2023 Volume 63 Issue 6 Pages 1011-1016
Blast furnace sludge contains carbon, which can originate from both coal-char and coke. Naturally, it becomes difficult to assign this unutilized carbon to a specific source. Conventional chemical analysis can only predict the total carbon content. This work therefore focuses on the quantification of carbon from the mixture of two different carbon sources using thermo-gravimetric methodology. To establish the methodology, synthetic char has been prepared under different conditions and suitably chosen for this study. Prepared char and coke fines have been heated separately to understand their individual performance. Further, coal-char and coke are mixed in known proportions (wt.%) and subjected to controlled heating under combination of synthetic air and inert atmosphere. Optimized heating profile consists of heating the mixture under inert environment, followed by an isothermal zone of around 12 hrs. Subsequently, the mixture is heated again in inert condition and followed by an isothermal zone of around 4 hrs. The controlled heating and holding time ensure weight loss of known carbon sources occurring separately. Weight loss of the mixture at lower isotherm is solely from carbon from coal-char, and at higher isotherm it is due to coke fines. The ratio of measured weight loss due to carbon sources has been agreed well with the known proportion inside the mixture. This derived process parameters have been found to be equally applicable for the complete range of mixing proportion. Subsequently, developed methodology is applied for different blast furnace sludge samples for quantification of carbon sources.
Blast furnace is a counter-current metallurgical reactor where iron-bearing materials are reduced to liquid iron and tapped out in certain intervals. For reduction purpose, cokes in different sizes are charged from the top. However modern blast furnaces operate with (additional) auxiliary fuel in the form of pulverized coal injection1) to reduce coke rate. But some part of the injected coal and coke remains unutilized, come out from blast furnace and are collected after the cleaning process called as Blast furnace sludge (BF Sludge). Different operational condition leads to different amount of carbonaceous material in top gas and most importantly different percentage of unused carbon from coal and coke in BF sludge. Several experimental studies have been performed to investigate the behaviour of unconsumed pulverized coal in Blast furnace.2,3) Different Computational-fluid dynamics (CFD) based modelling has been reported to examine the amount of unburnt char exited from raceway region.4,5) It has been observed that with increasing coal injection rate, the coal combustion efficiency goes down and reports more unburnt carbon to sludge BF Sludge. Therefore, controlling the coal injection rate as well as other process parameters are important for stable operation. The unused carbon in BF Sludge can provide the insight regarding overall carbon utilization. Moreover, to be specific, it is important to know the ratio of unutilized carbon from coal and coke so that operator can figure out which operation regime deteriorates the coal combustibility, and which results in more degradation of coke.
The size fraction distribution of the carbonaceous part in Blast furnace flue dust (BF flue dust) have been studied by Gupta et al.6) where the size fraction greater than 250 μm has been assigned to coke fines. Moreover, Gupta et al.7) have correlated the change in the stack height of crystalline structure of carbon with thermal treatment on coke fines with the help of X-ray diffraction technique. It helps to identify thermal history and generation of coke fines inside blast furnace either through physical degradation or gasification behavior of coke particle. They have also developed a calibration curve for stack height measurement based on char percentage8) for quantitative analysis, Machado et al.9) have calibrated the equation to find out char percentage in flue dust based on the pure char and coke samples used. They have adopted X-ray diffraction methodology. Their findings show that the carbonaceous part of BF flue dust mainly originates from coke fines and a very small amount is present as char fines.
Several literatures10,11) have adopted chemical and petrographic microanalysis to differentiate the char and coke fines in BF flue dust. Coke and coke fines are divided in different categories and their corresponding amount contributes to the overall ratio of char and coke fines. But the procedure is tedious. Alternatively, Yu et al.12) and Zhong et al.13) have demonstrated that Raman spectroscopy can be an attractive monitoring approach as it requires lesser time. The ratio of different intensity band from Raman spectra has been calibrated using synthetic mixtures of different char percentages and then applied to BF flue dust to derive relative quantity of coal and coke fines. Their result founds to be consistent with the result of petrographic analysis.
