Editor’s Note: To read this paper’s abstract, references, and acknowledgements, please go to the AFS Library and search for paper number 24-090-07.

Higher CSR coke has higher hot strength implying that coke with those properties will travel farther down the cupola shaft, which can enhance metal temperature and carbon pick up. Furthermore, the stronger coke in the presence of heat and CO2 will better support the burden in the cupola.

The researchers discovered a technical paper presented in 1982  on measuring different coke reactivity effects in cupola melting. At the time, blast furnace coke (BF coke) was higher in CRI than foundry coke. Since then, BF coke was reformulated to have lower CRI and higher CSR than foundry coke. CRI of typical foundry coke today is in a range of 32 to 43 and CSR a range of 7 to 12. CRI of typical BF coke today is less than 25 and CSR is greater than 60.
Hypothesis: Lower CRI/higher CSR coke will perform better than higher CRI/lower CSR coke, which is currently produced in the U.S. for cupola furnaces.

A test in a production cupola was proposed to substitute foundry coke in progressively higher amounts with low CRI/high CSR coke of appropriate size for the cupola (4-in x 6-in) or (100mm x 150 mm). If serious operational issues were encountered the test would be terminated. Otherwise, progressive substitutions would continue to full substitution with 100% low CRI/high CSR coke. The cupola would continue with the test coke until the supply was exhausted or operational issues that could not be corrected were encountered.

Coke Discussion
Definitions

Coke Reactivity Index (CRI)––percentage weight loss of a coke sample heated for 120 minutes at 2012F (1100C) in the presence of CO2.

Coke Strength After Reaction (CSR)––percentage weight of the coke sample left in a drum after tumbling that is greater than 3/8 in (9.525 mm).

CRI/CSR Measurement History

The initial research for evaluating the effect on coke reactivity as it relates to coke strength in the blast furnace was done by Nipon Steel in the 1970s. Early in the 1980s, a procedure was agreed to per the ASTM procedure as outlined in the definition section. The heating cycle time soaking the coke in the presence of CO2 at 2012F (1100C) for 120 minutes was established. 

Since typical cupola dwell time is approximately 60 minutes, the authors questioned if the ASTM procedure is relevant specifying 120 minutes. A research study was conducted in 1980 comparing different heat cycle times for different coke samples. Included at the end of this paper are two graphs (Figs. 18 and 19) from this study (Reference 4), for CRI and CSR. Both show a linear relationship that proves that the test soak time of 120 minutes is relevant to cupola melting. 

Of interest but not a claim in cupolas, based on coke testing in blast furnaces, there is a rule of thumb that for every 1% increase in CSR, a 3-lb. reduction of carbon from the coke up to 55 CSR can be realized, thereafter, a 1% increase in CSR can result in a 1-lb. reduction of carbon from the coke.

Laboratory Tests

Comparing the analysis in Table 1, coke ash is higher for both +4 in. BF coke and typical BF coke. It is important to note that the same coke producer made the foundry and BF coke. The higher volatile content of the +4in. BF coke was perplexing. Before proposing the possible reason for the higher volatile content, it is important to understand the method used to produce these cokes.

The traditional method for producing BF coke is in a byproduct slot oven battery. Normal BF coke produced in a slot oven has a top size of 2 in. (50 mm). The supplier of the +4 in. BF coke produces coke in a beehive type oven. This method produces BF coke top size that is larger. Through extensive efforts, the coke producer was able to scalp the coke larger than 4 in. (100 mm) for this cupola experiment.

One possible reason for the higher volatile in the +4 in. BF coke is that the larger coke origin could be from the center of that oven. If so, perhaps the coal was not fully degasified, hence the higher volatile content in the coke. The other possibility is error in the sampling of the coke. In any event it does not appear that higher volatile content negatively affected the performance of +4 in. BF coke in the cupola.
A coke strength analysis is provided in Table 2. The CRI for +4 in. BF coke is lower than foundry coke but not as low as typical BF coke produced by the supplier. Measuring CSR, which is a measure of hot strength shows much higher for the +4 in. BF coke but not as high as normal BF coke.

Stability and Hardness indexes show the mechanical and abrasion strength of the cokes. The results show a remarkable improvement compared to foundry coke. The test results comparison of the +4 in. BF coke were very close to the normal BF coke. The porosity of the cokes was also measured as it is a function of strength. The +4 in. BF coke was slightly less than the foundry coke. Per discussion with experts in making BF coke, changing the CRI of the coke has no bearing on the porosity of the coke.

A coke ash mineral analysis is presented in Table 3.

The formula from the AFS “Cupola Handbook” 6th edition (Chapter 8, page 17) was used to calculate the catalytic index based on the reported mineral content of the ash. Experts in the making of low CRI coke state that the coke ash mineral analysis will be less basic than higher CRI coke. This test report concurs and shows that the BF coke ash mineral analysis is less basic than the foundry coke used in the cupola trial. It appears that during the substitution with BF coke in the cupola, had the limestone in the charge been reduced along with the reduced BF coke charges, it may have prevented an increase in cupola slag basicity that occurred during the testing.

Coke Ash Fusion

The relationship between the ash composition and the ash softening temperature (Ts) is shown in Table 4 with calculations.

The Cupola Handbook (Chapter 8) states that the lower the ash softening temperature the better the carbon pick- up from the coke.

The +4 in. BF coke Ts was 113F (62.8C) higher, but this cupola test showed a higher increase of carbon than foundry coke.

Another test in a hot blast cupola may or may not corroborate these results.

Coke ash fusion temperatures for initial, softening, hemispherical, and fluid states are shown in Table 5. Tests were run in a laboratory furnace, both in oxidizing and reducing atmospheres due to both conditions existing in the cupola.

