- Email: [email protected]
From a mineralogical analytical view to a mechanism evaluation of cement kiln rings
Engineering Failure Analysis 95 (2019) 289–299
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
From a mineralogical analytical view to a mechanism evaluation of cement kiln rings
T
Sahar Belgacema, , Haykel Galaib, Houcine Tissc ⁎
a
Application Laboratory of Chemistry resources and natural substances and the environment, Faculty of Sciences Bizerte, 7021, Jarzouna, Bizerte, Tunisia Laboratory Materials, Processing and Analysis, INRAP, BiotechPole, Sidi Thabet 2020, Ariana, Tunisia c Production laboratory, 7000. Les Ciments de Bizerte, Bizerte, Tunisia b
ARTICLE INFO
ABSTRACT
Keywords: Kiln inlet Rings formation Carbonates Sulfates Quantification Operating data
Ring growth in cement rotary kilns is a complex and a dynamic phenomenon, where the agents responsible for their formation cannot be easily controlled. This is reflected by the mineralogical heterogeneity of the deposit material. In this context, eight samples collected from two rings, formed in the same kiln, but at different periods (during 22 and 3 days), were subjected to qualitative and quantitative mineralogical study. The consolidation of these rings was the result of the combined deposition of the carbonates and the sulphates phases. The analyzes were focused mainly on spurrite Ca5(SiO4)2CO3, calcite CaCO3, sulfospurrite Ca5(SiO4)2(SO4), Ca –Langbeinite K2Ca2(SO4)3 and chlorellestadite Ca10(SiO4)3(SO4)3Cl2. Their identifications and quantifications joined to the variation of sulfur amount and the calcination degree of the hot meal, permitted to better modulate the mechanism of ring formation in the kiln inlet.
1. Introduction The coating formation on the rotary cement kilns walls is a critical phenomenon. The adhering material plays an essential role in the proper functioning of the entire operation. It protects the refractory lining from chemical attacks, thermal shocks and abrasion by the kiln bed, which enhances the reactor lifetime [1,2]. However, the thickness of the crust is not the same throughout the oven. It can reach very elevated heights in some areas and extend for some meters. Thus, this crusting is known as rings [3], which represents a frequent problem in the cement industry by generating an increase in energy consumption, a drop in the furnace production and, in extreme cases; it stops the kiln [1,2,4]. The chemical and mineralogical composition of the rings and their location in the furnace vary according to the process operating conditions. In reality, several agents interfere: kiln dimension, variation of the raw mix chemistry, fuel and heat combustion engineering and selection of refractory [1]. Besides, while the kiln is functioning, volatiles cycles stands; which is considered as an important cause that activates the rings' appearance [5]. The elements concerned are sulfur, potassium, sodium and chlorine. They originate from the raw materials and the fuel, and they circle throughout the system in repetitive evaporation –condensation cycles [6,7]. Cement makers adopt the ASR index (ASR = (Na + K)/(2S/ Cl)) to estimate the composition of volatiles. Problems arise when the value falls outside the [0.8–1.2] range. Based on the chemistry formation, deposit materials are called sulfur rings when there is an excess of sulfur. If the opposite is true, they are called alkali rings. The sulfur excess is the most frequent and severe case [8,9]. Lastly, there is another type of ring known as spurrite ring.
⁎
Corresponding author. E-mail address: [email protected] (S. Belgacem).
https://doi.org/10.1016/j.engfailanal.2018.09.034 Received 25 May 2018; Received in revised form 19 September 2018; Accepted 25 September 2018 Available online 25 September 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Therefore, the in –situ investigation of build-up and rings represents the first key to obtain data about how all parameters, which involved in the clinkering procedure, perform during the process. Few studies have investigated deposits that stick on the cement kiln walls, in order to predict the mechanisms that lead to the deposit formation. Indeed, ternesite Ca5(SiO4)2(SO4) has been recognized as the main mineralogical phase, which is responsible for the material consolidation and the ring formation in the furnace upstream [4]. Other studies revealed the presence of spurrite Ca5(SiO4)2CO3 in ring formed in the pre –burning zone of a wet process cement kiln 10 and the preheater system of short kilns [11,12]. In the present study, the issue under scrutiny is the mineralogical composition variation in cement kiln rings. The qualitative and quantitative analysis of the phases involved allowed us to estimate the mechanisms that govern the ring formation. 2. Experimental 2.1. Kiln characteristics It is a full-dry kiln with a nominal flow rate of 4000 tons / day, length 83 m and internal diameters 5 m, 5.18 m, 5.80 m with four stages of cyclones and a grid cooler. Petroleum coke is the fuel used to generate the flame. 2.2. Sampling The new kiln of the company “Les Ciments de Bizerte” has often developed rings. In this study, the interest has been focused on two rings that appeared in the kiln inlet. The first ring (A) formed after 22 days of service, with a maximum height of 62.5 cm and that extends from 53 m to 63 m of the furnace exit. After the total removal of the ring, a second ring (R) formed after 72 h of operation in the same position as A and with a maximum height of around 40 cm. Five samples were collected from ring (A) noted A1, A2, A3, A4 and A5 and three from ring (R) noted as R1, R2 and R3. The sampling was carried out in the brick –flame direction following the color variation of the build –up as illustrated in Fig.1. In addition, a hot meal sample was taken from the last cyclone of the preheater system and its corresponding raw mix. 2.3. Operating data Table 1 gives the range of hot meal temperatures and the amount of O2 (wt%) at the preheater exit as well as the oxygen quantity at the kiln inlet, during the period associated to the formation of each ring. The Lime Saturation Factor (LSF), Silica Ratio (SR) and Alumina Ratio (AR), characteristics of the raw powder, are also given. Fig. 2 shows the variation of the calcination degree, related to the raw meal entering the kiln, during the period preceding the ring formation. From this data, it is clear that the raw powder, which entered the kiln, was characterized by a non-stable decarbonation rate. 2.4. Analytical techniques The chemical composition was determined via X-ray fluorescence and for the mineralogical investigation, XRD patterns of different samples were recorded by an ARL 9900 diffractometer operating with monochromated Co Kα1 radiation (1.788996 Å). Later, a Rietveld method was applied on all samples, using X'Pert High Score plus, to estimate their mineralogical composition in percentage by weight (QXRD). A thermogravimetric (TGA) analysis was performed using “Setaram Setsys” equipment with a heating rate of 10 °C/min in an inert atmosphere for five samples (A1, A3 A5, R1 and R2). 