Totten, Use of Quenching Factor for
Predicting the Properties of Polymer Quenching Media. Metal Science and Heat
Treatment, 1996. Vol. 38(No. 5-6): p. 248-251.
10. Bates, C.E., T. Landig, and G. Seitanikis, Quench Factor Analysis: A powerful
Tool Comes of Age. Heat Treating, 1985(December): p. 13-17.
11. Bates, C.E. and G.E. Totten. Application of Quench Factor Analysis to Predict
Hardness Under Laboratory and Production Condition. in Proceedings of the
First International Conference on Quenching and Control of Distortion. 1992
September. Chicago, Illinois.
12. Bates, C.E., G.E. Totten, and L.M. Jarvis, eds. Quench Factor Analysis: Polymer
vs. Hot Water., ed. I. Tenaxol. 1994: Milwaukee, Wisconsin.
13. Bates, C.E., G.E. Totten, and G.M. Webster, Cooling Curve and Quench Factor
Characterization of 2024 and 7075 Aluminum Bar Stock Quenched in Type 1
Polymer Quenchants. Heat Transfer Research, 1998. Vol. 29(No. 1-3).
14. Bernardin, J.D. and I. Madawar, Validation of the Quench Factor Technique in
Predicting Hardness in Heat Treatable Aluminum Alloys. International Journal of
Heat and Mass Transfer, 1995. Vol. 38: p. 863-873.
15. Staley, J.T., Effect of Quench Path on Properties of Aluminum Alloys. JTS Alcoa
Laboratories.: p. 396-407.
16. Staley, J.T., Quench Factor Analysis of Aluminum Alloys. 1987, The Institute of
Metals: London, England. p. 923-934.
17. Bates, C.E., G.E. Totten, and G.M. Webster. Aluminum Quenching with Polymer
Quenchants: An Overview. in 17th ASM Heat Treating Society Conference
Proceedings Including the 1st International Induction Heat Treating Symposium.
1997 September. Indianapolis, Indiana.
18. Maniruzzaman, M., J.C. Chaves, and R.D. Sisson. CHTE quench probe system - a
new quenchant characterization system. in 5th International Conference on
Frontiers of Design and Manufacturing. 2002. Dalian, China.
19. Maniruzzaman, M. and J.R.D. Sisson. Investigation of bubble nucleation site
density during quenching heat treatment process using video imaging. in TMS
Annual Meeting 2002. 2002. Seattle, WA.
20. Sisson, R.D., J.C. Chaves, and M. Maniruzzaman. The Effect of Surface Finish on
the Quenching Behavior of 4140 Steel in Mineral Oils. in Proceedings of the 21st
Heat Treating Society Conference. 2001. ndianapolis, Indiana.
21. Fontecchio, M. Effect of Bath Temperature and Agitation Rate on the Quench
Severity of 6061 Aluminum Quenched in Distilled Water. in 13th Annual
International Federation
for Heat Treatment & Surface Engineering (IFHTSE) Congress. 2002. Columbus, Ohio.
22. Chaves, J.C., M. Maniruzzaman, and J.R.D. Sisson. A New Quench
Characterization System for Steels. in Proceedings of the 21st Heat Treating
Society Conference. 2001. Indianapolis, Indiana.
75
23. Maniruzzaman, M. and R.D. Sisson. Bubble Dynamics During Quenching of
Steel. in ASM Heat Treating Society. 2001. Indianapolis, Indiana.
24. Chaves, J.C., The Effect of Surface Finish Conditions and High Temperature
Oxidation on Quenching Performance of 4140 Steel in Mineral Oil. 2001
November 20, Worcester Polytechnic Institute.
25. Han, S.W., G.E. Totten, and S.G. Yun. Continuously Variable Agitation in
Quench System Design. in Proceedings From the First International Conference
on Quenching and Control of Distortion. 1992 September. Chicago, Illinois.