Advanced Materials Research Online: 2013-08-30

ISSN: 1662-8985, Vols. 753-755, pp 1857-1861

doi:10.4028/www.scientific.net/AMR.753-755.1857
© 2013 Trans Tech Publications, Switzerland

Application in the Experimental Data Analysis of the Slag Viscosity with

Minitab Software
Shuqing ZHANG 1, a, Xiuling Lv 1, Lihong ZHANG2, Nahui ZHANG
1.
HEBEI UNITED UNIVERSITY, China
2.
TANGSHAN SPECIAL EQUIPMENT SUPERVISION ANDINSPERVISION, China
a
[email protected]

Key words: Minitab, slag viscosity, basicity, chemical composition

Abstract: The single factor variance of viscosity experimental data of blast furnace slag was
calculated by Minitab software. The influence of slag basicity and chemical compositions on the
slag viscosity was obtained. The results show that there have insignificant effects on the slag
viscosity when the slag basicity is over than 1.02, MgO content is less than 11.95% or TiO2 content
is less than 12.57%. While the slag basicity can be influenced significantly with the slag basicity
less than 1.02, MgO content over than 13.95% and TiO2 content over 13.57%. The Al2O3 content of
the slag should be less than 12.75%.

Introduction
The suitable viscosity of BF slag is one of the key factors to affect the BF smelting process.
Especially in recent years, the quantity of iron ore inward continually increase and the Al2O3
content in the ore is generally on the rise. The present state of raw materials make the slag viscosity
rise and desulfurization ability decrease [1]. Based on the viscosity determination of ChengGang BF
slag[2], the viscosity data obtained are analyzed by using the Minitab software in order to find the
relationship between these data and obtain the influence of various factors on slag viscosity further.

Calculation method
The Minitab software consists of data processing, graphics and analysis process which have the
characteristics of simple, direct and clear. At present, it has been widely used in process quality
control, experimental data analysis, etc [3-4].
When using Minitab software to analyze data, the statistical test conclusion is the P (Probability)
values. The P value is inversely proportional to the reliability of result. The smaller the P value, the
confidence of results is higher, namely. In many research fields, P is equal to 0.05, which is often
considered as acceptable error to the greatest extent. Generally, the P value is set before starting the
analysis, then P is compared with the level value to determine significance: if the P is less than or
equal to the level, the original hypothesis will be denied, and the alternative hypothesis can be
accepted; If the P value is greater than the level, the original hypothesis cannot be denied [5].
The variance analysis in Minitab is used to analyze the relationship between the response variable
and one or more predictive variables, and to model for the above relationship. In the analysis of
variance, four different methods are often used to provide confidence interval of the difference
between the average which are Tukey, Fisher, Dunnett and MCB,respctively. The MCB method is
used to calculate the confidence interval in this paper.

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans
Tech Publications, www.ttp.net. (ID: 132.239.1.231, University of California, San Diego UCSD, La Jolla, USA-29/02/16,05:55:33)
1858 Materials Processing and Manufacturing III

Results and discussions

Variance analysis of viscosity and basicity
At the condition of constant MgO, Al2O3 and TiO2 content, 5 groups of experiments were done to
get the η-t curve of the slag, the basicity of which were 0.92, 1.02, 1.12, 1.22 and 1.32, respectively.
The effect of basicity on slag fluidity was determined by variance analysis to find a turning point
that makes slag liquidity sharply decline with the increase of slag basicity.The single factor method
is applied to analyze the slag viscosity (µ) of 5 groups data with different basicity value and the
results are shown in Table 1.

