Automation and Control System of Reheating Furnaces
Automation and Control System of Reheating Furnaces
The reheating furnace is one of the key equipments in a hot rolling mill. It also consumes the majority of energy needed for the rolling of steel. The reheating furnace is normally used to heat the steel stock (billets, blooms, or slabs etc.) to the rolling temperatures which are suitable for plastic deformation of steel and hence for rolling in the hot rolling mill. Reheating furnace is an important equipment for the hot rolling of steel. It is the heart of any hot rolling mill. The heating process in a reheating furnace is a continuous process. Fuel used in these furnaces can be pulverized coal, liquid fuel, or gaseous fuel. Types of the reheating furnaces used in the rolling mills are (i) pusher furnace, (ii) walking beam furnace, (iii) walking hearth furnace, (iv) roller hearth furnace, and (v) rotary hearth furnace. The rotary hearth furnace is normally used in pipe rolling mills.
The steel stock to be rolled is charged at the entrance of the reheating furnace. During its travel in the reheating furnace, the steel stock is pre-heated, heated, and soaked as it passes through pre-heating, heating and soaking zone of the reheating furnace. At the end of the soaking zone of the furnace, the steel stock is discharged from the furnace by ejector for rolling in the rolling mill. The temperature of the heated steel material at the time of discharged depends on several factors and it can vary in the range of 1,100 deg C to 1,250 deg C.
The production process of a rolling mill needs strict controls on the temperature of steel stock from the reheating furnace for meeting the requirements for rolling. The reheating furnace, which heats the steel stock to the rolling temperature can meet these requirements on a continuous and reliable basis only when the control over the temperature and the output quantity are well coordinated. The reheating furnace moves the cold material from the entering side to the exit side, and during its travel through the furnace, the steel stock is heated by the heat generated because of the combustion of the fuel.
When the steel stock enters the furnace from its entry side, it is first heated to temperatures ranging from 750 deg C to 850 deg C in the preheating zone by the exhaust gases leaving the furnace. The heating zones provide heat directly to the steel stock and normally have temperatures ranging from 950 deg C to 1,150 deg C. Finally, the steel stock enters the soaking zone where it heats further and soaks the heat to have uniform temperature across its cross-section to make it ready for the rolling. The soaking zone maintains the zone temperatures ranging from 1,100 deg C to 1,250 deg C for meeting the requirements of the rolling process.
The charging temperature of the steel stock can range from ambient temperature to 800 deg C. The target exit temperature of the steel stock is governed by the requirement of the process of rolling which is dependent on the rolling speed, stock dimension, and steel composition. Steel quality aspects put constraints on temperature gradient and surface temperature of the steel stock.
The reheating furnaces used for heating the steel stocks are normally considered to have high energy consumption. They also emit good quantity pollutants in the atmosphere since the process used for heat generation is the combustion of fuel. Reheating process also has considerable influence on the economics and the working of the rolling mill.
The reheating furnaces normally have several zones depending on the capacity of the furnace and steel stock to be heated. These zones fall into three categories namely (i) preheating zone, (ii) heating zones, and (iii) soaking zones as shown in Fig 1. Except the preheating zone, all other zones have a number of burners to meet the energy need of the steel stock being heated.
Fig 1 Schematic diagram of a reheating furnace
Heat energy is transferred to the steel stock during its travel through the furnace mainly by means of convection and radiation from the burner gases and the furnace walls and by conduction within the steel stock. Heat energy is transferred to the steel stock by means of convection from the hot burner gases which are in direct contact with the steel stock and by means of radiation from the heated furnace walls and heated furnace roof. The transfer of the heat energy by the radiation is the most efficient way of the transfer of the heat energy. Radiation transfer of heat energy takes place through the useful heat transfer area created by the bed of the steel stock.
The reheating furnaces have high energy consumption. The term specific energy consumption refers to the quantity of energy used for reheating a unit mass of the processed material. The specific energy consumption is a key performance parameter for the reheating furnace and it directly influences carbon emissions and operational costs. Thermal efficiency of the reheating furnace is the ratio of heat energy delivered to the steel stock to be heated heat energy supplied to the furnace through the burners.
In the reheating furnace, the thermal efficiency and uniform heating play an important role in the reduction of energy cost and minimization of metal defects. It is essential to improve the efficiency of furnace by saving energy and to get high yields, less unwanted grain coarsening, and more homogeneity in the products as well as to get better thermo-mechanical properties of the steel. To achieve these requirements, it is essential to adopt complete automation and control system in the furnace. Development and evaluation of the automation and control system of the reheating furnace need precise analysis of the furnace dynamics. The reheating furnace is required to have an efficient automation and control system for the minimization of the energy cost for the generated heat.