3. Results 3.1. Chemical composition There is a noticeable concentration of sulfur and alkali (Table 2) as volatile elements compared to those introduced in the raw meal, which is a sign of the high activity of volatile cycles in the kiln inlet. Sulfur and potassium levels fluctuate significantly between [3.68–9.84] for ring A and [4.78–7.47] for ring R. Yet, Na2O seems stable in both rings. 3.2. XRD results A typical XRD profile of heat meal is given in Fig.3. It shows that the major phases, entering into the kiln inlet, were the free lime, the calcite and chlorellestadite. Quartz and C2S also existed in considerable amounts. As minor phases, there were magnesite, spurrite and sulfo-spurrite. In addition, the sample contained a trace of sulfated compounds such as arcanite K2SO4 and K2Ca2(SO4)3. The X-ray diffraction analysis (Figs.4 and 5) revealed the notable heterogeneity of the different samples removed from the two rings. Indeed, they contain the main clinker phases including alite (Ca3SiO5), belite (Ca2SiO4), aluminates and Brownmillerite. Carbonated phases, especially, calcite CaCO3 and spurrite Ca5(SiO4)2CO3 and even magnesite MgCO3. The XRD has permitted the detection of different sulfated phases: chlorellestadite Ca10(SiO4)3(SO4)3Cl2 is present in all samples, the ternesite Ca5(SiO4)2(SO4), and Ca –langbeinite K2Ca2(SO4)3 are absent in sample A1. Portlandite Ca(OH)2 was formed, most likely, from the free lime during the 290
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 1. Photographs of the eight samples and their position in respect of direction brick –Flame. Table 1 Parameters variation affecting the process functioning Parameters T
Preheater outlet
% O2 % O2 Kiln feed
(°C)
(Wt%) (Vol%)
Preheater outlet Kiln inlet
LSF⁎ SR⁎⁎ AM⁎⁎⁎
Period preceding Ring A
Period preceding Ring R
820–860 2.8–3.9 7–8.8 99.43–105.79 2.16–2.5 1.42–1.78
3.1–3.2 5.8–7 99.79–101.74 2.20–2.28 1.63–1.67
LSF = [CaO/ (2.8 SiO2 + 1.18 Al2O3 + 0.65 Fe2O3)] × 100 [13] for % MgO < 2% [14]. SR = SiO2 / (Al2O3 + Fe2O3) [13]. ⁎⁎⁎ AM = Al2O3 / Fe2O3 [13]. ⁎
⁎⁎
kiln cooling [12]. The amounts of spurriteS and sulfo-spurrite illustrated by QXRD are reported in Table 3. Spurrite exists in high amounts that exceed those of sulfospurrite in most samples. The sulfospurrite amount exceeds that of spurrite in two samples, A4 and R3. Only in A2, their percentages are practically equal. The percentages variation of the three major sulfate phases as a function of the sulfur content, measured in the samples, revealed 291
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 2. Variation of calcination degree related to the hot meal entering to the kiln. Each point is the average of three analyses per day in the case of ring A. Table 2 SO3, K2O and Na2O amounts (%Wt). Samples
SO3
K2O Na2O
Raw mix Hot meal Ring A
0.47 2.62 3,68 9,84 8,12 6,51 7,48 7,14 4,78 7,47
< 0.5 0.74 0,78 1,14 1,23 0,68 1,05 0,97 0,82 0,74
Ring R
A1 A2 A3 A4 A5 R1 R2 R3
0.52 0,62 0,81 0,7 0,7 0,64 0,76 0,82 0,78
Fig. 3. XRD pattern of hot meal sample entering in the kiln inlet.
that the ternesite content varies in an antagonistic manner with that of chlorellestadite. This might give a hint about the decomposition –formation of chlorellestadite and ternesite respectively. Also, the amounts of these phases do not vary linearly with the increase in sulfur content. Despite the fact that samples A5 and R3 contain almost the same sulfur percentage, they do not enclose identical amounts of Ca5(SiO4)2(SO4). Furthermore, there is an additional sulfate phase, yeelimite Ca3Al6O12·CaSO4, which was 292
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 4. XRD pattern of ring A samples A1, A2, A3, A4 and A5.
Fig. 5. XRD pattern of ring R samples R1, R2 and R3.
detected only in R3 in a substantial amount. This provides information about the divergence conditions, which led to the development of each ring layer. The temperature had probably reached a sufficient level, during the period attributed to R3 formation, allowing the appearance of yeelimite. 3.3. Thermal analysis Fig.6 illustrates the thermal behavior of a raw mix under an inert atmosphere. The DTA curve presents an important endothermic effect which had been manifested by a mass loss corresponding to calcite decomposition between 567 °C and 835 °C. The two other endothermic effects are attributed to the clinker melting after 1300 °C. At 895 °C, an exothermic effect appears, which is recognized 293
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Table 3 Mineralogical compositions (% Wt) of the collected samples, achieved by QXRD analysis. Phases Ca5 (SiO4)2CO3 CaCO3 MgCO3 CaO Ca(OH)2 Ca5 (SiO4)2 (SO4) Ca10(SiO4)3(SO4)3Cl2⁎ K2Ca2 (SO4) –β K2SO4 K2Ca(SO4)2(H2O) Ca3Al6O12·CaSO4 C3S C2S-β C2S-α C4AF Ca9 (Al6O18) Ca5NaAl3O9 (CaO) 12 (Al2O3)7 Ca2Al2SiO7 Ca3Mg(SiO4)2 MgO SiO2 ⁎
Hot meal Mineralogical name Spurrite Calcite Magnesite Lime Portlandite Ternesite Chlorellestadite Ca –langbeinite Arcanite syngenite Yeelimite Alite Belite Belite Brownmillerite Aluminate cubic Na-Aluminate Mayenite Gehlenite Merwinite Periclase Quartz
2.3 16.1 1.6 28.3 – 0.7 9.9 – 0.7 – – 8.4 7.8 10.6 – – – 0 0.5 2.7 6.6 – 3.7
Ring A
Ring 2
A1
A2
A3
A4
A5
R1
R2
R3
21,8 9,5 6,7 1,6 2,4 – 3,8 0,2 – – – 11,9 23,2 4,5 9 – – 1 – 4,4 – –
16,5 4,5 5,2 0,9 0,4 18,5 8,3 5,5 – – – 5,4 13,8 6,6 7,1 2,4 – 2,5 – 2,5 – –
21,2 17,5 3,6 – 3,6 11,3 5,4 5,4 – 1,1 – 1,6 13,7 3,4 5,8 – 2,2 4,1 – – – –
2,2 0,1 0,4 2,5 12,7 23,7 6,3 0,9 – – 1 0,2 31,6 3,8 7,7 – 0,9 4,4 – 1,2 0,5 –
41 6,3 4,8 – 3,6 1,9 9,4 4,1 – 0,1 0,1 1,7 10,1 4,2 8,7 – – 2,1 – 1,8 – –
30,9 7,1 1 1,8 3,2 8,3 13,8 2,8 – – – 1,6 13,5 2,1 6,1 – – 3,7 – 3,2 0,9 –
10 1,1 3,8 5 4,2 2,4 9 1,7 – – – 1,9 40,4 2,3 6 1,6 1,4 7 – 2,2 – –
1,3 – – 7,6 0,4 16,7 6,8 5,3 – 1,8 4,4 1,6 39,6 2,2 10,2 – – – – 1,5 1,2 –
Chlorine presence was confirmed by Ionic Chromatographic. Example: R2 contain 0.79% of Cl−.
Fig. 6. DTA/TG of the raw mix.
for the silicate formation. The TG curves of the analyzed samples are shown in Fig.7. The weight loss occurs in all samples in three major steps. The first one is located between 372 °C and 443 °C; it corresponds to water elimination, which is linked to Ca(OH)2. The second weight loss corresponds to carbonate disintegration in 450–930 °C temperature range. The CO2 is generated from three carbonate phases. In fact, the decomposition was initiated with MgCO3 from 450 °C to 620 °C, where the calcite decarbonatation started, to then end at 725 °C. Finally, the release of CO2 was completed by spurrite dissociation, which takes place within the 725–930 °C temperature range [10]. 294
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 7. TG curves of the studied rings samples: A1, A3, A5, R1 and R2.
