Thermal stress analysis was made to study crack initiation mechanism of refractories (carbon block) at hearth of blast furnace. In the analysis, 'Gap-link Model' was proposed to consider the effect of joint-mortar, stamp and shell which construct the boundary of carbon blocks. Calculated results show good agreement with those obtained by experiment of 1/4 scale testing hearth model furnace. The main results obtained by 'Gap-link analysis' are as follows. 1 In heating process, adjacent carbon blocks come into partial contact with each other. 2 Radial tensile stress occurs at the end of contact region, and this tensile stress is considered to cause cracking at side faces of carbon blocks of the hearth model furnace.
For the purpose of obtaining the maximum coke strength by optimum size preparation of coal charge, a mathematical theory is developed, which clarifies the relation between coke strength, inhomogeneity of charge and grain size structure of charge. Inhomogeneity of charge is represented by variance (V) of caking property parameter (X) and of coal rank parameter (Y), and related to observed coke strength (S) by the following equation, S=S+1/2 ∂2S/∂X2Vxx+∂2S/∂X∂YVxy+1/2 ∂2S/∂Y2Vyy where S is coke strength of homogeneous charge. Furthermore the mathematical solution is derived for each variance, which is expressed by volume and coking properties of each coal particle. The preliminary analysis of coke strength shows that the second derivatives are always negative. The mathematical solution shows that the variance is effectively lowered by the selective fine crushing of coal. The validity of theory is examined by experiments. For practical purposes, the theoretical equation is approximated based on the experimental results.
Tests were conducted on SRC (solvent refined coal) both in a test coke oven and commercial coke ovens to evaluate SRC as an additive in coke making. Our findings are as follows: 1) The addition of SRC to coal charges makes it possible to produce coke suitable for use in large-size commercial blast furnaces even if the original coal charge has low DI15015 and CSR. 2) Compared with other coal charge pretreatment methods, the addition of 7-8% SRC can produce the same coke quality as that obtained from BBCP (Briquette-blend coking process) and the addition of 2-3% SRC can produce the same coke quality as that obtained from the coal preparation control process (CPCP). 3) Noncoking coal can be used for up to 20% of the coal charge under the condition that the amount of SRC addition is equal to half of non coking coal. As mentioned above, it has been ascertained that SRC is very effective as a caking additive. The reason is that SRC contributes to change pore texture and improves the optical texture of the coke produced. It has been also proved that coke strength (DI15015) of coal charges which include SRC can be estimated by applying the SI-CBI method based on coal petrology.
Recent studies on simulation model for coke oven dealt with only heat transfer in half a chamber and the differential equations in them were solved only on the basis of observed values as boundary conditions. Therefore, the effect of the adjacent chamber could not be evaluated and coke oven operation was not satisfied to be predicted. In this article, the authers presented a new simulation model including oven chamber, vertical and horizontal flue, and regenerator with fuel flow rate supplied as boundary conditions. In this model, not only a half chamber, but also a number of chambers were able to be treated. Using this model, some programmed heating systems were evaluated and a new system was proposed. Besides the theoretical studies using the model, the systems were inspected by using both a test oven and a commercial oven.
In order to evaluate the coarse ores which become nuclei in quasi-particles and unmelted particles in sinter, their mineralogical properties were investigated. The granulating ability and the reducibility of coarse ores were characterized on the basis of the porosity of pores with diameters smaller than 100 μ before and after rapid heating up to 1350°C, respectively. The fusibility of a nucleus particle was determined according to the recrystallization degree of hematite grains after heating, which was related closely to the porosity. Sintering tests were carried out by using some kinds of quasi-particles which differed in the porosity of nucleus particles to clarify the influences of the mineralogical properties of coarse ores on the production and the quality of sinter. The productivity was reduced with increasing the fusibility of coarse are as a consequence of the decrease in the permeability of the melting zone in the sintering bed and the strength of sinter cake. The reducibility of sinter depended on that of coarse ores.