To start with, the reheating furnace is to be equipped with basic monitoring instruments such as (i) thermocouples at the appropriate locations for measuring of the temperatures, (ii) on-line oxygen analyzer for measuring the oxygen percentage in the exhaust gas, (iii) pressure transducers for the measurement of the furnace pressure, (iv) control instruments like variable frequency drives in forced draught and induced draught fans, (v) solenoid valves in air and fuel lines, and (vi) motor driven screw feeder and a control circuit. All the instruments and control settings are to be in a closed loop. Feedback from the monitoring instruments is to be received by the control circuit to control the different parameters needed for the optimization of the furnace efficiency.
The objective of reheating furnace automation and control system is to achieve the predefined temperature profile of the steel stock when it is discharged from the furnace for rolling while minimizing the energy consumption needed for the steel heating process. The main functions to be performed by the basic automation and control system for the reheating furnace normally cover at least (i) temperature control for each zone, (ii) gas flow control for each zone, (iii) gas-combustion air ratio control for each zone including the correct quantity of excess air, and (iv) furnace pressure and draught control. Normally reheating furnace control is implemented on a two-level hierarchy, with the first (lower) level aimed at the optimal combustion control (air / fuel ratio control) and the second (higher) level to carry out the set point optimization in order to achieve the minimum energy consumption.
Typically, a reheating furnace is divided into 3 burner zones to 8 burner zones. The temperature of the steel stock through each zone of the furnace is controlled by varying the zone temperature, which normally is assumed to be constant in the zone. The furnace temperature is controlled by varying the fuel flow to the burners of the zone. The temperature control system of the reheating furnace can be divided into three parts on the basis of function. The first part is the heat transfer system of the reheating furnace, the second part is the fuel flow control system through the servo valve, and the third and the final part is a PI (proportional integral) controller.
The automation and control system of the reheating furnace has two types of controls namely (i) sequential control, and (ii) technological control. Sequential control system controls the furnace kinematics, loading and downloading manoeuvres and the control of auxiliaries, like hydraulics, lubrication, and furnace cooling etc. Technological control system controls all those which are related to the instrumentation and regulation of the furnace, like PID (proportional integral derivative) temperatures loops, air / fuel cross-regulation, with precise control of air / fuel ratio, and pressure control, etc.
Automation and control system for the reheating furnace is carried out at three levels namely (i) Level 0 which is the on-off control, (ii)) Level 1 which is a PID controller-based system, and (iii) Level 2 which is a programmable logic controller (PLC) based system. Besides there is Level 3 at the remote location (manager’s office) for monitoring and production planning.
The process control is normally accomplished by the digital control system of the instruments and the routine control strategy is realized through continuous PID controller. Several thermocouples are arranged on the top and side wall of each control section of the furnace respectively to gather the actual temperature in each section of the furnace and then the sampling values are sent to PLC so as to realize the continuous PID control through the difference between the measured values and the set value of gas and air flow, then the opening degree of the burner nozzle in each section is adjusted to control the flow of gas so as to control the temperature.
However, since it is not clear about the combustion of the gas, if this method is adopted, the utilization efficiency of heat medium is very low and the energy consumption is very large. Hence, a kind of improved dual-crossing amplitude limiting full-auto combustion control is introduced and its basic principle is to carry out the control over combustion at the upper part and lower part of each section in normal working time. If necessary, the temperature adjustment signal of the upper part can be regarded as the set control of the dual-crossing amplitude limiting loop and the temperature detection value of the lower part can be used for monitoring the furnace condition.
This principal and subordinate control mode can better coordinate the balance of combustion and heat supply between the upper section and lower section of the furnace to make the combustion of the upper and lower section even. At the same time, it considers the combustion of fuel, playing a good role in energy-saving. A PLC / PC (personal computer) is used to realize this control function.
The process control system of a reheating furnace is an advance system combining with advanced computer software and hardware technology. It adopts replaceable operation station, strong relational distributed database support platform, highly-reliable process control station with large capacity redundancy and distributed I/O (input / output) control sub-system of PROFIBUS process field bus technology. It has advanced a sort of completely-integrated automatic solution and can provide all kinds of automation applications with a uniform technical circumstance.