The last weight loss appearing at a temperature beyond 1000 °C, is attributed to SO3 distributed among the Ca –langebinite, ternesite and chlorellestadite. All these phase’ decompositions show the same intermediate, CaSO4, in their decomposition. 3.4. Inter-comparison of the quantitative analysis In order to verify our results, QXRD, TGA and XRF were compared (Figs. 8 and 9) for the carbonate and sulfate phases, confirming
Fig. 8. Calculated amount of CO2 found by thermal analysis and comparison with those estimated by QXRD analysis. 295
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 9. Calculated amount of SO3found by thermal analysis and comparison with those estimated by QXRD analysis and XRF.
the estimation of CO2 and SO2/SO3. Comparing the total CO2 amount determined by ATG to that anticipated by the QXRD analysis, shows the good agreement between the obtained results. If we consider each phase separately, this observation is valid mostly for spurrite. CaCO3 and MgCO3 present divergence, especially, in sample A1. This outcome probably arises, from the overlap that exists between the two phases. Regarding the sulfur, the agreement between the ATG and XRF results was more important than that achieved by the QXRD analysis (Fig.9). This may be due to an XRD limit detection (LD = 1–2%) and a miss crystallization of sulfate phases. Subsequently, they were not detected by XRD, which is seen in the remarkable discrepancy between the Rietveld analysis results and the other methods. 4. Discussion 4.1. Operating data variation In the present study, it is necessary to take into account the parameter variations that influence the clinkering reactions:
• The limestone decomposition is the important reaction of the process, since it determines the kiln operation stability. The control •
•
of this reaction returns to check out the preheater–precalciner kiln system's efficiency by calculating the calcinations degree (C) [7]. The confrontation of the CaCO3 amounts (Table 3), in each ring layer, and the decarbonation variations (Fig.2) clearly show that there is correlation between them. Cement moduli are the main tools to control the composition variation of the raw meal. It is noticed that the fuel consumption varies with the kiln feed burnability, which is linked to the variation in the raw mix chemical composition [14,15]. In the present case, the calculated moduli show that there was a large variation in the chemical composition of the kiln feed; in the 72 h during which ring R formed, this variation was less noticeable than that of ring A. As a consequence, there was an increase in the burning temperature when that was needed to compensate the resulting deviation and to achieve the desired mineralogical composition of the clinker. Preserving the operation at a high temperature increased the sulfur and the potassium volatility in favor of Ca –L and ternesite formation. This could explain the difference in the amounts of these phases in the two rings. As it is known, temperature and oxygen take action during sulfation/desulfation reactions. This arises from the fact that the sulfate phases' stability depends on the partial pressure of SO3 which, in turn, is a function of the SO2 and O2 fugacity and temperature [16]. From Table 1, it can be seen that there is a large variation of oxygen, especially in the kiln inlet. This irregularity determines the chemical reaction in the kiln inlet.
The local variability of gases' flux, loaded with alkalis SO2 and/or CO2, is certainly responsible for the heterogeneity of the rings' composition. Hence, it is interesting to focus on the analysis of carbonates and sulfates. 4.2. Carbonates and sulfates formation Theoretically, by advancing along the furnace (> 1000 °C), CaCO3 must be fully decomposed in CaO and CO2 and where the first clinkering reactions have already started. However, this phase was detected by XRD in A1, A3, A5 and R2 in significant levels (Table 3). The explanation of its presence at this point of the kiln can be explained through several hypotheses:
• Thermal and mass transfer has been limited by the large distribution of the meal particles [17]. • As the calcination kinetics is sensitive to the CO partial pressure, an increase in the carbon dioxide concentration in the 2
296
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
atmosphere around CaCO3 particles can inhibit its decomposition [17]. Thus, the calcination process will require a higher temperature to be completed and uncalcined grains will be transported to the hottest areas in the furnace (> 1000 °C). From the literature [8,18], there are several possible pathways to form spurrite: either from calcite CaCO3 with quartz SiO2 or C2S according to reactions (1) and (2), or following the alite and/or belite decomposition when the atmosphere becomes rich in CO2 and the spurrite reaches its stability domain, according to reactions (3) and (4).)
2SiO2 + 5 CaCO3 2 Ca2SiO4 + CaCO3
(2)
Ca5 (SiO4 )2 CO3 Ca5 (SiO4 ) 2 CO3
(3)
Ca5 (SiO4 )2 CO3 + CaCO3
(4)
2 Ca2SiO4 + CaO + CO2 (g) Ca3SiO5 + CO2 (g)
(1)
Ca5 (SiO4 ) 2 CO3 + CO2
Reaction (1) is the most privileged in the last cyclone since the hot meal contains sufficient amounts of calcite and quartz. Yet, in the kiln inlet, the spurrite, most likely, originated from all or one of the reactions (2), (3) and (4). In a mixed gas (SO2/CO2), reactions with sulfur are favored instead of those with CO2 [8]. Consequently, the sulfated phases are primarily formed. If K2O exists in the solid, the Ca-Langbeinite appears at 940 °C [19] following reaction (5). This phase was identified as the precursor of the sulfospurrite formation [4] according the Eq. (6). With the presence of free lime, Ca5 (SiO4)2(SO4) can appear following Eq. (7) [20].
2CaO + 3SO3 + K2O
(5)
K2 Ca2 (SO4 )3
4 Ca2SiO4 + K2 SO4 . 2CaSO4
(6)
2(2CaSiO4 . CaSO4) + K2SO4
2Ca2 SiO4 + CaO + SO2 (g) + ½ O2 (g)
(7)
Ca5 (SiO4 )2 (SO4 )
Also, ternesite might be taking place from chlorellestadite decomposition, at a temperature of about 1000 °C [11], according to reaction (8). Ca10 (SiO4)3(SO4)3Cl2 → CaCl2 (g) + Ca5 (SiO4)2(SO4) + Ca2SiO4 + 2(CaSO4) (8). At about 900 °C, the free lime catches about 100% of SO2 (g) released from fuel combustion and thermal heating of the raw mix forming CaSO4 [6]. Other than that, there is no anhydrite detected in the studied samples; the CaSO4 phase probably formed in the material as an intermediate phase which explains its absence. 4.3. Proposed model for ring formation The aim of this section is to present a model illustrating the ring formation in the kiln inlet (Fig. 10) supported by a representation of the principal chemical reaction responsible for the ring appearance (Fig. 11). Based on what has been discussed, the chlorellestadite appearance was favored in the preheater system instead of spurrite and ternesite. In fact, there is an initial slow appearance of spurrite in the hot meal inflowing to the kiln upstream according to reaction (1).The kiln inlet is charged by volatile elements, mainly sulfur and potassium, transported from the burning zone by the hot gases. The temperature of the last one (Tg) is higher than that of the hot meal (Tm) coming from the preheater system. As a consequence, the volatile species condensed (Step 1) at a lower temperature point 12 on the cold powder, thus, generating Ca –langbeinite by a reaction with free lime. This created a sticky surface (Step 2) around the powder grains and/or on the kiln walls, which facilitated their adhering on the cold kiln walls. As third step, there are two possible scenarios: Scenario I: After the material took arrangement on the kiln wall as the ring formed, the mineralogical transformation (Step I-4) began through the repetitive exposure to the hot gases (SO2/CO2) as well as the dust particles. The richness of such environment accelerated the conversion of the starting phases, which are CaO, C2S, CaCO3 and the undecomposed chlorellestadite. Thus, the spurrite formation occurs via CO2 desorption by C2S and/or C3S following reactions (3) and (4) respectively. Ternesite appears after reaction (5) takes place and/or when the partial pressure of SO2 becomes enough to boost the belite sulfurization according the Eq. (6). It should be noted that the spurrite and sulfospurrite amounts depend on the CO2 and SO2 levels in the local atmosphere as well as the temperature. Scenario II: The melt formation enhances the reactions leading to spurrite and sulfospurrite formation (Step II-3), according to the equations mentioned above. These reactions are competitive as they depend on temperature and the partial pressure of CO2 (g) and SO2 (g). The sticky surface catches the particles in suspension (Step II-4) and the ring appears. As a final step, the ring consolidation (Step 5) owes to the combination effect of the phase's nature and to the addition and heat removal cycles resulting from the kiln rotation. The ring thickness increases as the kiln rotates and it acquires its compactness through the heating cycles. Nevertheless, the growth rate is related to temperature, presence of a molten phase, SO2 and CO2 partial pressure. 5. Conclusion Extracting more information about the conditions, which govern the ring formation process, represents an efficient tool to better understand the rings appearance and their growth. In this perspective, an analytical investigation was performed on two kiln rings material, which formed almost in the same position (kiln inlet), as well as the operating data that characterized the period associated to each ring. 297
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 10. Rings formation model in the kiln inlet.
