. . . From the Electric Furnace Conference

Tundish Nozzle Blockage in Continuous Casting by G.

c. Duderstadt, R. K. Iyengar, and J. M. Matesa

In the continuous casting process, teeming through the tundish constitutes one of the more critical steps, since control of temperature and deoxidation during teeming affects both the internal and surface quality of the product. While the continuous casting of silicon-killed steel is approaching a standardized practice, casting of aluminum-killed steel has been limited because of constriction of the tundish nozzle. An illustration of the problem encountered is shown in Fig. 1.' In-

creasing aluminum additions as well as decreasing the temperature can thus lead to a condition which requires lancing. Note that the amount of aluminum that can be tolerated is less than required for production of fine grain steel. To permit the use of aluminum, two alternate casting methods which circumvent the problem of tundish nozzle blockage have been proposed. One method involves the addition of aluminum wire to the mold. However, this technique will find its limitation with respect to stringent cleanliness requirements unless the oxygen content is appreciably reduced prior to the aluminum addition. The second method utilizes the concept of an oversized tundish nozzle coupled with a stopper rod to vary the flow area, thus com-

G. C. DUDERSTADT, R. K. IYENGAR, and J. M. MATESA are, respectively, research supervisor, research engineer (formerly with Jones & Laughlin, Graham laboratory), and research engineer, Graham Laboratory, Jones & Laughlin Steel Corp., Pittsburgh, Po.

pensating for any nozzle restriction."a Because of the undesirable stream characteristics associated with this technique, submerged teeming has been proposed. In an attempt to establish a practical method for casting of aluminum-killed steel, a study was initiated to investigate the mechanism of nozzle blockage in tundish nozzles. The study included investigation of the role of aluminum content, steel temperature, nozzle material and diameter, reoxidation in the tundish, and heat transfer in the nozzle. Also investigated was teeming through stopper rod-controlled oversized nozzles to determine the practicability of this technique. It was hoped that knowledge gained from this study would pro-

2840 %

rn Q

z

::l

~ Z

LLI

0:: ::I

• •

00

2800

No Lancing 0

o

00

~

< 0:: LLI

0

o

~ 2760~_---'·

LLI ~

•

..J

W W

6

~

rn

6

6

6

I· 2 Ton Ladle

2· Tundish

6

2720L-____L -_ _ _ _~----~-----L-----J 0.4 0.6 1.4 ALUMINUM ADDED FOR DEOXIDATION LB/TON Fig. I-Effect of aluminum and tion.

ste~1

3· Tundish Overllow 4· Burner

5· Thermocouple 6· Shieldinl Device 7. Mold

temperature on casting operaFig . 2-Experimental set-up for tundish nozzle experiments. APRIL 1968, JOURNAL OF METALS-89

2950 u..

w

'"

::;, I-

«

'" w

c..

:;: w

I-

TEEMING TIME Imin.)

Fig. 3-Typical example of the change of the steel temperature in the tundish (V2-in. nozzle) as encountered in the experiments.

vide information for design of our future continuous casting facility. EXPERIMENTAL PROCEDURE

The experimental setup used in this study is shown in Fig. 2. Steel, similar to the composition of C-I020 Grade, was supplied from a two-ton electric furnace .. Deoxidation was accomplished in the ladle with ferrosilicon (0.2-0.3 % Si) and varying amounts of aluminum (0-0.068% AI",J). The steel was teemed into a one-strand tundish which was lined with an insulating layer of AP Green 28 and a working lining of Ram 90. The capacity of the tundish was 600 lb of steel, resulting in a metal head of 14 in. Normal preheat temperature of the tundish was 2250°F. At this metal head, the theoretical casting time was calculated to be 12 and 5 min. for nozzle diameters of 1f2 and 3f4 in., respectively. To establish standard casting conditions, the temperature of the steel was measured in the ladle as well

as in the tundish at regular time intervals. A typical example of the change of the temperature in the tundish is given in Fig. 3. After 10 min. of casting, the steel temperature in the tundish usually reached 2830°F. For analysis of chemical composition, oxygen, and inclusion content, steel samples were taken at the following stages: ladle, tundish, and tundish stream below the nozzle. To avoid oxidation of the stream during sampling, the space between tundish and ingot mold was enclosed and purged with nitrogen prior to the start of and during teeming. To document the effect of operating variables on the characteristics of the emerging tundish stream, photographs were taken at close intervals.