The items included in the process control system are unified data management, communication, and configuration and programming software. All kinds of technologies can be integrated in the overall system with a global database on the same interface with users. Engineers or technicians can configurate and programme in the same platform for all kinds of applications. The process control system makes the PLC incorporated in the distributed control system (DCS) more easily, embodying the genuine characteristics of the computerized automatic control.
DCS and PLC are useful for the reliable process control system of the reheating furnace. The instrument control system and electric control system both can be realized by using DCS and PLC. The control system of reheating furnace is divided into furnace area PLC control system, roll PLC control system, and reheating furnace DCS control system. The controller of instrument and electric connect to the server of the operator station (OS) by the Ethernet. By using DCS and PLC, the technology of network field bus technology combines the process control system and process computer, intelligent instrument, and drive system to realize the instrument and electric control in one system.
For the reheating furnace controlling process, control system which is a complicated transmit heat and character process, has several peculiarities such as disturb, strong coupling, and great pure lag etc. PID control is a comprehensive controlling system, because of its maturity, familiarity to the technicians and operators, frequently used, broad application in computer DDC (display data channel), data controlling equipment and especially complicated controlling system, and practice proved since it is adapted in different industrial controlling processes.
The key equipments of the reheating furnace are (i) furnace charging equipment, (ii) furnace discharging equipment, (iii) walking system in case of walking beam furnace and walking hearth furnace, transfer roll tables, burner control system, and the tracking of the steel material. The whole control system of the reheating furnace is normally configured with double CPU (central processing unit) redundancy, redundancy ring industrial Ethernet, PROFIBUS DP (decentralized peripherals) bus and PLCs with distributed I/O structure. Fig 2 shows typical detailed configuration.
Fig 2 Configuration of control system of reheating furnace
The process control system of the reheating furnace is also required to carry out the material tracking function. For this, it is required that the system carry out the data transfer function regarding the measuring and weighing data, detecting, charging of material in the furnace and the discharging of material from the furnace with the downstream and upstream processes. Further it has to control the scale loss in the furnace for achieving the different objectives of the reheating furnace which are high yield, low fuel consumption, and low environment emissions so as to achieve these objectives in the automatic mode.
The hardware configuration (Fig 2), shows that the distributed control system of the reheating furnace has two remarkable characteristics. The first one is that the basic automation level adopts redundancy ring structure so as to improve the reliability and stability of the system and the second one is that the information level adopts standard Ethernet structure to make the system having large capacity data communication ability and convenient extensibility. The CPU is the core of the control system of the reheating furnace which jointly form the redundancy system control station together with distributed I/O structure.
The monitoring system adopts industrial control computer and pure-flat colour display unit as the operator station (OS) and the engineer station (ES). In OS, the operator monitors and controls the process through HMI (human machine interface). The operator stations of the system work in online and coordination mode and they are fully transparent and fully fault tolerant, and are normally replaceable with each other. This mode makes this system advantageous in reliable operation, reasonable data distribution, quick operation speed, friendly human-machine interface and convenient to use.
The key of reheating furnace control system is a series of PLCs. There are three kinds of communication networks. These are internal memory reflection network, Ethernet, and PROFIBUS network. The communications among heating, roughing mill, and finishing mill is normally realized by internal memory reflection network. The communication among reheating furnace inner PLC is realized by Ethernet. The communications between PLC, OS, and ES are realized by Ethernet. The communication between PLC and long-distance station is realized by PROFIBUS.
There are a number of PLCs used in reheating furnace. They are used for the operation of instruments, electrical devices, and air recuperator etc. At the OS function menu is carried out. All kinds of production process parameters of temperature, pressure, and flux and so on can be monitored by relative function menu in the OS. The power supply of all PLC systems in the reheating furnace field is realized by a UPS (uninterrupted power supply).
The three-level network system structure (Level 1, Level 2, and Level 3) composing of standard Ethernet network and PROFIBUS DP field bus ensures high reliability, no blockage, and high speed of the data communication of the system. The double-ring fault-tolerant optical fibre network ensures more effectively the field anti-interference ability and high reliability of the data transmission of the network system.
The process control system need to have software configuration which uses Microsoft windows as the operating system software and is to be configurated on engineer station as the software to control the lower-position computer and monitor the upper-position computer. In addition, it is to be provided with Industry Ethernet communication software and redundancy system software. The operator station is to be provided with upper-position computer monitoring software. Besides the functions which can be realized by the common monitoring software, it is to be possible to add other softwares with different controls and programmes for accomplishing more complicated functions.