Using the various sinter raw materials, the main controlling factors for the granulation and the bedpermeability of sinter raw materials were investigated. The results were obtained as follows: 1) The particle size, the water-absorbability and the added water content were the main controlling factors for the granulation and the bed-permeability of sinter raw materials. 2) By the use of these factors, the granulation and the bed-permeability of the sinter raw materials were uniformly formulated with the numerical model. 3) This model simulation was in good agreement with the observed values. This model can estimate the suitability of material combination for sinter.
A mathematical model for sintering process at Mizushima Works was developed using a super capacity computer (M-200), and some operation cases were simulated to obtain useful operation guidance by this model. One of characteristics of the model is a calculation method for heat capacity and diameter of particles of raw mixture for sinter. The present conception that particles of ore, lime, and coke are considered to be homogenized is the same one as in treating the heat transfer behavior of perfectly mixed gas. Analyses of optimum conditions of temperature pattern in ignition furnace, volume distribution of suction air along strand and coke contents distribution in direction of sintering bed height by the simulation model gave results as follows; (1) For saving combustion gas consumption for ignition furnace, it is effective to keep the peak temperature high and the peak time short. (2) To reduce coke consumption, it is especially important to control the size distribution of coke within a narrow rang. (3) To maintain the same heat hysteresis along the height of sintering bed, the difference of coke contents between upper layer and lower layer must not exceed 0.5%. (4) By the control of air volume distribution along strand and of coke contents distribution in direction of sintering bed height, fluctuations of heat index (Q900) and cooling rate index (CT1100) in direction of sintering bed height can be eliminated, and precise quality control can be promoted in near future.
The solid solution state and the crystal structure of calcium ferrite formed in lime-fluxed iron ore sinters and pellets have been extensively investigated in the synthetic quaternary CaO-SiO2-Al2O3-Fe2O3 system. The results obtained are summerized as follows: (1) The calcium ferrite is a solid solution consisting of CaSiO3-Ca(Fe, Al)6O10. The SiO2 component is found to dissolve into the ferrite continuously up to 12.5 mol% at 1250°C and the formation of SiO2-poor ferrite requires an increase in Al2O3 content. On the other hand there exists a lower limit of CaO·3Al2O3 component at about 2 mol% for the formation of the ferrite. It is also confirmed that the dissolution of SiO2 and Al2O3 components is the essential condition for the formation of the ferrite in the light of phase boundary between hematite and calcium ferrite determined in the present study. (2) Powder and single crystal x-ray diffraction studies show that these ferrite phases crystallize in triclinic throughout the wide range solid solubility of SiO2 and Al2O3 and the lattice parameters of the most SiO2-rich ferrite are found to be a=10.057 Å, b=10.567 Å, c=9.092 Å, α=95.45°, β=114.33°and γ= 64.13°. From Patterson synthesis of measured structure factors, a considerable number of Fe3+ ions are suggested to be situated in the tetrahedral sites. It is proposed that the structure may be stabilized by substituting the Fe3+ ions in the tetrahedral sites by Si4+ and Al3+ ions with smaller ionic radii.
The formation of melts which greatly affect the structure and quality of sinter has been studied by interrupted sintering pot tests. On the basis of the experimental results, it is clarified that the SiO2 content in sinter can be decreased by addition of finely ground quartz. (1) The melt is mainly formed of fine ore particles which adhere to nucleus of quasi-particles. The melt bonds unmelted coarse ore particles together to produce sinter. (2) Addition of finely ground quazrt increases silicate melt and decreases calcium ferrite melt. (3) Addition of quartz ground into less than 1 mm size helps to decrease the SiO2 content in sinter while maintaining the desired quality and production rate of sinter.
From the standpoint of mineralogy, mainly morphology of crystal, hematite in self-fluxed sinter is classified into eight types, and it is confirmed that size degradability during the reduction of sinter intensively depends on such morphology. Especially skeletal rhombohedral hematite, which has not been yet described among natural hematite ores, has an extraordinary destructive force during the reduction. Skeletal rhombohedral hematite crystallizes out at a falling temperature stage of sintering reaction when phases of solid magnetite and liquid slag transit into a temperature range of hematite crystallization in phase diagrams, and magnetite changes to hematite abruptly. The more rapid cooling of sintering, the less such hematite and so size degradation during the reduction is improved. Intensive destructive force is considered to be closely related to mineralogical characteristics, those are, skeletal form and locally generation of such crystal groups in state of parallel intergrowth near open pores of sinter. Also one of stable operations applied above principle has been introduced.
