Study on Heating Behaviour of Coal during Carbonization in Non-Recovery Oven

The coking process is based on the transformation of coal into coke at high temperature. In non-recovery coke oven, heating is usually asymmetric, this may be due to the fact that heating is provided by controlled combustion of volatile matter of charged cake at oven crown and sole flue.In the present investigation, a simplified temperature profiles were simulated to typical temperature profiles of the non-recovery coke oven. The selection criterions of ovens were based on the oven temperature and coking time. Based on measurement of heating pattern, a numerical methodology was proposed for predicting temperature of the intermediate point by using Lagrange interpolation method. Further, a MATLAB-based algorithm was used to predict the same heating pattern based on actual temperature profile in a non-recovery oven. The model was tested and validated with actual temperature profile of six industrial ovens. The model also reproduces the main feature of the measured temperature profile and shown good agreement with experimental data of industrial oven. The pred icted temperature profile can discriminate d ifferent operating conditions and works well even at low level of temperature deviation. These variations will affect the heating rate and coking cycle, superimposed on the oven crown and sole temperature pattern. However, the operating temperatures were the most important factor fo r normal operation and found that the heating rate of normal operation was varied in the range of 1.06 -0.62°C/min at a height of 200mm from top & bottom and centre of the charged cake.


Introduction
The coal carbonizat ion is the physico-chemical process, depends on the coking rate, operating parameters, coal blend properties and the transport of thermal energy. The heating rate of coal influences the strength and the fissuring properties of coke. In order to arrive at a ho mogeneous quality, the heating of the coal cake in a coke oven should therefore be uniform over the total length and height of the oven. In addition to this, the plastic layer migration rate influences the level of thermal stress in the resolidified mass and therefore, the level of fissuring.
The strength of coke depends to a large extent on thermal condit ion p rev ailing du ring carbon izat ion. The thermal condit ions are influ enced by oven cro wn and so le flue temperature, negative pressure (suction) and carbonization time. The oven temperature and coking time adopted in a normal operation are not independent factors and they vary inversely. At constant temperature, extension of coking time beyond what might appear strictly necessary is known as the 'soaking time" and allowing the coke to remain in the hot oven during this time is called soaking of coal.
In the coking process, the maximu m coke temperature is partially decided by the total heat supplied between charging of the coal mass in the oven and discharging the coke. Heating rate is strongly associated with the pattern of heat supply during coke making process. A suitable heat requirement and the pattern of heat supply should be selected fro m the point of view of co ke quality as well as minimizing the total heat consumption.
The plastic layer of coal is a highly heterogeneous where intricate physical and chemical equilibria between solid, liquid and gaseous components make the study complex [1]. The formation of a coal plastic layer during coking is a pledge to obtain a coking residue, and the gas pressure arises a layer predetermines a magnitude of coking pressure. The zones, which vary in viscosity [2], in a magnitude of the force to perforate the layer [3], can be separated from the plastic layer. A relatively small amount of the work is connected with the investigation of the "active" role of a coal plastic layer in coking process [4].
Softening, devolatilization, swelling and resolidification are closely related. These all mentioned phenomena are highly dependent on the degree of heating rate. It has already beenreported that all coals irrespective of their rank, can be devolatilized without showing any swelling provided the heating rate is sufficiently slow. The carbonizat ion process may be schematically characterized by the fo llo wing typical equations [5]: Coal → Metaplast (1) Metaplast → Semi-coke + primary vo latile matter (2) Se mi-coke → Coke + Secondary volatile matter (3) For any part of bulk coal charge, co mposition of the volatile matter in it is defined in terms of nine species, like, CH 4 , CO, CO 2 , C 2 H 6 , coal tar, H 2 , H 2 O, NH 3 , H 2 S.

Parameters Influencing Heat Transfer in the Charge
Heat transmission rate to a coal charge in a coke oven is affected by several factors, such as, coal blend, mo isture content, bulk density, oven crown and sole temperature, etc. These factors influence the thermal phenomena, the most important ones are shown in Figure 1.

The Moisture Content
The moisture contentthroughout the coal charge influences the heating; on the one hand a large amount of heat is required for the evaporation of water and on the other hand thermal effects arise from the condensation of water. In addition, the distribution of mo isture largely controls the bulk density distribution within the charge. However, the final temperature of the coke is affected by this to a limited amount.