The results yielded by this study convincing evidence that the variation of operating parameters explains the diversity of the mineralogical phases and their proportions, in the rings material. Indeed, the calcination degree is one of the agents that govern the presence of spurrite and calcite. It is imperative to preserve a stable decarbonation amount of the hot meal. Yet, what is most important is to optimize the tolerable range of its variation. Also, it is imperative to continuously check and optimize the oxygen content in the kiln in order to have command over the sulfur cycle. Consequently, this will pilot the formation of ternesite, Ca –langbeinite and chlorellestadite. Also, the amounts determination of aggressive elements (S, K, Cl− and Na) is not enough to deal with the ring appearance. Thus, qualitative and quantitative analysis of the phases was achieved for a better estimation of the reactions governing the appearance of carbonates and sulfates at the kiln inlet.
Fig. 11. Illustration of chemical reactions involved in rings formation in the kiln inlet. 298
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Acknowledgments “Les Ciments de Bizerte” is acknowledged for their support. References [1] Maria del Mar Cortada Mut, Linda Kaare Nørskov, Flemming Jappe Frandsen, Peter Glarborg, Kim Dam-Johansen, Review: Circulation of Inorganic elements in Combustion of Alternative Fuels in Cement Plants, Energy Fuel 29 (2015) 4076–4099. [2] Jagendran Ravindran, Soundarajan Krishnan, Studies on Thermal Analysis of Cement Rotary Kiln Based on Clinker Coating Materials on Refractories, Energy and Monetary Savings. Congress on Recent Development in Engineering and Technology (RDET-16). August 22–24, (2016) Kuala Lumpur (Malaysia). [3] K.E. Peray, J.J. Waddell, The Rotary Cement Kiln, Chemical Publishing Co, New York, 1972 ISBN: 978-08-20603-67-4. [4] J. Nievoll, S. Jorg, K. Dosinger, Corpus, Sulphur, spurrite, and rings-Always a headache for the cement kiln operator ? J. World. Chem. (2009) 78–83. [5] D. Gastaldi, E. Boccaleri, F. Canonico, In situ Raman study of mineral phases formed as by-products in a rotary kiln for clinker production, J. Raman Spectrosc. 39 (2008) 806–812. [6] Maria del Mar Cortada Mut, Linda Kaare Nørskov, Peter Glarborg, Kim Dam-Johansen, SO2 release as consequence of alternative fuels combustion in cement rotary kiln inlets, Energy Fuel 29 (4) (2015) 2729–2737. [7] Anjan K. Chatterjee, Chemistry and engineering of the clinkerization process-Incremental advances and lack of break throughs, Cem. Concr. Res. 41 (2011) 624–641. [8] A.K. Chatterjee, Role of Volatiles in Cement Manufacture and in the Use of Cement, Advances in Cement Technology – Critical Reviews and Case Studies on Manufacturing, Quality Control, Optimization and Use, (2011), pp. 203–236. [9] Haykel Galai, Sarra Ben Hafdhallah, Mineral quantification of preheated cement raw mix, Adv. Cem. Res. 25 (5) (2013) 295–300. [10] G. Goswami, B. P. Padhy and J. D. Panda, Thermal analysis of spurrite from a rotary cement kiln, J. Therm. Anal. Vol. 35 (1129–1136). [11] I. Recio Dominguez, J. Gómez-Millán, M. Alvarez, S. De Aza, L. Contreras, A.H. De Aza, Build-up formation and corrosion of monolithic refractories in cement kiln preheaters, J. Eur. Ceram. Soc. 30 (2010) 1879–1885. [12] W. Kurdowski, M. Sobon, Mineral composition of build -up in cement kiln preheater, J. Therm. Anal. Calorim. Vol. 55 (1999) 1021–1029. [13] Samira Telschow, Flemming Frandsen, Kirsten Theisen, Kim Dam-Johansen, Cement Formation, A Success Story in a Black Box: High Temperature phase Formation of Portland Cement Clinker, Ind. Eng. Chem. Res. 51 (2012) 10983–11004. [14] Mohamed A. Aldieb, G. Hesham, Proceedings of the World Congress on Engineering and Computer Science. 2010 Vol II WCECS 2010, October 20–22, (2010) (San Francisco, USA). [15] T. K. Chatterjee, Burnability and Clinkerization of Cement Raw Mixes, Advances in Cement Technology – Critical Reviews and Case Studies on Manufacturing, Quality Control, Optimization and Use, pp. 69–113. [16] Isabel Galana, Theodore Haneinb, Ammar Elhowerisa, Marcus Bannermanb, Fredrik P. Glassera, Phase compatibility in the system CaO-SiO2-Al2O3-SO3- Fe2O3 and the effect of partial pressure on phase stability, Ind. Eng. Chem. Res. 56 (2017) 2341–2349. [17] B.R. Stanmore, P. Gilot, Review—Calcination and carbonation of limestone during thermal cycling for CO2 sequestration, Fuel Process. Technol. 86 (2005) 1707–1743. [18] H. Bolio-Arceo, F.P. Glasser, Formation of spurrite, Ca5(SiO4)2CO3, Cem. Concr. Res. 20 (1990) 301–307. [19] F.W.Taylor Harold (Ed.), The chemistry of Portland cement Manufacture, Cement Chemistry, second Ed. Thomas Telford, London, U.K., 1997ISBN: 9780727725929. [20] Theodore Hanein, Isabel Galan, Fredrik P. Glasser, Solon Skalamprinos, Ammar Elhoweris, Mohammed S. Imbabi, Marcus N. Bannerman, Stability of ternesite and the production at scale of ternesite-based clinkers, Cem. Concr. Res. 98 (2017) 91–100.