Evaluation of sinter properties during reduction has been studied by a new method on quantification of sinter microstructure (mineral components and porosity). (1) Reduction degradation index is expressed by number of 100300 μm grain size of hematite. (2) Reduction index is described by quantity of Ca-ferrite and porosity consisted of pore size less than 150 μm. (3) Pressure drop during softening test under load is indicated by quantity of slag and total porosity. The pressure drop increases with increasing slag quantity and decreasing the total porosity. (4) Temperature at beginning of melt-down is estimated by quantity of Ca-ferrite and total porosity. The above results are not only statistical, but also substantial because these are supported by investigation during reduction. So these are applied to indices of sinter evaluation.
Isothermal softening tests of pre-reduced sinters under a constant load were carried out in order to clarify their softening properties in terms of softening viscosity. (1) Softening viscosity decreases with decrease in sinter size and shows a minimum at a certain prereduction degree. (2) Softening viscosity of slag phase is smallest in the mineral phases which compose the sinter. It is suggested that formation of slag phase having high softening viscosity improves softening properties of sinter. (3) Softening properties are improved by controlling gangue minerals, especially by increase in CaO/SiO2 ratio. (4) Shrinkage degrees during reduction process under simulated conditions of temperatures and gas compositions in a real blast furnace are in good agreement with the values obtained by integrating the shrinkage rate based on the softening viscosity with respect to the shrinking time.
The energy cost of pelletizing process has become higher than that of sintering process after the first and second oil crisis. In order to solve this problem, the following improvements were made on: (1) The firing system of indurating process has been changed into the pulverized coal injection from oil and coke oven gas injection. (2) The coke breeze has been inserted into the green pellets. (3) The modification for the grate process and process fan has been carried out. As the result of these improvements, the energy cost has become almost equal to sintering process. Also, for the improvement of shape character, crushed pellets have been developed. The operation test was carried out with this new pellets in Kakogawa No. 2 BF, and the successful result was obtained.
Induration test of pellet with addition of carbonaceous material was carried out by using an experimental pot grate furnace. A simulation model for analyzing induration phenomena of the pellet was also developed, in which consideration was given to the combustion of carbon, the decomposition of limestone, the oxidation of magnetite and the heat transfer between gas and solid. (1) Simulation analysis results agreed comparatively well with experimental ones. Therefore this model was considered to be applied to determine the operation conditions of the pot grate furnace and estimate the heat pattern in the pellet layers for commercial pellet plant. (2) As the carbonaceous material content in pellet increased, the pellet temperature of bottom zone became higher than that of upper zone due to the transference of combustion heat of carbon from upper to bottom layer. This resulted in difference in pellet properties between upper and bottom layer. (3) From this result, it was concluded that the limit of the content of carbonaceous material was about 1.0%. (4) It was estimated that the reduction of heat consumption was about 80000 kcal/t for the addition of 1.0% carbonaceous material. (5) The separate addition of carbonaceous material to the pellet bed and the granulation of double layers as to the content of carbonaceous material were found very effective to improve pellet properties.
Conception and evaluation test methods on the qualities of raw materials for gaseous direct reduction as well as the qualities of direct reduced iron are described. Emphasis is on the clarification of the difference of desirable qualities of the direct reduction feeds from those of traditional blast furnace feeds. The raw materials for the gaseous direct reduction should have, 1) high iron content with small amount of such impurities as phosphorous, sulphur and alkali, 2) high reducibility to make retention time of the burden in the furnace shorter, 3) less fines generation characteristic to decrease the furnace pressure drop lower, 4) less clustering tendency to guarantee smooth decending of the burden in the furnace. To evaluate those properties, the static bed reduction test, Linder rotating reduction test and the clustering test were adopted. Small stainless steel basket which had been charged from top of an actual operating shaft furnace were taken out from the discharge point and the properties of reduced materials were investigated. Test results of actual shaft furnace showed good coincidence with those of laboratory scale.