Table 1 Single factor test results of slag viscosity with different basicity Pa·s
Temperature/℃ First group Second group Third group Fourth group Fifth group
1 460 0.705 0.654 0.691 0.744 0.760
1 455 0.712 0.675 0.708 0.744 0.790
1 450 0.718 0.688 0.716 0.762 0.801
1 445 0.760 0.699 0.727 0.794 0.805
1 440 0.777 0.706 0.740 0.820 0.816
1 435 0.777 0.710 0.751 0.844 0.827
1 430 0.783 0.729 0.764 0.872 0.839
1 425 0.783 0.740 0.773 0.893 0.842
1 420 0.826 0.766 0.781 0.908 0.867
1 415 0.863 0.792 0.792 0.91 0.885
1 410 0.887 0.796 0.818 0.912 0.900
1 405 0.945 0.809 0.835 0.921 0.932
1 400 1.006 0.824 0.850 0.945 0.939
1 395 1.023 0.839 0.870 0.995 0.949
1 390 1.04 0.861 0.889 1.003 0.980
1 385 1.081 0.880 0.906 1.019 0.981
1 380 1.113 0.898 0.919 1.021 1.000
1 375 1.161 0.915 0.939 1.068 1.019
1 370 1.219 0.939 0.962 1.051 1.031
1 365 1.273 0.956 0.980 1.081 1.044
1 360 1.303 0.962 1.003 1.129 1.075
1 355 1.347 0.982 1.027 1.146 1.105
1 350 1.381 1.003 1.044 1.183 1.126
1 345 1.394 1.034 1.077 1.213 1.148
1 340 1.405 1.064 1.098 1.245 1.178
1 335 1.107 1.096 1.131 1.267 1.228
1 330 1.409 1.122 1.165 1.306 1.265
1 325 1.457 1.152 1.204 1.336 1.308
1 320 1.506 1.187 1.241 1.353 1.355
1 315 1.610 1.228 1.277 1.388 1.392
1 310 1.700 1.277 1.314 1.426 1.420
1 305 1.746 1.331 1.351 1.472 1.474
1 300 1.808 1.383 1.388 1.521 1.515
1 295 1.871 1.439 1.429 1.569 1.571
1 290 1.933 1.504 1.470 1.621 1.623
1 285 2.009 1.597 1.493 1.664 1.685
 Advanced Materials Research Vols. 753-755 1859

The data that is higher than the melting temperature in η-t curve is selected to analyze. The
original hypothesis:µ1=µ2=µ3=µ4=µ5，and the alternative option hypothesis: µ1, µ2, µ3, µ4, µ5
are not all equal. The variance analysis results of slag viscosity with different basicity by Minitab
software are calculated, the results of which is DOF =4, SS =1.3068, MS =0.3267, F =3.52, P =
0.009. The value of P is 0.009，less than 0.05 significantly, so the original hypothesis is rejected
and the alternative hypothesis is accepted. These shows there are significant differences among 5
kinds of slag with different basicity levels. The MCB method is used to analyze the variance of
different slag.
The smaller the viscosity value, the better the slag fluidity. Therefore, it thinks that the minimum
is the best choice when comparing. The overall error rate is 0.05, and the critical value is 2.18 after
calculated by MCB method of Minitab. The viscosity average of 5 groups is shown in Table 2.
Table 2 Average value and standard deviation of slag viscosity of 5 groups Pa·s
Groups Sample points Average value Standard deviation
1 36 1.230 5 0.408 1
2 36 0.998 6 0.278 9
3 36 1.017 1 0.253 0
4 36 1.131 2 0.278 0
5 36 1.113 8 0.281 2
It can been seen from Table 2 that the average value of second group is the minimum, then
Minitab will compare average value of the first, third, 4th, and 5th group with that of the second
group. The confidence intervals gained by average value of the first, third, 4th and 5th group
minusing that of the second group is shown in Table 3. Table 3 shows that the slag liquidity has
obvious difference between the first group and the other four groups. All confidence intervals of the
first group are greater than zero. Therefore, the slag liquidity of first group is the worst. But it can't
show the slag liquidity of other four groups has significant differences, because of the confidence
interval containing 0 (no difference).
It indicates that when slag basicity is less than or equal to 0.92, the change of slag basicity has
remarkable effect on slag fluidity; while it has not obvious influence on slag liquidity when slag
basicity is great than or equal to level 2.
Table 3 Confidence intervals gained by average value of the first, third, 4th and 5th group
minusing that of the second group
level lower limit center Upper imit
1 0.000 0 0.231 8 0.386 0
2 -0.172 6 -0.018 4 0.135 8
3 -0.135 8 0.018 4 0.172 6
4 -0.021 7 0.132 5 0.286 7
5 -0.039 0 0.115 2 0.269 4