299
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
From a mineralogical analytical view to a mechanism evaluation of cement kiln rings
T
Sahar Belgacema, , Haykel Galaib, Houcine Tissc ⁎
a
Application Laboratory of Chemistry resources and natural substances and the environment, Faculty of Sciences Bizerte, 7021, Jarzouna, Bizerte, Tunisia Laboratory Materials, Processing and Analysis, INRAP, BiotechPole, Sidi Thabet 2020, Ariana, Tunisia c Production laboratory, 7000. Les Ciments de Bizerte, Bizerte, Tunisia b
ARTICLE INFO
ABSTRACT
Keywords: Kiln inlet Rings formation Carbonates Sulfates Quantification Operating data
Ring growth in cement rotary kilns is a complex and a dynamic phenomenon, where the agents responsible for their formation cannot be easily controlled. This is reflected by the mineralogical heterogeneity of the deposit material. In this context, eight samples collected from two rings, formed in the same kiln, but at different periods (during 22 and 3 days), were subjected to qualitative and quantitative mineralogical study. The consolidation of these rings was the result of the combined deposition of the carbonates and the sulphates phases. The analyzes were focused mainly on spurrite Ca5(SiO4)2CO3, calcite CaCO3, sulfospurrite Ca5(SiO4)2(SO4), Ca –Langbeinite K2Ca2(SO4)3 and chlorellestadite Ca10(SiO4)3(SO4)3Cl2. Their identifications and quantifications joined to the variation of sulfur amount and the calcination degree of the hot meal, permitted to better modulate the mechanism of ring formation in the kiln inlet.
1. Introduction The coating formation on the rotary cement kilns walls is a critical phenomenon. The adhering material plays an essential role in the proper functioning of the entire operation. It protects the refractory lining from chemical attacks, thermal shocks and abrasion by the kiln bed, which enhances the reactor lifetime [1,2]. However, the thickness of the crust is not the same throughout the oven. It can reach very elevated heights in some areas and extend for some meters. Thus, this crusting is known as rings [3], which represents a frequent problem in the cement industry by generating an increase in energy consumption, a drop in the furnace production and, in extreme cases; it stops the kiln [1,2,4]. The chemical and mineralogical composition of the rings and their location in the furnace vary according to the process operating conditions. In reality, several agents interfere: kiln dimension, variation of the raw mix chemistry, fuel and heat combustion engineering and selection of refractory [1]. Besides, while the kiln is functioning, volatiles cycles stands; which is considered as an important cause that activates the rings' appearance [5]. The elements concerned are sulfur, potassium, sodium and chlorine. They originate from the raw materials and the fuel, and they circle throughout the system in repetitive evaporation –condensation cycles [6,7]. Cement makers adopt the ASR index (ASR = (Na + K)/(2S/ Cl)) to estimate the composition of volatiles. Problems arise when the value falls outside the [0.8–1.2] range. Based on the chemistry formation, deposit materials are called sulfur rings when there is an excess of sulfur. If the opposite is true, they are called alkali rings. The sulfur excess is the most frequent and severe case [8,9]. Lastly, there is another type of ring known as spurrite ring.
⁎
Corresponding author. E-mail address: [email protected] (S. Belgacem).
https://doi.org/10.1016/j.engfailanal.2018.09.034 Received 25 May 2018; Received in revised form 19 September 2018; Accepted 25 September 2018 Available online 25 September 2018 1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Therefore, the in –situ investigation of build-up and rings represents the first key to obtain data about how all parameters, which involved in the clinkering procedure, perform during the process. Few studies have investigated deposits that stick on the cement kiln walls, in order to predict the mechanisms that lead to the deposit formation. Indeed, ternesite Ca5(SiO4)2(SO4) has been recognized as the main mineralogical phase, which is responsible for the material consolidation and the ring formation in the furnace upstream [4]. Other studies revealed the presence of spurrite Ca5(SiO4)2CO3 in ring formed in the pre –burning zone of a wet process cement kiln 10 and the preheater system of short kilns [11,12]. In the present study, the issue under scrutiny is the mineralogical composition variation in cement kiln rings. The qualitative and quantitative analysis of the phases involved allowed us to estimate the mechanisms that govern the ring formation. 2. Experimental 2.1. Kiln characteristics It is a full-dry kiln with a nominal flow rate of 4000 tons / day, length 83 m and internal diameters 5 m, 5.18 m, 5.80 m with four stages of cyclones and a grid cooler. Petroleum coke is the fuel used to generate the flame. 2.2. Sampling The new kiln of the company “Les Ciments de Bizerte” has often developed rings. In this study, the interest has been focused on two rings that appeared in the kiln inlet. The first ring (A) formed after 22 days of service, with a maximum height of 62.5 cm and that extends from 53 m to 63 m of the furnace exit. After the total removal of the ring, a second ring (R) formed after 72 h of operation in the same position as A and with a maximum height of around 40 cm. Five samples were collected from ring (A) noted A1, A2, A3, A4 and A5 and three from ring (R) noted as R1, R2 and R3. The sampling was carried out in the brick –flame direction following the color variation of the build –up as illustrated in Fig.1. In addition, a hot meal sample was taken from the last cyclone of the preheater system and its corresponding raw mix. 2.3. Operating data Table 1 gives the range of hot meal temperatures and the amount of O2 (wt%) at the preheater exit as well as the oxygen quantity at the kiln inlet, during the period associated to the formation of each ring. The Lime Saturation Factor (LSF), Silica Ratio (SR) and Alumina Ratio (AR), characteristics of the raw powder, are also given. Fig. 2 shows the variation of the calcination degree, related to the raw meal entering the kiln, during the period preceding the ring formation. From this data, it is clear that the raw powder, which entered the kiln, was characterized by a non-stable decarbonation rate. 2.4. Analytical techniques The chemical composition was determined via X-ray fluorescence and for the mineralogical investigation, XRD patterns of different samples were recorded by an ARL 9900 diffractometer operating with monochromated Co Kα1 radiation (1.788996 Å). Later, a Rietveld method was applied on all samples, using X'Pert High Score plus, to estimate their mineralogical composition in percentage by weight (QXRD). A thermogravimetric (TGA) analysis was performed using “Setaram Setsys” equipment with a heating rate of 10 °C/min in an inert atmosphere for five samples (A1, A3 A5, R1 and R2). 