Reduction experiments of wustite pellets with H2-N2 and CO-N2 gas mixtures of various compositions were carried out at 1000°C. The wustite pellets were prepared by partial reduction of pure hematite pellets (about 3.9 g in weight, about 0.62 cm in radius, and 0.200.25 in porosity) in 50%CO-50%CO2 gas mixture at 1000°C. The rate parameters contained in the unreacted-core model were determined by applying the methods of the mixed-control plot and the data-fitting to the reduction data. In the reduction with H2-N2, the rate parameters for both the reactant gas and the product gas should be estimated. The relation between the effective diffusion coefficient in the product layer and the N2 gas concentration XN2, is explained in terms of the dependency of the ternary diffusion coefficient on XN2, but the relation between the apparent chemical reaction rate constant and XN2 cannot be solely explained by that effect. The change of the specific surface area of solid reactant during the reduction should also be taken into account. In the reduction with CO-N2, the effect of N2 gas concentration on the reduction mainly results from the change of the driving force of the reduction.
Activity measurement of FeTO in molten blast furnace(B.F.) type slags by the method of equilibrating molten slag with iron melt or gas involves great difficulties due to fairly low oxygen potential or a small content of FeTO in molten slags. Then, in this work, EMF method was adopted for aFeTO measurement in B.F. type slags. The activity of FeTO (aFeTO) was calculated from EMF of the following cell. Fe(pure solid)/CaO-SiO2-Al2O3(-MgO)-FeTO/O2, Pt. Oxygen pressure in the cell at EMF measure-ment was kept at approximately 10-10 atm, at 1400°C by CO-CO2 mixture. A Pt-O2 electrode with a Pt ring was employed as the standard electrode. The results obtained are summarized as follows: (1) aFeTO in the melt of CaO-SiO2-Al2O3-FeTO (CaO/SiO2=0.41.2 mol% ratio, Al2O3=12.5 mol%) indicates a positive deviation from the ideality in the range of FeTO content 0.210 mol% at 1400°C. Activity coefficient (γFeTO) at 1 400°C increases from 1.4 to 2.0 with increase in CaO/SiO2 ratio from 0.4 to 1.2. (2) The value of aFeTO in the melts of CaO-SiO2-Al2O3-MgO-FeTO indicates less positive departure from the ideal solution behavior in the range of FeTO content 0.22 mol% at 1400°C than those in the melts without MgO. (3) Thermodynamic quantities of mixing of FeTO were calculated by using the data of log aFeTo versus reciprocal absolute temperature (1/T). Gibbs's partial molar free energy of mixing of FeTO in the melts of CaO-SiO2-Al2O3 increases from-17 to-5 kcal/mol with increasing FeTO content in the range of 0.310 mol%.
Chemical affinities of the transfer reactions of Si, Mn, S, and Fe between slag and pig iron in the blast furnace (ASiO2, AMnO, As and AFeO) were calculated from oxygen potential (PO2) and temperature of slag and pig iron and their compositions. The PO2 and temperatures were measured by using oxygen concentration cells for slag and pig iron during tapping from the furnace. The following results were obtained; 1) PO2 in pig iron was larger than that in equilibrium with C and CO gas and thus, oxygen in pig iron was supersaturated. 2) PO2 in slag was one order of magnitude larger than that in pig iron. This PO2 was equal to that in equilibrium with pig iron and FeO in slag. 3) Because ASiO2<0 and AS<0, SiO2 in slag could not be reduced and pig iron could not be desulfurized by slag. Because AMnO≅0, Mn and MnO in slag would be in equilibrium. 4) The magnitude of ASiO2 depended on the characteristics of each of the three blast furnaces examined. 5) The content of FeO in slag could be estimated from PO2 and temperature in slag by using oxygen concentration cells for slag.