Many literatures concentrate on utilization of thermo-gravimetric analyzer for coal and biomass14,15,16) and others focused on analyzing BF flue dust extensively, whereas Ng et al.17) have developed a unique thermo-gravimetric procedure to quantify the percentage of carbon coming from coal and coke, and further modified owing to high char content in Blast furnace sludge compared to BF flue dust. However, for all the cases studied the carbon coming from coal-char always remains on lower side compared to carbon originating from coke. Moreover, present authors have tested the described thermo-gravimetric programme for different BF sludges. But many cases, the methodology fails to produce distinct peak for coal char and coke. This may be due to presence of even higher char amount in BF sludge.
To the best of authors’ knowledge, more work has been concentrated on BF flue dust rather than on BF sludge. Hence, this work focuses on the aspect to differentiate and quantify char and coke percentage in BF sludge. While all the literatures have contributed towards development of different methodologies using synthetic char and coke samples, but none has explicitly revealed the preparation process of synthetic char which can be an important aspect for such development. Present manuscript delineates the parameters of synthetic char preparation in detail. For the quantification of relative carbon percentage in BF sludge, this work uses thermo-gravimetric technique as a tool in comparison with other tedious methods which has not been deployed so extensively for such analysis. In addition, this work attempts to provide a systematic approach based on thermo-gravimetric technique to arrive at the final thermal-gas profile for quantifying carbon of char and coke fines of interests for a wide range of mass ratios. At first, the behavior of synthetic char and coke samples have been studied separately to get an idea of their mass loss behavior under different conditions. Subsequently, with iterations, an optimum thermal-gas profile has been developed. Further it has been verified for synthetic mixture composed of different ratio of coal-char and coke fines. Finally, carbonaceous material from blast furnace sludge has been quantified to check the applicability of the developed methodology.
Carbonaceous part of BF sludge mostly consists of char and coke fines. Char is getting generated due to incomplete combustion of pulverized coal in raceway region of blast furnace. It is unlikely to simulate such harsh condition and short residence time in lab scale apparatus to prepare the char. Authors have first searched different literatures to identify standard procedures available to prepare synthetic char. Several articles18,19,20,21) have reported the usage of Drop Tube Furnace (DTF) at different temperatures. Liming Lu et al.18) have characterized the char prepared from Australian black coal in a Drop Tube Furnace at a temperature of gas 900°C, 1200°C, 1500°C in a slightly oxidizing environment 1(%) v/v O2. The residence time and heating rate of the sample inside the furnace was 1 s and 104°C/s. Casal et al.19) have also prepared char from different types of coal (highly bituminous to anthracite, Vitrinite to inertinite rich) in two different devices: Flat Flame Burner and Drop Tube Furnace. The temperature is maintained at 1300°C for both the devices with different gas environments. The feeding rate, residence time of sample is kept at 1 g/min and 200 ms. It can be inferred form literatures that different approaches have been used for preparation of char and no standardized technique is available. Following part describes the details of preparing the char samples and synthetic mixture.
Coal samples as well as metallurgical coke samples are collected from blast furnace which are often used in daily operation. Coal sample is grounded down to below 200 mesh (< 74 μm) which represents actual size fraction used in pulverized coal injection through tuyere in blast furnaces. In similar line coke samples are also grounded to similar fineness. Characterization of different coal, coke, coal-char samples were carried out by proximate analyzer (TGA 801) and ultimate analyzer (CHNS 628). The physical composition of coal and coke samples are shown in Table 1
| Sample | Proximate Analysis (%) | Ultimate Analysis (%) | ||||||
|---|---|---|---|---|---|---|---|---|
| VM | Ash | FC | C | H | N | S | O | |
| Coal | 15.5 | 10 | 74.5 | 84.05 | 3.05 | 1.43 | 0.64 | 0.83 |
| Coke | 2.07 | 16.55 | 81.38 | 82.2 | 0.04 | 0.4 | 0.68 | 0.13 |
For our purpose, coal samples are being put in Horizontal Tube furnace and different holding temperature has been tried out to observe the impact on coal-char preparation. To prepare coal-char, around 100 gm of coal sample is heated in an inert atmosphere of N2 from room temperature to several levels of holding temperature of 400°C, 600°C, 800°C and 1100°C. The sample is being held up to 3 hrs at the highest temperature and subsequently allowed to cool up to room temperature inside the furnace. Char prepared at different temperatures show significant variation, as can be seen in Table 2.