3. Results 3.1. Chemical composition There is a noticeable concentration of sulfur and alkali (Table 2) as volatile elements compared to those introduced in the raw meal, which is a sign of the high activity of volatile cycles in the kiln inlet. Sulfur and potassium levels fluctuate significantly between [3.68–9.84] for ring A and [4.78–7.47] for ring R. Yet, Na2O seems stable in both rings. 3.2. XRD results A typical XRD profile of heat meal is given in Fig.3. It shows that the major phases, entering into the kiln inlet, were the free lime, the calcite and chlorellestadite. Quartz and C2S also existed in considerable amounts. As minor phases, there were magnesite, spurrite and sulfo-spurrite. In addition, the sample contained a trace of sulfated compounds such as arcanite K2SO4 and K2Ca2(SO4)3. The X-ray diffraction analysis (Figs.4 and 5) revealed the notable heterogeneity of the different samples removed from the two rings. Indeed, they contain the main clinker phases including alite (Ca3SiO5), belite (Ca2SiO4), aluminates and Brownmillerite. Carbonated phases, especially, calcite CaCO3 and spurrite Ca5(SiO4)2CO3 and even magnesite MgCO3. The XRD has permitted the detection of different sulfated phases: chlorellestadite Ca10(SiO4)3(SO4)3Cl2 is present in all samples, the ternesite Ca5(SiO4)2(SO4), and Ca –langbeinite K2Ca2(SO4)3 are absent in sample A1. Portlandite Ca(OH)2 was formed, most likely, from the free lime during the 290
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 1. Photographs of the eight samples and their position in respect of direction brick –Flame. Table 1 Parameters variation affecting the process functioning Parameters T
Preheater outlet
% O2 % O2 Kiln feed
(°C)
(Wt%) (Vol%)
Preheater outlet Kiln inlet
LSF⁎ SR⁎⁎ AM⁎⁎⁎
Period preceding Ring A
Period preceding Ring R
820–860 2.8–3.9 7–8.8 99.43–105.79 2.16–2.5 1.42–1.78
3.1–3.2 5.8–7 99.79–101.74 2.20–2.28 1.63–1.67
LSF = [CaO/ (2.8 SiO2 + 1.18 Al2O3 + 0.65 Fe2O3)] × 100 [13] for % MgO < 2% [14]. SR = SiO2 / (Al2O3 + Fe2O3) [13]. ⁎⁎⁎ AM = Al2O3 / Fe2O3 [13]. ⁎
⁎⁎
kiln cooling [12]. The amounts of spurriteS and sulfo-spurrite illustrated by QXRD are reported in Table 3. Spurrite exists in high amounts that exceed those of sulfospurrite in most samples. The sulfospurrite amount exceeds that of spurrite in two samples, A4 and R3. Only in A2, their percentages are practically equal. The percentages variation of the three major sulfate phases as a function of the sulfur content, measured in the samples, revealed 291
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 2. Variation of calcination degree related to the hot meal entering to the kiln. Each point is the average of three analyses per day in the case of ring A. Table 2 SO3, K2O and Na2O amounts (%Wt). Samples
SO3
K2O Na2O
Raw mix Hot meal Ring A
0.47 2.62 3,68 9,84 8,12 6,51 7,48 7,14 4,78 7,47
< 0.5 0.74 0,78 1,14 1,23 0,68 1,05 0,97 0,82 0,74
Ring R
A1 A2 A3 A4 A5 R1 R2 R3
0.52 0,62 0,81 0,7 0,7 0,64 0,76 0,82 0,78
Fig. 3. XRD pattern of hot meal sample entering in the kiln inlet.
that the ternesite content varies in an antagonistic manner with that of chlorellestadite. This might give a hint about the decomposition –formation of chlorellestadite and ternesite respectively. Also, the amounts of these phases do not vary linearly with the increase in sulfur content. Despite the fact that samples A5 and R3 contain almost the same sulfur percentage, they do not enclose identical amounts of Ca5(SiO4)2(SO4). Furthermore, there is an additional sulfate phase, yeelimite Ca3Al6O12·CaSO4, which was 292
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 4. XRD pattern of ring A samples A1, A2, A3, A4 and A5.
Fig. 5. XRD pattern of ring R samples R1, R2 and R3.
detected only in R3 in a substantial amount. This provides information about the divergence conditions, which led to the development of each ring layer. The temperature had probably reached a sufficient level, during the period attributed to R3 formation, allowing the appearance of yeelimite. 3.3. Thermal analysis Fig.6 illustrates the thermal behavior of a raw mix under an inert atmosphere. The DTA curve presents an important endothermic effect which had been manifested by a mass loss corresponding to calcite decomposition between 567 °C and 835 °C. The two other endothermic effects are attributed to the clinker melting after 1300 °C. At 895 °C, an exothermic effect appears, which is recognized 293
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Table 3 Mineralogical compositions (% Wt) of the collected samples, achieved by QXRD analysis. Phases Ca5 (SiO4)2CO3 CaCO3 MgCO3 CaO Ca(OH)2 Ca5 (SiO4)2 (SO4) Ca10(SiO4)3(SO4)3Cl2⁎ K2Ca2 (SO4) –β K2SO4 K2Ca(SO4)2(H2O) Ca3Al6O12·CaSO4 C3S C2S-β C2S-α C4AF Ca9 (Al6O18) Ca5NaAl3O9 (CaO) 12 (Al2O3)7 Ca2Al2SiO7 Ca3Mg(SiO4)2 MgO SiO2 ⁎
Hot meal Mineralogical name Spurrite Calcite Magnesite Lime Portlandite Ternesite Chlorellestadite Ca –langbeinite Arcanite syngenite Yeelimite Alite Belite Belite Brownmillerite Aluminate cubic Na-Aluminate Mayenite Gehlenite Merwinite Periclase Quartz
2.3 16.1 1.6 28.3 – 0.7 9.9 – 0.7 – – 8.4 7.8 10.6 – – – 0 0.5 2.7 6.6 – 3.7
Ring A
Ring 2
A1
A2
A3
A4
A5
R1
R2
R3
21,8 9,5 6,7 1,6 2,4 – 3,8 0,2 – – – 11,9 23,2 4,5 9 – – 1 – 4,4 – –
16,5 4,5 5,2 0,9 0,4 18,5 8,3 5,5 – – – 5,4 13,8 6,6 7,1 2,4 – 2,5 – 2,5 – –
21,2 17,5 3,6 – 3,6 11,3 5,4 5,4 – 1,1 – 1,6 13,7 3,4 5,8 – 2,2 4,1 – – – –
2,2 0,1 0,4 2,5 12,7 23,7 6,3 0,9 – – 1 0,2 31,6 3,8 7,7 – 0,9 4,4 – 1,2 0,5 –
41 6,3 4,8 – 3,6 1,9 9,4 4,1 – 0,1 0,1 1,7 10,1 4,2 8,7 – – 2,1 – 1,8 – –
30,9 7,1 1 1,8 3,2 8,3 13,8 2,8 – – – 1,6 13,5 2,1 6,1 – – 3,7 – 3,2 0,9 –
10 1,1 3,8 5 4,2 2,4 9 1,7 – – – 1,9 40,4 2,3 6 1,6 1,4 7 – 2,2 – –
1,3 – – 7,6 0,4 16,7 6,8 5,3 – 1,8 4,4 1,6 39,6 2,2 10,2 – – – – 1,5 1,2 –
Chlorine presence was confirmed by Ionic Chromatographic. Example: R2 contain 0.79% of Cl−.
Fig. 6. DTA/TG of the raw mix.
for the silicate formation. The TG curves of the analyzed samples are shown in Fig.7. The weight loss occurs in all samples in three major steps. The first one is located between 372 °C and 443 °C; it corresponds to water elimination, which is linked to Ca(OH)2. The second weight loss corresponds to carbonate disintegration in 450–930 °C temperature range. The CO2 is generated from three carbonate phases. In fact, the decomposition was initiated with MgCO3 from 450 °C to 620 °C, where the calcite decarbonatation started, to then end at 725 °C. Finally, the release of CO2 was completed by spurrite dissociation, which takes place within the 725–930 °C temperature range [10]. 294
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 7. TG curves of the studied rings samples: A1, A3, A5, R1 and R2.