Table I. Experimental Data on the Relationship Between AI Content and Nozzle Blockade

AlsoL. %

0 lSi-killed) <0.004 0.005 0.008 0.016 0.036 0.048 0.056 0.068

EXPERIMENTAL RESULTS

Casting Time, min

Steel Cast, Ib

Max. Avg. Tundlsh Casting Temp., Speed, OF Ib/mln

O'-in. zircon nozzles) 3500 2900 12.0 17.0 17.0 18.0 4.3 3.8

2900 3200 2500 400 200

2900 2910 2905 2890 2910

(3,4-ln. zircon nozzles) 3300 2920 13.0 2920 8.5 1600 2920 1500 10.0

-300 _230 _230 _190 _ 95 _ 55

_280 _190 _150

constriction of the stream occurred more rapidly resulting in complete interruption at aluminum concentrations typical of fine grain steel. The analyses of samples taken from the tundish and emerging stream at various time intervals are summarized in Fig. 5. With aluminum greater than 0.025%, the amount of alumina and oxygen present remains almost constant independent of the aluminum concentration. Also, there is a considerable decrease in soluble aluminum from the tundish to tundish stream. However, there is no significant difference in the concentration of alumina and oxygen between the emerging stream and the tundish. This suggests that an appreciable amount of aluminum is oxidized in the tundish and the resulting alumina is either transferred into the slag or precipitated on the nozzle. Several nozzles were sectioned after clogging and prepared for

Role of Aluminum Content The aluminum content of the steel was gradually increased from 0% to 0.03% soluble aluminum to determine its effect on nozzle blockage. The results obtained from these experiments are shown in Table I. At low aluminum levels (less than 0.005% Al soluble), nozzle blockage did not occur. As the aluminum content was increased, the casting rate was considerably reduced and nozzle blockage was frequently encountered at 0.008% Al soluble. Above this, blockage occurred in all trials and in a relatively short time. For example, at 0.036% Al soluble, blockage occurred after 3.8 min. of casting, resulting in an average casting rate of 55 lb of steel per min., compared with a theoretical casting rate of 300 lb per min. The effect of nozzle blockage on stream characteristics is illustrated in Fig. 4. With increasing aluminum,

I

1I

m

.005 %AI Sol.

.008 %AI Sol.

.016 %AI Sol.

.036 %AI Sol. MINUTES o.

8 11131517

Fig. 4-Effect of aluminum content in steel on tundish stream characteristics (V2-in. zircon nozzle).

90-JOURNAL OF METALS, APRIL 1968

TEEMING TIME (MIN.) OTUNDISH, /j, TUNDISH STREAM

Fig. 5-Changes in oxygen, aluminum (soluble), and alumina concentration in tundish and tundish stream as a function of time.

Table II. Comparison of Experimental Data for Shielded and Open T undish

Casting Time, Al sol., % min

0.016 0.036 0.030 0.040 0 .056

Steel Cast, Ib

Max. Tundlsh Temp., OF

(Open Tundlsh, ~'-In. nozzles) 4.3 400 2890 3.8 200 2910

Avg. Casting Speed, lb/mln

-

95 55

(Shielded Tundlsh, ~'-In. nozzles) 9.3 l420 2900 _150 10.0 1140 2900 -115 (Open Tundlsh, %-In. nozzle) 8 .S 1600 2920

-190

(Shielded Tundlsh, %-In. nozzle) 0.064 10.5 3500 2900 _400

Fig . 6-Cross-section of a typical nozzle clog indicating the dendritic growth of metal from the nozzle wa II into the bore.