| Sample details | Maximum holding temperature (°C) | Proximate Analysis (%) | ||
|---|---|---|---|---|
| VM | Ash | FC | ||
| Char_400 | 400 | 13.7 | 11.4 | 74.9 |
| Char_600 | 600 | 6.8 | 12.6 | 80.6 |
| Char_800 | 800 | 2.9 | 13.9 | 83.2 |
| Char_1100 | 1100 | 0.55 | 12.4 | 87.05 |
From Table 2 it can be observed that with increase in maximum holding temperature VM% of the char samples decreases. At lower temperatures, volatile matter does not get released in contrast with the expected complete de-volatization of pulverized coal subjected to raceway environment. Authors have also tried to increase the holding time from 3 hrs to 6 hrs at 600°C to get more de-volatization. However, that process does not yield any improvement in terms of volatile release. At temperature 800°C and 1100°C, VM% has been decreased considerably.
Further analysis has been performed to examine the crystalline properties of different samples through X-ray Diffraction technique (XRD). It has been proven that coke shows higher intensity compared to coal and char samples due to their oriented crystal structure.9) Figure 1 illustrates the XRD pattern of coke and char prepared at 1100°C where similar pattern and same range of crystallinity with higher intensity is observed between them whereas char prepared at 800°C shows lower degree of crystallinity having low intensity. This show coke and char (1100°C) have same coking properties. Therefore, char prepared at 800°C temperature has been taken as input material and mixed in different proportions with coke fines to make synthetic mixtures for further testing.
XRD pattern of Coke and Char prepared at different temperatures. (Online version in color.)
This section demonstrates the development of thermo-gravimetric methodology and subsequent applicability of this technique to different synthetic coal-char and coke mixtures. Subsequently BF sludge has been tested with the developed technique to reveal amount of char generation with respect to coke fines.
3.1. Development of Thermo-gravimetric MethodologyThe principle of this methodology rests on the difference in reactivity for coal-char and coke causing different combustion characteristics under atmospheric environment. And hence at first, individual samples are heated separately in Thermo-gravimetric analyzer (Themys TG) system from room temperature to around 850°C to observe the mass loss behavior under synthetic air atmosphere (79 (wt.) v/v% Ar +21 (wt.) v/v% O2). The main objective is to find out the temperature corresponding to maximum mass loss rate for individual samples.
Figure 2 demonstrates TG and DTG profiles against temperature for both samples. The heating rate of both samples is also varied from 1°C/min to 5°C/min, to check the sensitivity of heating rate on temperature where mass-loss rate is maximum. It can be seen, mass loss for char samples initiates much earlier compared to coke samples substantiating its lower reactivity compared to char-samples. However, as the heating rate is increased, the mass loss rate becomes faster for both samples, but the peak has been shifted towards higher temperature. While coal-char sample shows peak mass loss rate at around 480°C ± 40°C (henceforth referred as TCoal-char), coke fines display it around 650°C ± 40°C (henceforth referred as TCoke) depending on heating rate.
Variations in TG and DTG profiles for individual sample of (a) coal-char (b) coke at different heating rates i.e. 1°C/min, 3°C/min and 5°C/min. (Online version in color.)
It is clearly seen that there is a difference in reactivity of both samples by observing the difference in TCoal-char and TCoke. However, if one look closely at TCoal-char mass loss for coke fines has also started although the rate is low. In addition, it will be required to hold a homogenous mixture, for a sufficient time at a certain temperature level to ensure complete mass loss of char fines before even coke oxidization starts. Therefore, the objective is to fold- find out both suitable temperature and holding time to obtain distinct and separate peak for both carbon sources at two temperature levels.