The last weight loss appearing at a temperature beyond 1000 °C, is attributed to SO3 distributed among the Ca –langebinite, ternesite and chlorellestadite. All these phase’ decompositions show the same intermediate, CaSO4, in their decomposition. 3.4. Inter-comparison of the quantitative analysis In order to verify our results, QXRD, TGA and XRF were compared (Figs. 8 and 9) for the carbonate and sulfate phases, confirming
Fig. 8. Calculated amount of CO2 found by thermal analysis and comparison with those estimated by QXRD analysis. 295
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 9. Calculated amount of SO3found by thermal analysis and comparison with those estimated by QXRD analysis and XRF.
the estimation of CO2 and SO2/SO3. Comparing the total CO2 amount determined by ATG to that anticipated by the QXRD analysis, shows the good agreement between the obtained results. If we consider each phase separately, this observation is valid mostly for spurrite. CaCO3 and MgCO3 present divergence, especially, in sample A1. This outcome probably arises, from the overlap that exists between the two phases. Regarding the sulfur, the agreement between the ATG and XRF results was more important than that achieved by the QXRD analysis (Fig.9). This may be due to an XRD limit detection (LD = 1–2%) and a miss crystallization of sulfate phases. Subsequently, they were not detected by XRD, which is seen in the remarkable discrepancy between the Rietveld analysis results and the other methods. 4. Discussion 4.1. Operating data variation In the present study, it is necessary to take into account the parameter variations that influence the clinkering reactions:
• The limestone decomposition is the important reaction of the process, since it determines the kiln operation stability. The control •
•
of this reaction returns to check out the preheater–precalciner kiln system's efficiency by calculating the calcinations degree (C) [7]. The confrontation of the CaCO3 amounts (Table 3), in each ring layer, and the decarbonation variations (Fig.2) clearly show that there is correlation between them. Cement moduli are the main tools to control the composition variation of the raw meal. It is noticed that the fuel consumption varies with the kiln feed burnability, which is linked to the variation in the raw mix chemical composition [14,15]. In the present case, the calculated moduli show that there was a large variation in the chemical composition of the kiln feed; in the 72 h during which ring R formed, this variation was less noticeable than that of ring A. As a consequence, there was an increase in the burning temperature when that was needed to compensate the resulting deviation and to achieve the desired mineralogical composition of the clinker. Preserving the operation at a high temperature increased the sulfur and the potassium volatility in favor of Ca –L and ternesite formation. This could explain the difference in the amounts of these phases in the two rings. As it is known, temperature and oxygen take action during sulfation/desulfation reactions. This arises from the fact that the sulfate phases' stability depends on the partial pressure of SO3 which, in turn, is a function of the SO2 and O2 fugacity and temperature [16]. From Table 1, it can be seen that there is a large variation of oxygen, especially in the kiln inlet. This irregularity determines the chemical reaction in the kiln inlet.
The local variability of gases' flux, loaded with alkalis SO2 and/or CO2, is certainly responsible for the heterogeneity of the rings' composition. Hence, it is interesting to focus on the analysis of carbonates and sulfates. 4.2. Carbonates and sulfates formation Theoretically, by advancing along the furnace (> 1000 °C), CaCO3 must be fully decomposed in CaO and CO2 and where the first clinkering reactions have already started. However, this phase was detected by XRD in A1, A3, A5 and R2 in significant levels (Table 3). The explanation of its presence at this point of the kiln can be explained through several hypotheses:
• Thermal and mass transfer has been limited by the large distribution of the meal particles [17]. • As the calcination kinetics is sensitive to the CO partial pressure, an increase in the carbon dioxide concentration in the 2
296
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
atmosphere around CaCO3 particles can inhibit its decomposition [17]. Thus, the calcination process will require a higher temperature to be completed and uncalcined grains will be transported to the hottest areas in the furnace (> 1000 °C). From the literature [8,18], there are several possible pathways to form spurrite: either from calcite CaCO3 with quartz SiO2 or C2S according to reactions (1) and (2), or following the alite and/or belite decomposition when the atmosphere becomes rich in CO2 and the spurrite reaches its stability domain, according to reactions (3) and (4).)
2SiO2 + 5 CaCO3 2 Ca2SiO4 + CaCO3
(2)
Ca5 (SiO4 )2 CO3 Ca5 (SiO4 ) 2 CO3
(3)
Ca5 (SiO4 )2 CO3 + CaCO3
(4)
2 Ca2SiO4 + CaO + CO2 (g) Ca3SiO5 + CO2 (g)
(1)
Ca5 (SiO4 ) 2 CO3 + CO2
Reaction (1) is the most privileged in the last cyclone since the hot meal contains sufficient amounts of calcite and quartz. Yet, in the kiln inlet, the spurrite, most likely, originated from all or one of the reactions (2), (3) and (4). In a mixed gas (SO2/CO2), reactions with sulfur are favored instead of those with CO2 [8]. Consequently, the sulfated phases are primarily formed. If K2O exists in the solid, the Ca-Langbeinite appears at 940 °C [19] following reaction (5). This phase was identified as the precursor of the sulfospurrite formation [4] according the Eq. (6). With the presence of free lime, Ca5 (SiO4)2(SO4) can appear following Eq. (7) [20].
2CaO + 3SO3 + K2O
(5)
K2 Ca2 (SO4 )3
4 Ca2SiO4 + K2 SO4 . 2CaSO4
(6)
2(2CaSiO4 . CaSO4) + K2SO4
2Ca2 SiO4 + CaO + SO2 (g) + ½ O2 (g)
(7)
Ca5 (SiO4 )2 (SO4 )
Also, ternesite might be taking place from chlorellestadite decomposition, at a temperature of about 1000 °C [11], according to reaction (8). Ca10 (SiO4)3(SO4)3Cl2 → CaCl2 (g) + Ca5 (SiO4)2(SO4) + Ca2SiO4 + 2(CaSO4) (8). At about 900 °C, the free lime catches about 100% of SO2 (g) released from fuel combustion and thermal heating of the raw mix forming CaSO4 [6]. Other than that, there is no anhydrite detected in the studied samples; the CaSO4 phase probably formed in the material as an intermediate phase which explains its absence. 4.3. Proposed model for ring formation The aim of this section is to present a model illustrating the ring formation in the kiln inlet (Fig. 10) supported by a representation of the principal chemical reaction responsible for the ring appearance (Fig. 11). Based on what has been discussed, the chlorellestadite appearance was favored in the preheater system instead of spurrite and ternesite. In fact, there is an initial slow appearance of spurrite in the hot meal inflowing to the kiln upstream according to reaction (1).The kiln inlet is charged by volatile elements, mainly sulfur and potassium, transported from the burning zone by the hot gases. The temperature of the last one (Tg) is higher than that of the hot meal (Tm) coming from the preheater system. As a consequence, the volatile species condensed (Step 1) at a lower temperature point 12 on the cold powder, thus, generating Ca –langbeinite by a reaction with free lime. This created a sticky surface (Step 2) around the powder grains and/or on the kiln walls, which facilitated their adhering on the cold kiln walls. As third step, there are two possible scenarios: Scenario I: After the material took arrangement on the kiln wall as the ring formed, the mineralogical transformation (Step I-4) began through the repetitive exposure to the hot gases (SO2/CO2) as well as the dust particles. The richness of such environment accelerated the conversion of the starting phases, which are CaO, C2S, CaCO3 and the undecomposed chlorellestadite. Thus, the spurrite formation occurs via CO2 desorption by C2S and/or C3S following reactions (3) and (4) respectively. Ternesite appears after reaction (5) takes place and/or when the partial pressure of SO2 becomes enough to boost the belite sulfurization according the Eq. (6). It should be noted that the spurrite and sulfospurrite amounts depend on the CO2 and SO2 levels in the local atmosphere as well as the temperature. Scenario II: The melt formation enhances the reactions leading to spurrite and sulfospurrite formation (Step II-3), according to the equations mentioned above. These reactions are competitive as they depend on temperature and the partial pressure of CO2 (g) and SO2 (g). The sticky surface catches the particles in suspension (Step II-4) and the ring appears. As a final step, the ring consolidation (Step 5) owes to the combination effect of the phase's nature and to the addition and heat removal cycles resulting from the kiln rotation. The ring thickness increases as the kiln rotates and it acquires its compactness through the heating cycles. Nevertheless, the growth rate is related to temperature, presence of a molten phase, SO2 and CO2 partial pressure. 5. Conclusion Extracting more information about the conditions, which govern the ring formation process, represents an efficient tool to better understand the rings appearance and their growth. In this perspective, an analytical investigation was performed on two kiln rings material, which formed almost in the same position (kiln inlet), as well as the operating data that characterized the period associated to each ring. 297
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Fig. 10. Rings formation model in the kiln inlet.