metallographic examination to establish the mechanism by which nozzle blockage occurs. A crosssection of a clogged nozzle illustrating the dendritic growth of metal from the nozzle wall into the bore is shown in Fig. 6. An enlargement of tre transition zone between nozzle wall and metal is shown in Fig. 7, indicatmg the formation of oxide branches close to the wall strengthened by freezing steel. A microscopic study of the nonmetallic branches revealed that they consist primarily of alumina embedded in a matrix of MnAISi-oxides (Fig. 8) . Furthermore, the area closer to the center of the nozzle was found to contain only pure alumina. While this was found to be typical for steel containing about 0.01 % AI, the transition zone of complex MnAISi-oxide branches tended to disappear when high Al concentrations (about 0.02/0.03%) were em-

ployed. In this case, the entire clog contained almost pure alumina, sometimes in large white pockets, with very little complex oxides present. This analysis, therefore, suggests that the mechanism of clogging is a complex process involving the precipitation of alumina in a MnAISi-oxide matrix on the nozzle reinforced by freezing steel. Furthermore, the directional growth of the steel dendrites indicates a heat loss through the nozzle wall which favors the build-up. Role of Reoxidation

As noted earlier, an appreciable amount of aluminum is lost in the tundish, while oxygen and alumina concentration remain nearly constant between tun dish and tun dish stream. To investigate the role of reoxidation and its effect on nozzle blockage, the tundish and ladle stream were completely enclosed

Fig. 7-Enlargement of the transition zone between nozzle wall (right side) and metal clog. SOX, reduced 2S%.

and the system purged with nitrogen prior to teeming and during the entire period of casting. The results obtained with the shielded tundish are summarized in Table 2 and illustrated in Fig. 9. During casting of steel containing 0.04% Al soluble through an open tundish, clogging occurred after about 4 min. Shielding the tundish extended the casting time to 10 min. and doubled the casting rate. However, clogging could not be avoided; while shielding reduces reoxidation, it does not provide a complete solution to the problem of nozzle blockage. Role of Nozzle Design

Earlier, it was noted that the directional growth of steel dendrites in the clog indicated a heat loss to the surrounding nozzle wall. In an attempt to minimize this heat loss, an upward protruding zirconia nozzle, shown in Fig. 10, was tested. This nozzle was supplied through courtesy of the Research and Development Dept. of the Atlas Steels Co. The design was based on the assumption that the molten metal surrounding the protruding part

Fig. 8-Photomicrograph of oxide branches consisting primarily of alumina embedded in MnAISi -oxide matrix. SOOX, reduced 2S%. APRIL 1968, JOURNAL OF METALS-91

0.07 0.06 0.05 0 .04

.

0.03 0.02

0.01 300 CASTING RATE IlB./MIN .)

Fig. 9-Effect of aluminum content and tun dish shielding on the casting rate for a V2-in. and 3,4-in. nozzle at a metal head of 14 in.

Fig. IO-Schematic drawing of a protrudingtype nozzle supplied through courtesy of Atlas Steels Co. Note that the portion containing the bore is above the bottom of the lining.

would supply sufficient heat to the bore to prevent the steel from freezing. The results obtained with this nozzle were rather discouraging since clogging still occurred. However, an examination of this nozzle design from the point of view of heat flow suggests that (1) the protruding part should have the highest thermal conductivity, and

(2) the bottom portion of the nozzle should be highly inSUlated. Based on this, a new nozzle was developed incorporating a highly conductive top (zircon, alumina, graphitized alumina, or magnesia) and a bottom portion of insulating refractory (Fig. 11) . A total of 12 experiments was conducted with this composite nozzle employing a standard aluminum addition of 1.7 lb/ton and a tundish preheat temperature of 2475°F. Under well controlled practice, it was possible to drain the tundish completely at 80 % of the theoretical casting rate without encountering nozzle blockage. Several important observations were made c'uring these trials. First, it was found that a high preheat temperature is essential for a successful cast. If the preheat temperature is low, e.g. around 2250°F, a reduced casting rate with subsequent clogging is encountered.