The effect of FeO concentration, H2 % and H2O % in reducing gas on the reduction rate of FeO in blast-furnace-type slag has been studied by measuring the weight-loss change of slag in MgO crucible with time at 1450°C. The results obtained are as follows: 1) The reduction rate is expressed as, V=4.6×10-4√PH2(NFeO-0.72PH2O/PH2)g-oxygen/cm2·s 2) A value of 3 × 10-4 cm2/s is suitable for the diffusion coefficient of iron oxide in slag. 3) The order of reduction rate of FeO in molten slag by H2 gas under a blast furnace condition is almost the same as that by solid carbon.
The thermal state and reduction conditions of ores in a lumpy zone have considerable effects on the shape of a cohesive zone and the thermal state in the lower part of a blast furnace. Therefore, a throat diameter probe and a shaft diameter probe were developed as new detection devices for surveying the state of the lumpy zone, and a mathematical model was also developed for estimating the state of the lumpy zone and the outside surface of a cohesive zone in a blast furnace. The inductive model mentioned above was composed of five simultaneous ordinary differential equations concerning temperature and gas compositions, and these were solved numerically by making use of Runge-Kutta-Gill method. Simulations based on the model were carried out using actual operating data of Sakai No. 2 blast furnace, and it was confirmed that the accuracy of the model was fairly satisfactory in regard to the estimation of gas velocity and gas compositions. Furthermore, static characteristics of reduction reactions of ores by CO and H2 gas and temperature distributions in the lumpy zone were discussed on the basis of estimation results.
In order to clarify the dynamic behavior of the cohesive zone in the blast furnace, a simulation model experiment was carried out. One-sixth scale sector model unit was built. Quasi are and coke were charged at the top and hot air (180°C) was injected from the tuyeres. The formation process of the cohesive zone was observed via a glass wall fitted on the front of the unit. (1) The cohesive zone is found to be classified into inverted U, inverted V and normal W types, each of which seems to be attributable to the characteristics of gas flow around the root of the cohesive zone. Among them, inverted V type is most desirable for its efficient heat exchange and low thermal load. (2) As concerns burden distribution, a high ore/coke ratio in the periphery makes the cohesive zone inverted U type and a low ore/coke ratio results in W type. In order to obtain inverted V type, an appropriate permeability of the root, especially at the wall side, is necessary. (3) The cohesive zone tends to become inverted U type with a decrease in the thermal flow ratio. An increase in the blast volume raises the level of the cohesive zone and causes its width to become broad. The permeability of the cohesive layers also exerts a large effect on their formation. The mechanism of the cohesive-zone formation has been discussed.
A mathematical model for the simulation of Si transfer process in the blast furnace was developed. The whole regions of the furnace from the stock level to the hearth were one-dimensionally simplified, and mass and heat transport and main reaction rates in each region were included in the simulation. The feature of the model is that instead of the slag-metal reaction in the hearth the reaction between SiO-gas and iron-droplets in the dropping zone is exclusively taken into account as the Si transfer reaction. When the temperature distribution is so simulated as to be well coincide with the actual in-furnace distribution, the calculated Si contents in hot metal are proved to be in good agreement with the observed values.
By the use of a mathematical simulation of Si transfer via SiO in the blast furnace, main operational factors affecting Si concentration in pig iron were examined. (1) The concentration ranging from 0.4% up to 2.5% was well simulated on the basis of the operational data of Chiba No. 2 and No. 3 furnaces. (2) The observed relation between Si concentration and hot metal temperature was clarified by the concept of thermal flow ratio, which gave unified explanation to the relation. (3) In iron production for steelmaking, the main source of SiO was from ash in coke, and the rate determining step of Si transfer into iron was Si absorption by pig iron from SiO. (4) In the case of foundry iron containing more than 2% Si, the amount of SiO generated from flowing-down slag accounted for over 80% of the generated SiO, and Si transfer rate was determined by SiO generation. (5) Simulation of an operation with a low flame temperature in front of the tuyeres was tried, and it was proved that even in such a small furnace as Chiba No. 2 furnace (1380 m3) an operation with 0.2% Si under a hot metal temperature of 1490°C was possible by reducing SiO generation from coke ash.