Variance analysis of viscosity and chemical composition

At the condition of constant basicity, Al2O3 and TiO2 content, 5 groups of experiments is done to
get the η-t curve of the slag, the MgO content of which are 6.0%,9.95%,11.95%,13.95% and
15.95%, respectively. The effect of MgO content on slag fluidity were obtained by variance
analysis to find a turning point that makes slag liquidity sharply decline with the increase of MgO
content.
1860 Materials Processing and Manufacturing III

According to the results of slag viscosity, it calculates the variance of samples with different
MgO content, in which DOF =4, SS =7.4538, MS =1.8634, F =22.87, P = 0. It is seen from above
data that the value of P is less than 0.05 significantly, which indicates that 5th group slag have
significantly difference. The variance of slag with different MgO content is analyzed by MCB
method. It obtained that the overall error rate is 0.05, and the critical value is 2.17. The center
valure of confidence intervals in first group is 0.5885, second group 0.2286, third group 0.2069, 4th
group -0.1248 and 5th group -.01248. The above data presents that there are significant difference
for the slag liquidity between 1~3 groups and 4~5 groups. All confidence intervals of 1~3 groups
are greater than zero, so the slag liquidity is worse than that of 4~5groups. It indicates that when the
MgO content is less than or equal to 11.95%, the change of MgO content has insignificant effect on
slag fluidity; while it has obvious influence on slag liquidity when slag basicity is great than or
equal to level 4.
When the slag has lower MgO content, improving the content of MgO can make part of
Si x O yz − in slag depolymerize forming complex compound with low melting point such as
monticellite. Therefore, the melting point of the slag is reduced with the increase of MgO content.
The slag liquidity of 4th group is the best and its confidence interval contains not only negative
value but also the minimum average. Therefore the MgO content in slag is not higher than lever 4,
that is 13.95%.
According to the above calculation method, single factor variance analysis of slag viscosity with
different Al2O3 and TiO2 content is done respectively by using Minitab software. The results show
that the suitable content of Al2O3 is 12.75%, and it should not exceed 14.75%. When TiO2 content
is less than 12.57%, it has no significant effect on the slag liquidity. While the TiO2 content
significantly affects the slag liquidity when it is more than 13.57%, the conclusion of which is
consistent with that of literature 4.

Conclusions
Single factor variance analysis of experimental data of slag viscosity was carried on using Minitab
software, which verifies the correctness of using Minitab software to process a large quantity of
experimental data.
When slag basicity is less than 0.92, the change of slag basicity has significant effect on slag
liquidity; When slag basicity is greater than 1.02, it has a little effect on slag fluidity. When w(MgO)
is less than 11.95%, it affects the slag liquidity insignificantly; while it is greater than 13.95%, the
change of MgO content would seriously affect the liquidity of slag. The suitable content of Al2O3 in
the slag is 12.75%, the content of which should not exceed 14.75%. The change of TiO2 content has
insignificant effect on slag liquidity when w(TiO2) is less than 12.57%, but it obviously affects the
slag liquidity with w(TiO2) greater than 13.57%.
 Advanced Materials Research Vols. 753-755 1861

References:
[1] Xiayu HU, Liquan LIN, Liaosha LI. Laboratory Research on Property of Higher Al2O3 Content
Slag[J]. China Metallurgy, 2006, 16(2)：36-41(in Chinese).
[2] Hongwang MU, Shuhui ZHANG, Lü Qing, Yanqin SUN. Experimental Study on Viscosity of
Blast Furnace Slag with Medium Titania and High Alumina Content under Neutral Atmosphere
[J]. Iron & Steel, 2012, 47(6)：18-21, 50(in Chinese).
[3] Hailiang WANG, Derong ZHU, Bin GUI,etc. MINITAB Soft Application in Quality Control of
Process Control[J]. Coal Mine Machinery, 2006, 27(10)：114-116(in Chinese).
[4] Zhanfei MU.Study on Improving Reduction Disintegration Index of the Low-temperature
Sinter[D].Shenyang: Northeastern University, 2008 (in Chinese).
[5] Chi ZHANG. Minitab: Six Sigma solution [M]. Guangzhou: Guangdong Economic Publishing
House, 2003(in Chinese).
Materials Processing and Manufacturing III
10.4028/www.scientific.net/AMR.753-755

Application in the Experimental Data Analysis of the Slag Viscosity with Minitab Software
10.4028/www.scientific.net/AMR.753-755.1857