I~~-- 5"----.,·~1

7"

3"

1........t-----..;.,.5.8---t~~1

~

z

UJ

u

L&. L&. UJ

0

u UJ

(!J

a:: :z:::

<

u

en c

Fig. 12-Conical stopper and nozzle used for oversized nozzle studies. 92-JOURNAL OF METALS, APRIL 1968

1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.3 0.2 0.1 0 0

Fig. II-Schematic drawing of a protrudingtype nozzle made of highly conductive material at the top and highly insulating material at the bottom.

Second, the temperature of the steel in the tundish should remain above 2800°F to avoid blockage. Third, the nozzle materials studied do not show any significant difference in operating performance. Finally, although the casting stream was solid and steady, canting of the stream occurred from time to time, indicating temporary formation of oxide build-up and subsequent removal. Even though casting through composite, protruding type, metering nozzles achieved partial success, it was concluded that the above limitations would probably exclude such nozzles from further consideration, particularly in view of stringent requirements for good stream characteristics under plant conditions. The potential of a variable orifice nozzle for casting aluminumkilled steel was therefore investigated.

~--..-~

.

Pear Shaped System

v.............. c-:J v

_ ....... __ ......,. __ ...

~ ..

~

0

Conical

s~em

G

• Water Level-16" To Top Of Nozzle • Water Level-l0" To Top Of Nozzle 0.3

0.4

0.5

0.6

HEIGHT OF STOPPER HEAD FROM NOZZLE SEAT-IN. Fig. 13-Discharge coefficient vs. stopper position.

100

o~----------------------~ 0 .5

height was held constant for a certain period of time. In this case, the flow rate decreased continually to 15% of its original value within 5% min., as demonstrated in Fig. 16. Clogging would have shortly occurred if the stopper rod had not been raised. Typical flow characteristics of a stopper rod controlled nozzle are presented in Fig. 17. The occasionally large variations in casting rate for a given stopper height can be attributed to the formation of build-up and subsequent removal. Also, the large variations at small stopper heights are primarily due to the sensitivity of small flow areas to build-up. Although the trials with oversized nozzles have demonstrated that a solution of the specific problem of nozzle blockage during C3.sting of AI-killed steel has been found, the application of this technique to commercial casting of high quality billets depends on the successful solution of the following problems: (1) the poor stream characteristics require a shroud extending from the tundish into the mold ; (2) a control system must be developed to prevent wide variation in metal discharge rate; and (3) the matter of build-up and subsequent removal of alumina requires careful attention with regard to its effect on surface and internal quality of the product.

are fixed by operating considerations, a larger range of stopper movement is permissible if the discharge coefficient is low. Fig. 13 illustrates that such a low discharge coefficient is offered by the conical nozzle-stopper head system, which was subsequently selected for actual casting trials with steel of C-I018 composition containing 0.03-0.04% Al soluble. Stopper Rod Controlled Teeming

CASTING TIME (min .'

Fig . 14-Casting rate vs. time at various stopper heights-Fire cloy nozzle and plumbago stopper head .

OVERSIZED NOZZLE STUDIES Water Model Tests

To determine the best stopperrod-nozzle combination for control of flow, a model of a tundish equipped with an oversized nozzlestopper system was constructed and tested using water as the fluid medium. The three systems selected for study were: a rounded nozzle with a spherical stopper head, the same rounded nozzle with an inverted pear-shaped head, and a conical nozzle-stopper head combination. Results of these experiments indicated that the best flow control was obtained with the conical nozzle and a similarly shaped stopper head, such as shown in Fig. 12. The rate of metal discharged through an orifice system is given by the following equation: W =

p •

CD . A" . y'2gh

(1)

where W = lb of steel/sec p = CD = A N= h =

density lb/ft" discharge coefficient flow area, ft" metal head, ft

Since the available flow area increases with stopper height and since casting rate and metal head