Burden distribution in blast furnace is formed through the following four processes: (1) Burden descends by the gravitational force in the large bell hopper, (2) is discharged from this hopper followed by falling onto the burden bed, (3) moves toward the furnace center to form the slope, and (4) changes the angle of the slope with its descent, becoming a base for next charge. The processes (1), (2), and (4) have been studied by experiments and theoretical calculations. The main results obtained are as follows: (1) Burden descent in the large bell hopper is characterized by such a flow pattern that burden descends predominantly in the wall side region of hopper, resulting in the burden mixing and therefore affecting the burden distribution. (2) The flow rate of burden from the large bell hopper is proportional to time to the 1.5th power after the discharge begun. (3) The surface angle of burden charged decreases with the burden descent. This behaviour will be explained by such a theory that the burden has a uniform descending velocity over the horizontal section of shaft.
Oil-less operation of the blast furnace has been carried out on the background of the recent oil market soar-up. During the early period of oil-less operation, the furnace encountered a temporary instability resulting from the increase in slips, but currently the stabilized operation has been maintained mainly due to the application of the humidified blast. In order to clarify the effect of hydrogen upon the blast furnace operation from the viewpoint of softening and melting process of ore burdens, softening under-load tests have been carried out under simulated blast furnace conditions. H2 addition improves the softening resistivity of burdens and shifts the beginning temperature of pressure drop to the higher temperature side by virtue of the reduction enhancement effect of H2 at high temperature. This results in the cohesive zone with narrowed width. Possible influence of the cohesive zone width on in-furnace permeability has been examined by using a simulation model for estimating the infurnace pressure distribution. The calculation in the case of non-humidified blast indicates the increase in pressure gradient in the region of cohesive zone. The humidified blast decreases the possibility of slip occurrence and helps the stable operation.
At Chiba No. 1 blast furnace, fine materials condensed from the gas phase were sampled by the use of a vertical probe inserted into the furnace from the top. After the sampling the materials were examined by chemical analysis and X-ray diffraction. On the basis of the examination the in-furnace circulations of zinc, alkalis and their compounds were discussed through thermodynamic calculation. The results of the dissection survey of blown-out furnace could be well explained by the calculation result based on the measurement. An important role of the circulating materials, especially zinc and its compounds, in the formation of scaffold and inactive zone near the wall also was pointed out.
In order to detect the melting zone profile in a blast furnace, a tracer method has been developed. Including tracers uniformly distributed in a given are layer and analyzing the change of the concentration in tapped metal, the melting process of tracer-containing layer can be estimated. And the profile of melting zone is obtained by transforming the melting ratio of tracer-containing layer to the melting radius at every height in the furnace. The change of the tracer concentration in tapped pig iron is analyzed assuming that the molten metal in the furnace drops smoothly to the hearth and the dilution of the tracer in the hearth acts on a system of first order lag. As the tracer, an Fe-P alloy was used. The melting temperature was obtained as 1350°C by the softening and melting test under load with elevated temperature. The tracer method was applied to Kakogawa No. 1 blast furnace (3090 m3), whose operating conditions were almost same as those of the shutting down, and the profile of melting zone was estimated. To verify the estimated results, the furnace was quenched with water and the internal state was investigated. The estimated profile of melting zone is in fairly good agreement with the inner profile of softening and melting zone observed after the dissection.
In order to seize the operational problems and clarify the various phenomena in a blast furnace operated at the extremely low fuel rate, Nippon Kokan K. K. had decided to execute a trial operation in No. 3 blast furnace (inner volume: 3223 m3, blow in; Jan. 1975) at Fukuyama works. As a result, a monthly mean fuel rate of 396 kg/t was recorded in Nov. 1981. Comparing with the previous result of the same furnace (428 kg/t, Jan. 1979), necessary heat at the lower part of a furnace could be much reduced. Both the high reducibility of sinter and adequite burden distribution control enabled to maintain the high shaft gas efficiency of 97.5%. As the heat flux ratio increasing, temperature level at the shaft lowered and the three stages of thermal reserve zones were observed. The level of cohesive zone also lowered and its shape changed from "inverse V" to "V" bia "W" shape, furthermore some of measured results indicated that melting line would be very close to the raceway. In its transition period, burden desending became a little unstable, but got well again when fuel rate reduced to around 400 kg/t. The essential hindrances to get further low fuel rate in future are seemed to be the reducibility of ore at the short effective zone for its reduction and the fusing capacity of ore around the raceway.