The Bulk Density
The bulk density of thecharge is an important factor affecting the operation of an oven, its throughput and coke quality. This parameter has mostly depends on the size distribution of the charge, the addition of moisture and binder to the charge. The bulk density in the chamber controls the local heat demand throughout the coke bed and also,the operation of coke oven and coke quality.

The Oven Regulation
The variation in temperature difference between coking processes in ovens depends largely on the width of the oven. The different heating rates were reported for different oven widths such as, 0.45m and 0.55m for 3.4 and 2.7 K/ min, respectively [6].
Considerable amounts of work on carbonization in top charge process at high temperature carried out in various countries have revealed that the application of higher coking rate results in improved coke strength and yield. However, the application of higher coking rate in stamp charged coal may give different results since the mechanism of coking is not same as that of top charge. The effects of coal characteristics and carbonization conditions, e.g. bulk density, flue temperatureand mo isturecontent on swelling pressure have been studied [7][8]. So me studies also reported that the heating rate influences the coke quality. Co ke Strength after Reaction (CSR) increasesand M40 of coke decreases quite significantly with increase in the heating rate of the coal charge [9].
The operating conditions and heating rates of both recovery and non-recovery coke oven are different due to asymmetry of heating. In recovery coke making, the coal mass is heated at a constant rate with the help of secondary fuel and operate with positive pressure. While the non-recovery oven is operated withnegative pressure, the crude gas produced from the coal in the non-recovery oven is first partially co mbusted in the free space of oven above the coal charge. This partially co mbusted crude gas is led through vertical ducts in the side walls (downcomers) into the heating flue system under the oven sole. Here the combustion is comp leted with further supply of air so that the coal layer is evenly heated fro m top and bottom [10]. The process control of the non-recovery coke making is accomplished by: ○ Monitoring crown and sole flue temperature ○ Adjusting sole flue and door induced air ports, and ○ Adjusting flue gas uptake dampers In non-recovery coke oven, it is expected that the heating rate will very with the varying depth of coal cake and temperature o f crown & sole flue. A lso, it is expected that the heating rate for a particular zone in the chamber will be influenced by the heat transfer among the neighbouring zones and the coking volu me of the respective zones. In the present investigation, temperature profiles and heating rates of charged coal cake in different zones in non-recovery coke oven were studied throughout the coking cycle.

Mathematical Model
The primary function of the model developed in this study is to make an estimate, fo r a given coal blend, of the temperature profile at a specified oven crown and sole flue temperature during carbonizat ion in non-recovery oven. In general, heat transfer in the charge is due to a combination of conduction, convection and radiation, as well as heat generated by the reactions and phase change. An equivalent conduction model approach is taken, following Merrick [11][12][13], where the actual thermal conductivity of the material is replaced by an effective thermal conductivity, which takes into account the effects of conduction and radiation. Different mathematical models are also used for different objectives, such as effect of heating rate, numerical simu lation fo r coal carbonization using PHOENICS, different heat transfer model used in coke oven for p redicting heat transfer in coke oven [14][15][16][17][18]. Fro m the literature, it is observed that, most of the works were carried out on recovery coke making and there is less work reported on non-recovery coke making. In the present study,a simplified mathematical model has been developed to predict the temperature profile of non-recovery coke oven. This model assumes transient heat transfer process in one-dimension and is based on fundamental princip les of heat transfer, thermodynamics and kinetics of reactions. A statistical correlation was developed with the experimental values.
For the imp lementation of the mathemat ical models, an in-house interactive MATLAB code named-Lagrange interpolationwas developed to solve the equations of the model numerically to predict the temperature profile. After preliminary processing of the input data for heat recovery oven, the code calculates the intermed iate point temperature at different height with respect to time.The code, which calculates the histories of intermediate point temperature, was extended, using the model of temperature profile to calculate the temperature profile of the charge coal cake. The code requires sole flue temperature, cro wn temperature and temperature profile at five different heights of coal cakeas an inputsand assuming all other parameters were same. The temperature profile of the oven at 50mm d istance with an interval of 2.5 hours can be predicted fro m the model. The description of heat flow fro m oven free space and sole flue to coal cake fro m top and bottom of the oven are shown in Figure 2.

Assumptions of the Model
Several assumptions were made to simplify the case for this study which are as follo w: ○ The charge coal cake depth was 1000mm ○ Total coking time was 64 hours (normal operation) ○ The moisture of coal cake was 9.5±1.5% ○ The properties of charged coal cake was constant ○ Bu lk density of the charge coal cake was constant. ○ Indirect heat fro m bottom (heating at the bottom through 150mm of silica bricks).
○ Direct heat fro m top (heating at the top due to the radiation fro m temperature).
○ Heat transfer in the process was considered in asymmetry condition.