~UlT. _ ·c .,,<-

~ ~.c.­

Average Casting Rate 4

The experimental tundish which was described earlier was modified by the installation of a load cell weighing system which provided a continuous recording of the weight poured during the casting period. A manually operated control device was installed on the tundish to manipulate the height of the stopper rod, thereby varying the flow area of the nozzle. Initial casting trials with a fire clay nozzle and plumbago stopper head indicated considerable erosion of the nozzle as demonstrated in Fig. 14. It can be seen that, while the stopper height was held constant, the flow was continually increasing. In commercial operation, such continued erosion of the nozzle would lead to a situation where control of the casting rate could become impossible. Therefore, further experiments were carried out with a plumbago nozzle which has higher wear resistance and higher refractoriness. Early trials with this system aimed at maintaining a constant metal flow rate of 300 lb/min indicated a restriction of the nozzle flow area similar to that encountered in regular metering nozzles. The fact that the flow continually decreases regardless of stopper head position is clearly illustrated in Fig. 15. An important observation is that different flow rates can be obtained for identical stopper head positions (compare positions A and B) indicating that build-up of nonmetallics and subsequent removal had occurred. To obtain a more positive indication of the build-up, the stopper

B

E

~~~

CONCLUSIONS

Examination of the phenomenon of nozzle blockage, encountered during casting of aluminum-killed steel, resulted in the following conclusions: 1. The mechanism of clogging is a complex process involving the precipitation of alumina in the bore of the nozzle augmented by dendritic growth of entrapped steel. 2. At low aluminum concentrations «0.01 % Al soluble), the alumina branches were found to be embedded in a matrix of MnAISioxides, while at higher concentrations (== 0.01 % Al soluble) the matrix tended to disappear with the

~Ulr-o, 800

t;~ :!:il!:

600

u

200


o~~

____

~

__________________

~~

0 .8 0.7 0.6 0.5

o CASTING TIME Imin .) Fig . IS-Casting rate vs. time at various stopper heights-Plumbago nozzle and stopper. Constant flow trial.

2

3

4

5

7

6

8

9

CASTING TIME Imin.) Fig. 16-Casting rate vs. time Plumbago nozzle and stopper.

at

various

stopper

heights-

APRIL 1968, JOURNAL OF METALS--93

400t-

300tto

ZW

.~

;::: .... E

",« "-

;3n:

~

200

+_ J~,I

J

lOOtO~~~

0.3

0.4

__~__- L__~~~ 0.5

0.6

0.7

STOPPER HEIGHT (in.)

Fig. 17-Effect of stopper height on casting rate for conical nozzle-stopper head system.

clog consisting only of alumina and frozen steel. 3. Although the tendency for build-up to occur was found even at low aluminum concentrations,

94--JOURNAL OF METALS, APRIL 1968

complete blockage of the nozzle was encountered only at concentrations ~ 0.008% Al soluble. 4. While reducing reoxidation of the steel in the tun dish through complete shielding decreased the rate of clogging, it did not completely eliminate the problem. 5. A new nozzle design with improved heat transfer characteristics eliminated blockage under strictly controlled operation. However, the stringent requirements for high tundish preheat (~ 2450°F) and a minimum steel temperature of 2800°F, together with the unresolved problem of canting of the stream, are believed to impose serious limitations under actual plant conditions. 6. The concept of a variable orifice for casting achieved by an oversized nozzle with stopper rod control represents a solution to the problem of nozzle blockage. However, application of this technique

to the casting of high quality billets depends on the significance of the following: (a) the practicability of a shrouded tundish stream ; (b) development of a control system to prevent large variations in casting rate: and (c) the effect of agglomerated alumina, washed out of the nozzle, on product quality. Acknowledgement

The authors are grateful to the Foundry personnel at Pittsburgh Works of the Jones & Laughlin Steel Corp., particularly to F. J. Smak, for cooperation in the experimental work and also to C. C. Gerding, J. M. Stoll, and K. Piene for helpful assistance in the design and execution of the experimental program.

REFERENCES 1

F. Tomsu and V. Dolezel: Steel Times,

Sept. 9, 1966. pP. 344-146

• A. Bungeroth and

E.

Scheufele:

Times, Sept. 18, 1964. pp. 402-403. 3 B. Tarmann: Hiirterei-Technische teilungen. 1962, Vol. 17. pp. 29-38.

Steel Mit-