The results yielded by this study convincing evidence that the variation of operating parameters explains the diversity of the mineralogical phases and their proportions, in the rings material. Indeed, the calcination degree is one of the agents that govern the presence of spurrite and calcite. It is imperative to preserve a stable decarbonation amount of the hot meal. Yet, what is most important is to optimize the tolerable range of its variation. Also, it is imperative to continuously check and optimize the oxygen content in the kiln in order to have command over the sulfur cycle. Consequently, this will pilot the formation of ternesite, Ca –langbeinite and chlorellestadite. Also, the amounts determination of aggressive elements (S, K, Cl− and Na) is not enough to deal with the ring appearance. Thus, qualitative and quantitative analysis of the phases was achieved for a better estimation of the reactions governing the appearance of carbonates and sulfates at the kiln inlet.
Fig. 11. Illustration of chemical reactions involved in rings formation in the kiln inlet. 298
Engineering Failure Analysis 95 (2019) 289–299
S. Belgacem et al.
Acknowledgments “Les Ciments de Bizerte” is acknowledged for their support. References [1] Maria del Mar Cortada Mut, Linda Kaare Nørskov, Flemming Jappe Frandsen, Peter Glarborg, Kim Dam-Johansen, Review: Circulation of Inorganic elements in Combustion of Alternative Fuels in Cement Plants, Energy Fuel 29 (2015) 4076–4099. [2] Jagendran Ravindran, Soundarajan Krishnan, Studies on Thermal Analysis of Cement Rotary Kiln Based on Clinker Coating Materials on Refractories, Energy and Monetary Savings. Congress on Recent Development in Engineering and Technology (RDET-16). August 22–24, (2016) Kuala Lumpur (Malaysia). [3] K.E. Peray, J.J. Waddell, The Rotary Cement Kiln, Chemical Publishing Co, New York, 1972 ISBN: 978-08-20603-67-4. [4] J. Nievoll, S. Jorg, K. Dosinger, Corpus, Sulphur, spurrite, and rings-Always a headache for the cement kiln operator ? J. World. Chem. (2009) 78–83. [5] D. Gastaldi, E. Boccaleri, F. Canonico, In situ Raman study of mineral phases formed as by-products in a rotary kiln for clinker production, J. Raman Spectrosc. 39 (2008) 806–812. [6] Maria del Mar Cortada Mut, Linda Kaare Nørskov, Peter Glarborg, Kim Dam-Johansen, SO2 release as consequence of alternative fuels combustion in cement rotary kiln inlets, Energy Fuel 29 (4) (2015) 2729–2737. [7] Anjan K. Chatterjee, Chemistry and engineering of the clinkerization process-Incremental advances and lack of break throughs, Cem. Concr. Res. 41 (2011) 624–641. [8] A.K. Chatterjee, Role of Volatiles in Cement Manufacture and in the Use of Cement, Advances in Cement Technology – Critical Reviews and Case Studies on Manufacturing, Quality Control, Optimization and Use, (2011), pp. 203–236. [9] Haykel Galai, Sarra Ben Hafdhallah, Mineral quantification of preheated cement raw mix, Adv. Cem. Res. 25 (5) (2013) 295–300. [10] G. Goswami, B. P. Padhy and J. D. Panda, Thermal analysis of spurrite from a rotary cement kiln, J. Therm. Anal. Vol. 35 (1129–1136). [11] I. Recio Dominguez, J. Gómez-Millán, M. Alvarez, S. De Aza, L. Contreras, A.H. De Aza, Build-up formation and corrosion of monolithic refractories in cement kiln preheaters, J. Eur. Ceram. Soc. 30 (2010) 1879–1885. [12] W. Kurdowski, M. Sobon, Mineral composition of build -up in cement kiln preheater, J. Therm. Anal. Calorim. Vol. 55 (1999) 1021–1029. [13] Samira Telschow, Flemming Frandsen, Kirsten Theisen, Kim Dam-Johansen, Cement Formation, A Success Story in a Black Box: High Temperature phase Formation of Portland Cement Clinker, Ind. Eng. Chem. Res. 51 (2012) 10983–11004. [14] Mohamed A. Aldieb, G. Hesham, Proceedings of the World Congress on Engineering and Computer Science. 2010 Vol II WCECS 2010, October 20–22, (2010) (San Francisco, USA). [15] T. K. Chatterjee, Burnability and Clinkerization of Cement Raw Mixes, Advances in Cement Technology – Critical Reviews and Case Studies on Manufacturing, Quality Control, Optimization and Use, pp. 69–113. [16] Isabel Galana, Theodore Haneinb, Ammar Elhowerisa, Marcus Bannermanb, Fredrik P. Glassera, Phase compatibility in the system CaO-SiO2-Al2O3-SO3- Fe2O3 and the effect of partial pressure on phase stability, Ind. Eng. Chem. Res. 56 (2017) 2341–2349. [17] B.R. Stanmore, P. Gilot, Review—Calcination and carbonation of limestone during thermal cycling for CO2 sequestration, Fuel Process. Technol. 86 (2005) 1707–1743. [18] H. Bolio-Arceo, F.P. Glasser, Formation of spurrite, Ca5(SiO4)2CO3, Cem. Concr. Res. 20 (1990) 301–307. [19] F.W.Taylor Harold (Ed.), The chemistry of Portland cement Manufacture, Cement Chemistry, second Ed. Thomas Telford, London, U.K., 1997ISBN: 9780727725929. [20] Theodore Hanein, Isabel Galan, Fredrik P. Glasser, Solon Skalamprinos, Ammar Elhoweris, Mohammed S. Imbabi, Marcus N. Bannerman, Stability of ternesite and the production at scale of ternesite-based clinkers, Cem. Concr. Res. 98 (2017) 91–100.
299