Methodol ogy
Fourier's law of heat conduction is applied across first half of the charge assuming symmet ry. The following equation is obtained. ρC Where, C is the specific heat capacity of the charge, T is the temperature of the charge, k is the thermal conductivity of the charge, x is the space variable and t is the time variable.
The thermal conductivity k can be rep laced by an effect ive thermal conductivity k e for this model.
The above equation is a parabolic equation and can be treated as an initial boundary value problem with following boundary conditions.

Boundary Condition
At the centre ∂ T ∂ x ( 500 , t ) = 0 (due to symmetry) At the surface of the charge near to wall; temperature is nearly equal to that of the wall T ( 0, t ) = 1000 (7) The k e is very difficult to determine, as it varies rigorously with x and t within the oven since the value of C and k is not constant during carbonization. Therefo re, mathematical model based on Lagrange extrapolat ion was chosen to determine temperature histories inside the oven.

Lag range Extrapolation
Where, x i is the points where the temperature was recorded using thermocouples, y i is the temperature at x i andprediction of temperature is to be made at a d istance x. Now, instead of solving the heat equation, the temperature historiesat different thickness of coal cake were pred icted by using the equation (8).A MATLA B code was written to solve intermediate temperature curve p lot of T vs x for different time (t ) intervals.

Experimental
All experimental works were carried out at Hooghly Met Coke (HM C), Tata Steel, India. The temperature profiles of charge coal cake, oven crown and oven sole temperature were recorded with the help of thermocouples throughout the cycle of all experimental oven. The cross-section view of non-recovery coke oven is shown in Figure 3. For wider spectrum of study, six different ovens were selected for present study. These ovens were chosen based on oven performance like good, normal and abnormal conditions. In the blend, crushing fineness varies in the range of 89-91% belo w 3.2 mm size, bulk density varies in the range of 1040-1060 kg/ m 3 and the mo isture was maintained in the range of 9.5-11%in all tests. The properties of charged coal blend, charging temperature of sole flue & oven crown and heating rates of all ovens which are used in this study are shown in Tables1-2.
It is a challenging task to conduct such experiments on an operating coke oven facility. In each of the experiments, a set of five thermocouples denoted T1-T5 were inserted directly into the coal charge, to a depth of 1200mm, via holes in the lower oven door for measuring the temperature patterns. Likewise, the sole flue and crown temperature were recorded with the help of thermocouples namely oven sole temperature (OST) and oven crown temperature (OCT) respectively. The thermocouples OST and OCT were inserted at the oven roofand sole fluechannel, and temperature were recorded with an in interval of 30 minutes at all sevenlocations. The locations of all five thermocouples used in experiments are shown in the Figure 4. Using these temperature profiles, seventeen additional temperature profiles at a particular time were predicted by using Lagrange ext rapolation method. For simp lificat ion of problems, the charged coal cake is div ided into two halves and equations were solved separately in two halves. Now using these data points, the temperature profile o f the oven was obtained at different time intervals.