Towards developing suitable heating profile, synthetic mixture of Coal-char and coke fines is prepared in the ratio of 80 to 20 wt.%. Mixed sample is heated to 450°C from room temperature at heating rate of 4°C/min. To ensure completion of Coal-char oxidation at that temperature, mixed sample is kept isothermally for 4 hrs Further, the mixture is heated to 600°C with slow heating rate at 1°C/min and kept again for 4 hrs for completing coke oxidation. The sample is heated again to 700°C and hold for 2 hrs just to ensure complete oxidation and residue only contains ash.
Figure 3(a) superimposes the temperature profile along with rate of mass loss data against the experimental time. It can be observed that there are three peaks in mass loss rate curve and therefore it is difficult to assign specific peak correspond to coal-char and coke oxidation separately. Moreover, it indicates there has been an overlap of oxidation zone of coal-char and coke. It might be because of insufficient hold time at TCoal-char as well as slow heating rate between temperatures 450°C and 600°C.
Temperature, TG and DTG profile of synthetic mixture with experimental time for (a) Profile-1 (b) Profile-2. (Online version in color.)
With learnings from previous exercise, the heating profile is modified by lowering the first isothermal region to 400°C and corresponding hold time has been increased to 6 hrs for coal-char oxidation. Remaining heating profile remains unchanged. Figure 3(b) describes the heating profile behavior along with mass loss characteristics. But unfortunately, the modified profile does not overcome the shortcomings of previous profile and therefore needs further tweaking.
Figure 4 illustrates the final heating profile developed for the same sample. The modification is being conducted at the lower isothermal region by increasing the temperature to 430°C and corresponding hold time to 12 hrs. The second adjustment is being done in the heating rate between 430°C and 600°C, which has been increased to 4°C/min. Also, the gas profile has been changed from synthetic air mixture to inert atmosphere during the ramp up phase from room temperature to 430°C as well as from 430°C to 600°C.
Temperature, TG and DTG profile of synthetic mixture with experimental time for final profile. (Online version in color.)
The heating profile can predict two distinct peaks of mass loss. It is also realized that the rate of mass loss asymptote towards zero at the end of the lower isothermal zone at 430°C which verifies the end of Coal-char oxidation and demonstrates that coke oxidation does not start. The rest of the mass loss from the sample belongs to coke fines. Comparison of carbon percentage from theoretical calculation and experimental standpoint is being conducted and shown in Table 3.
| Sample No | Theoretical proportion of Coal-char to coke (%) | Theoretical carbon coming from Coal-char (%) | Theoretical carbon coming from coke (%) | Experimental carbon coming from coal-char (%) | Experimental carbon coming from coke (%) |
|---|---|---|---|---|---|
| 1 | 80:20 | 66.40 | 17 | 66.25 | 17.30 |
| 2 | 80:20 | 66.40 | 17 | 68.38 | 16.66 |
| 3 | 80:20 | 66.40 | 17 | 67.19 | 16.05 |
Theoretical carbon is calculated by multiplying the Fixed Carbon (FC) of the char (prepared at 800°C) and coke fines with their corresponding weight percentage in the synthetic mixture as applicable. So, these values respectively indicate the carbon percentage from coal-char and coke fines in the synthetic mixture. Now, one can find out the mass loss (percentage) in the lower isothermal zone through dividing the amount of mass loss at this temperature with the initial mass of the sample (synthetic mixture). This percentage will belong to carbon coming from coal-char fines. The rest mass loss (percentage) is calculated in similar process for coke fines. The accuracy of the prediction deviates only by ±3% from the theoretical proportion whereas repeatability comes within ±2% range as shown in Table 3.
3.2. Implementation of Developed MethodologyPrevious section describes process of developing the thermo-gravimetric methodology for single synthetic mixture of definite proportion. To check the applicability of the methodology, different synthetic mixtures have been tested and compared with the known proportion of mixtures. Table 4 compares the experimentally derived ratio of Coal-char and coke with theoretical proportion in mixed sample.
| Sample No | Theoretical proportion of Coal-char to coke (%) | Theoretical carbon coming from Coal-char (%) | Theoretical carbon coming from coke (%) | Experimental carbon coming from coal-char (%) | Experimental carbon coming from coke (%) |
|---|---|---|---|---|---|
| 1 | 30:70 | 24.9 | 59.5 | 27.3 | 56.55 |
| 2 | 60:40 | 49.8 | 34 | 50.1 | 32.7 |
| 3 | 40:60 | 33.2 | 51 | 32.3 | 52.4 |
Further, samples of blast furnace sludge are collected from different blast furnaces at different days of operation to predict the carbon content from this developed technique and compare with ultimate analysis of the sample. Samples are dried in a bench-scale oven at 105°C (for 6 hrs) to remove the moisture and make it ready for testing.