A dynamic mathematical simulation model has been developed in order to study the dynamic behavior of a blast furnace. The model can predict the change of the internal state of the furnace by the numerical integration of a set of first order partial differential equations of the gas and solid state variables in the longitudinal direction. The numerical computation adopted in this model is the method of characteristics to change the partial differential equations to the ordinary differential equations. Ten kinds of chemical reactions including gaseous reduction of ore based on two-interfaces unreacted core model, the heat loss through the wall dependent on the heat capacity of the furnace body, and the heat exchange between gas and solid are considered in this model. The blow-in operation of a blast furnace has been simulated by the model. The good agreement of the simulated results of the change in the temperature of the furnace refractories as well as the top gas temperature and composition has verified the usefulness of the application of the model to the planning of a blast furnace blow-in operation.
In order to carry out a blow-out operation with emptying a blast furnace without those troubles as the channeling, the overheat of top gas and the abnormal drop of hot metal temperature, a system has been developed to establish an operation schedule by means of a two-dimensional gas flow model and an one-dimensional dynamic model. In this system, the two-dimensional gas flow model is used for the simulation of the blast furnace gas flow pattern at each stage in the operation to evaluate the critical blast volume to the channeling and the one-dimensional dynamic model is used to predict the transitions of the stock level, the top gas temperature and the hot metal temperature during the operation to build a blasting schedule to avoid the channeling and a water sprinkling schedule to avoid the overheat of top gas and to set the coke rate just before the operation to prevent hot metal temperature from the abnormal drop. This system was applied to the emptying blow-out operation of Kokura No. 2 blast furnace in March 1981, to give satisfactory result.
A fundamental investigation on combustion of pulverized coal has been carried out by using a vertical experimental furnace. (1) In this furnace, low secondary air velocity and low air ratio give insufficient mixing degree of pulverized coal and air. Consequently, experiments are carried out under the condition that the combustion rate is not influenced by the mixing process. (2) The combustion rate becomes higher with increase of volatile matter content of coal and with decrease of its particle size. (3) From the analysis of experimental results by one dimensional combustion model, the reaction rate constant of Lithgow coal is expressed as follows: ks=3.3×104exp(-24500/RT) (4) The combustion behavior of pulverized coal in the raceway of blast furnace is estimated by a simulation model.
The Oita No. 1 Blast Furnace (currently in its second campaigne) was blown in August 1979 and commenced so called "all coke operation" in March 1980. Further Nippon Steel induced the Pulverized Coal Injection (PCI) technology from ARMCO which uses coal as an auxiliary fuel in place of oil. This technology has been in practice on a commercial basis for many years at ARMCO. On June 26, 1981, Oita was successful in adapting this PCI technology to a large blast furnace of 4000 m3 inner volume involving a super high top pressure with a high blast temperature. Both the equipment and the operation have made good progress up to date. In this report, the mechanical out line and major characteristics of the PCI system have been presented as well as developments of technologies for adaption to a large furnace, especially in respect to the combustion and raceway characteristics and control for stabilized transport as affected by scaling up under a high pressure and a high blast temperature conditions. Furthermore, the details of full scale tests are described, which were conducted by utilizing a single-tuyere test facility which was installed on the Oita No. 2 BF. Operational results have shown a satisfactory distribution uniformity throughout tuyere lines around the furnace. As far start-up operation, the aimed injection levels, projected step up rates of injection and optimization of operational conditions are reported herein together with actual operational performance such as the replacement ratio of PC to coke, stability of hot metal quality, material and heat balance and circumferential balance.