Results and Discussions
In order to arrive at a homogeneous quality, the heating of the coal cake in the oven should be uniform. The heating rate of coal cake varies widely with varying temperature pattern of oven crown and sole flue throughout the coking cycle. Also, it is expected that the heating rate for a particu lar zone in the chamber will be influenced by the heat transfer among the neighbouring zones and the coking volume of the respective zones.The maximu m heat generation increase somewhat less than linearly with heating rate. As the heating rate is increases, the temperature corresponding to maximu m rate is shifted to higher value, but at the same time another side rate of heating decreases and hence uneven carbonization occurs. Therefore, the sole flue temperature is ma intained appro ximately 50 ℃ higher co mpare to oven crown temperature because at bottom heat passes through 150mm of silica bricks. Table 2 shows the heating rates of all six ovens which were used in present investigation. Results showed that the heating rates widely varied with varying sole flue and oven crown temperatures. Also, in some cases, the heat penetration rateswere not coinciding at the end of 64 hours coking cycle. This may be due to the difference in heating rate at different height of coal cake (200mm, 350mm and 500mm i.e. centre of charged cake fro m bottom and top).
Experimental and pred icted temperature p rofilesobtained using this model isshown in Figures 5-10. Results show that the predicted temperature profile and experimental temperature p rofile relationship are in good agreement. It may also be observed fro m the result that the predicted temperature p rofile can d iscriminate between operating conditions and works well even at lower level of temperature deviation (Figures 5b-10b). These variations will affect the heating rate and coking cycle, superimposed on the oven crown and sole temperature pattern. Therefore, at a lower oven sole and crown temperature, the centre line o f charged coal cake is shifted towards the lo wer temperature zone and the coke pushing is delayed.
The rate at wh ich coal charged is heated varies considerably with variations in both crown and sole flue temperatures. The part of the charge near the sole flue and top of cake will be heated at higher rate at early t ime than at later time, and theserates will be different to those experienced by the part of the charge near the oven centre. Results show that the order of heating rate of all six ovens at different distance of 200mm, 350mm and 500mm fro m bottom and top varies in the range of 0.60-1.00 & 0.78-1.22, 0.50-0.70 &0.50-0.70 and 0.50-0.60 ℃ / min, respectively. This may be due to the effect of sole flue and crown temperature of oven throughout the cycle. It appears that heating rate measured at different points coincide with the passage of the p lastic layers because temperature distribution in different transformat ion phase is not same. Therefore, the rate of carbonization decreases remarkably towards the centre of the charged coal cake. The similar heating pattern of oven no. 483 and 136 were also observed (Figures 6-7). Result shows that the heating rate at a point of 200mm fro m the top (1.22 and 0.99℃/ min) is higher co mpared to 200mm fro m the bottom (0.99 and 1.03℃/ min ) of oven no. 483.The temperature of oven crown rose rapidly, as soon as the coal cake was charged in the respective oven. During initial period of heating, the sole flue temperature was higher co mparing to oven crown temperature. However, after few hours, the oven sole flue temperature goes down or becomessimilar to oven crown temperature upto appro ximately 45 hours. This is due to the higher crown temperature of oven no. 483 (Figure 6a). Also, the similar sole flue temperature trend was observed in oven no. 136. This may be due to non-uniform temperature of all five points at the end of 64 hours (Figs. 6a and 7a) of the coking cycle. Figure 8 shows the temperature profile of oven number 450. Results show that the heating rate varies in the range of 0.60 -0.89℃/ min. It is also observed that the heating rate at the centre of the charge is almost same as oven nos. 303 and 136 but the heating rate at 200mm and 350mm distance from bottom/top are different. Th is may due to the uneven temperature during carbonization. Figures 8 (a) and 8 (b) show that variation of sole flue and crown temperature upto 38 hours, after that the temperature profile was found in normal condit ion and hence, the heating rate at centre was normal i.e. 0.60℃/ min.   These results show that the condition of oven number 406 is better co mpared to oven number 401. It is also observed that the oven regulations of the oven nos. 401 and 406 are not proper. These ovens are under repair because the suction ports of these ovens are chocked. Therefore, the progress of sole flue and crown temperature are not adequate as per desired requirements. Therefore, oven regulations or uniform heating of coke is an important key in coke making. Thus, homogeneous temperature distributions during the coking period are important for productivity, coke quality and battery life.

Conclusions
A mathemat ical model was used to predict the temperature profile of the non-recovery coke oven at different heights of coal cake. The three process parameters considered in this study were sole flue temperature, oven crown temperature and five different point temperatures of coal cake (200mm, 350mm and 500mm fro m bottom and top). The heating rate of normal operation was varied in the range of 1.06 -0.62℃ /min at a height of 200mm fro m top & bottom and centre of the charged cake i.e. 500 mm fro m top and bottom.The mathematical model was developed for predict ing the temperature profile at 50mm distance with an interval of 2.5 hours by using the set of experimental data and mathematical software package Matlab 7. The predicted values obtained using the models were in very good agreement with the experimental values. It was observed that the predicted temperature p rofile of the model can d iscriminate between operating conditions and works well even at lower level of temperature deviation. These variations will affect the heating rate and coking cycle, superimposed on the oven crown and sole temperature pattern.
The model can be also used to predict the performance of carbonization throughout the coking cycle from top and bottom o f coal cake towards the centre of the charge. It was found that an increase or decrease in sole flue and crown temperature pattern resulted in increase or decrease of the heating rates from top and bottom of the charged coal cake. The model also showed a decreasing trend of heating rate towards the centre of the charged coal cake which was actually observed in the plant.