The foundry coke ash softening temperature in an oxidizing atmosphere temperature is 108F (60) higher and 102F (56.7C) lower in a reducing atmosphere as compared to the formula calculation (Ts).

The difference for the BF coke was 180F (100C) higher in an oxidizing atmosphere and 79F (14.4C) lower in a reducing atmosphere as compared to the formula calculation (Ts).

Also, it is interesting to note that all temperature results were higher in an oxidizing than in a reducing atmosphere. Furthermore, for every test state, the +4 in. BF coke ash fusion temperatures were higher than the foundry coke. This clearly supports the theory that lower CRI coke has higher ash fusion temperatures.

A graph from the “Studies of a Quenched Cupola Part IV: Coke Behavior” paper is shown in Figure 1. This graph shows the coke temperature from a cupola computer simulation program. In addition, it shows L002 (average carbon stacking height) or the degree of graphitization. Coke enters the cupola as amorphous carbon. As the coke descends, it increases in temperature and starts graphitizing at the coke/liquid metal interface above the tuyeres. The authors added to this graph test results of coke ash fusion liquid temperatures in a reducing atmosphere depicted with heavy arrows. Comparing the two, BF coke was 150 (83.3C) higher than foundry coke. The ash fusion temperature in an oxidizing atmosphere was not plotted because the laboratory furnace was only capable of measuring a maximum of 2700F (1482C).

The authors believe the ash fusion liquid temperature is an important consideration. The theory is that perhaps the ash may insulate the carbon inside the coke until the ash is released when it becomes liquid thus exposing the carbon at a higher temperature. Furthermore, the higher coke temperature transforms more carbon from amorphous to graphitic carbon.

Experimental Trials

Coke Sampling Procedure: Five-gallon bucket samples of coke were collected randomly for the duration of the tests for both the foundry and +4in. BF coke. The buckets were then emptied into three 55-gallon (208 liter) drums each for foundry and BF coke. The total weight of each coke sample was approximately 600 lb. (272 Kg).

The original test sequence was to obtain base line data on the first day with foundry coke only. The second day was to start blending 30% +4 in. BF coke working towards 50%. The third day was to start at 50% BF coke and work towards 100% barring problems producing the required molten iron. 

It became apparent on the second test day that the next truckload of BF coke was not going to arrive in time to start testing on the third day. Based on the promising performance of the BF coke, a decision was made to accelerate to 100% BF coke.

In the afternoon, we noted that the cupola slag changed color, suggesting the slag was more basic. In addition, spout carbon increased dramatically while silicon began decreasing. This further suggested that perhaps the cupola slag was becoming more basic which could be detrimental to the refractory. 

Accordingly, further testing of the second truckload of BF coke was delayed until a complete chemical composition of the cupola slag and coke could be obtained. In addition, a review of the cupola performance including all charge materials and overall conditions of the cupola needed to be performed. 

The foundry practice where this testing was performed was to vary the blast rate of the cupola to match the molten metal demand at the molding line. Sometimes, low melt demand dictates a lower blast rate such that an under-blowing condition (stack gas ≤ 11% CO) is created. When this occurs, the slag dwell time in the well of the cupola increases, which normally would inhibit carbon gain, but if limestone is not reduced to compensate for lower basicity ash of the BF coke, the slag basicity and volume can increase enough to create an environment where carbon increases and silicon decreases. 

When the second truckload of BF coke was run, the foundry agreed to change the cupola operation practice during times of lower melt demand. The cupola blast rate was held constant and when less molten metal was needed, the blast rate was turned off until the molding line consumed enough metal. Then the blast rate was restored. Since the blast off-time was less than 15 minutes, it resulted in more consistent carbon and silicon concentrations at the cupola spout. 

Data collection/test days were:
    Baseline data foundry coke: 8/16/22
    1st day with BF coke: 8/17/22 (Fig. 2)
    2nd day with BF coke: 11/10/22 (Fig. 3)

Tracking the melting operation during the baseline and testing periods consisted of monitoring each charge and charge components, cupola blast variables, cupola variables and molten iron/slag results (Tables 6-9, Figs. 4-6). 
Upper stack gases were collected continuously and analyzed for CO and CO2 concentration.

Data collected from the cupola output included iron and slag data. Molten iron samples were obtained from the cupola trough at 20-minute intervals and analyzed for chemical composition by thermal analysis and spectrochemical analysis. Molten iron temperature was also obtained at 20-minute intervals. Slag samples were obtained on a less frequent interval and submitted to an external lab for analysis.

Experimental Results

Figures 7-15 and Tables 10-14 show selected operational variables comparing the baseline day using foundry coke with the two days of BF coke testing. 

Conclusions

The low CRI/high CSR coke (BF coke) performed better than low CSR/high CRI coke (Foundry Coke) as evidenced by:
    Improved carbon recovery resulting in less coke needed.
    Improved silicon recovery.
    No difference in spout temperature.

The authors believe the higher strength of BF coke versus foundry coke provided better performance due to larger coke pieces reaching the melt zone.

If it is desired to reproduce this test, the following items will be necessary:
    BF coke must conform to cupola size needs (i.e., 4 in. x 6 in. or 4 in. x 9 in. sizing).
    Proximate and ash analysis of the BF coke should be similar to BF coke used in this testing.
    Amount of limestone may need to be adjusted if less BF coke is required to prevent expected higher slag basicity.
    Coke breeze generated from handling should be tracked and measured.

The authors suggest further testing in a hot blast cupola to confirm whether similar or better results with BF coke can be obtained. Also, the authors extend a challenge to the coke producer industry to produce a suitable foundry coke with lower CRI and higher CSR. Based on the results of this test, it appears that cupola foundries could realize lower coke usage with this type of coke, which would be a potential cost savings, emissions reduction, and an increase in melt rate.