Table 5 reports carbon (percentage) in BF Sludge through ultimate analysis and compares the data with experimentally observed mass loss value. Further, with the developed methodology, the mass data has been assigned to Coal-char and coke fines percentage in the samples. It is observed that carbon percentage coming from ultimate analysis agrees well with the total mass loss (coal-char + coke) of the sample which can be seen in Fig. 5. Figure 5 shows the TG and DTG plots obtained for sample 2 and 6, which substantiates that the derived methodology works for BF sludge also.
| Sample No. (Blast furnace Sludge) | Total carbon through ultimate analysis (%) | Experimentally determined total mass loss (%) | Experimentally determined ratio of char to coke percentage inside the carbonaceous part (%) |
|---|---|---|---|
| 1 | 41.75 | 42.6 | 39:61 |
| 2 | 37.80 | 37.4 | 56:44 |
| 3 | 6.50 | 6.85 | 44:56 |
| 4 | 20.10 | 18.2 | 41:59 |
| 5 | 45.10 | 43.1 | 75:25 |
| 6 | 33.5 | 33.12 | 63:37 |
Temperature, TG and DTG profile of different blast furnace sludge with experimental time a) sample no. 2 and b) sample no. 6. (Online version in color.)
If the rate of Blast furnace sludge generation is known, this methodology will be able to derive the rate of unburnt char generation owing to insufficient combustion of pulverized coal injection in together with carbon loss from coke. Further, operators can analyze process parameters of blast furnace to optimize the coal injection rate in tandem with coke rate.
The article illustrates a thermo-gravimetric approach to distinguish any proportion of coal-char and coke percentage in a synthetic mixture. It utilizes the difference in reactivity behavior of different carbonaceous material to quantify their individual presence inside a mixture. Hence, it is important to understand the characteristics of individual material to get an initial impression of temperature levels where maximum mass loss rate happens. Char being the high reactive material, shows DTG peak at lower temperature compared to coke fines.
The thermo-gravimetric heating and gas profile has been developed systematically in a step-by-step manner by majorly understanding the mass loss rate at lower isothermal zone. Sufficient hold time at lower isothermal zone proves to be the key contributor in achieving distinguish peaks corresponding to Coal-char and coke fines. Also, switching off oxidizing environment during the ramp up from one isothermal zone to another completely arrests the chance combustion of coke fines and one can easily assigns the mass loss at both isothermal zones to two different carbon samples. The accuracy and reproducibility of the technique is also demonstrated. Here it is worth to note that the temperature-gas profile works fine for the complete range of mixing ratio of coal char and coke fines which is the key aspect before extending it on BF sludge.
This technique has been applied to analyze the carbon content in BF sludge collected at different operational runs. The methodology shows the total carbon amount predicted by the chemical analysis matches well with the carbon percentage predicted by the thermo-gravimetric technique. The analysis also reveals that depending on the operational philosophy char content in BF sludge can be more than coke percentage which hints at the improper combustion of pulverized coal inside Blast furnace.
It can be concluded that the scientific approach outlined in the present manuscript can be adopted quickly to analyze the relative carbon percentage in BF sludge. The importance of such systematic approach is explained in detail along with the capability of developed methodology. Combining the sludge rate with relative carbon percentage, one can calculate carbon loss per specific production rate which is sufficient to indicate utilization of different carbonaceous material inside Blast furnace.
The authors wish to thank Mr Ranjit Singh for carrying out experiments in TGA. The authors are also grateful to the management of Tata Steel Ltd. for granting permission to publish the work.