In order to make the operators master the operation condition of the whole control system of the computer in time, accurately control the stable operation of the system and be convenient in operation, the system is to establish system process flow assembly drawing, and different process flow charts. The process flow chart is the process flow display picture with industrial control parameters, vividly reflecting the production process state of the system and providing the display of the measured values of the detection points of the system. In addition, the system is also to establish the loop display picture and the centralized control over each loop which can be realized in the loop picture by clicking the detailed information including set value, process value, output value, manual and auto switch and upper and lower limit alarm of the process variables on the PID control panel.
The analog quantity is to display value and flow accumulation of the analog quantity in form of centralization. The thermometer is to display the main temperature parameters of the system. Functions of control system is to be as per programme flow of charging and automatic running.
Process control system can have difficulties in its functioning. The fuel for the reheating furnace can frequently haves unstable heat value, resulting in difficulty in control over the temperature. The control method of reheating the furnace temperature and the fuel gas flow as well as the air flow cascade dual-crossing amplitude is to limit the stabilized control set value. The detailed idea consists of taking the temperature of the reheating furnace as the main loop, the fuel gas flow and air flow as the auxiliary loop. With this, the output of the temperature loop of the reheating furnace is to be transformed into the initial set value of the fuel gas flow and air flow.
Important characteristics of the reheating furnace automation and control system
The plant operators are able to boost the overall performance of the reheating furnace with the reheating furnace automation and control system. The advanced combustion optimization system improves those parameters which can make the difference in the quality of the rolled steel product. The reheating furnace automation and control system allows for targeted improvements of the metallurgical properties, for example increased temperature uniformity or minimized negative effects such as steel decarburization or high level of scale formation.
The reheating furnace automation and control system helps to reduce fuel consumption by adopting those heating strategies which minimize the temperature set points and yet ensure the needed final target temperature of the steel stock. At the core of the reheating furnace automation and control system is normally a sophisticated mathematical model which is capable of simulating the heating curves inside the furnace for each steel stock charged. The reheating furnace automation and control system shows the overall furnace efficiency by visualizing the heat balance diagrams. It also provides consumption and emission trends which can be correlated to actual production data.
The reheating furnace automation and control system constantly monitors the heating history parameters of each single workpiece of the steel stock. The system works predictively, meaning it makes consistent projections to estimate the development of the heating path. On this basis, the reheating furnace automation and control system continuously chooses and modifies the temperature set points. When coupled with digital instrumentation, the system also optimizes the digital firing pattern.
The reheating furnace automation and control system is highly efficient in optimizing the combustion parameters during events like sudden stoppages, furnace heating-up time, or changes in the product campaign. The system ensures proper heating repeatability, even with inexperienced operators, although even highly experienced employees benefit substantially from the mathematical model guidance during transitory events.
The reheating furnace automation and control system tracks the consumption and the other efficiency related parameters of the furnace and stores the trends over time in order to allow prompt monitoring.
The reheating furnace automation and control system takes feedback from the roughing mill pyrometer. This function adapts the mathematical model according to the roughing mill pyrometer measurements. The temperature of the piece measured by the pyrometer is compared to the expected temperature and adapts the model accordingly.
The reheating furnace automation and control system normally has the advanced feature of ‘skid marks effect’ prediction. This function allows to monitor the ‘skid marks effect’ by means of a specific algorithm which simulates cold spot effect generated by water cooled skids, and manage the different parameters of the automation and combustion system according to the process requirements.
The reheating furnace automation and control system normally incorporate the pacing model. The model forecasts the time in which each workpiece of the steel stock inside the furnace will be ready to be discharged to fulfill the minimum heating / soaking time requirement of each piece inside the furnace and get the maximum productivity.
Modelling of the reheating furnace
In all types of the hot-rolling operation, the reheating furnace is a critical component determining quality of end-product. Hence, the reheating process needs precise control of the stock temperature and temperature uniformity over the entire heating period. While the energy consumption in a reheating furnace depends largely upon the production conditions such as stock dimension, material grade and throughput, and improved control of the furnace, the firing pattern can lead to indirect energy saving through improving the furnace set-point temperatures. However, the multi-zone cascaded construction of reheating furnaces and the associated thermal inertia of the furnace make the task of furnace temperature control very challenging, particularly on occasions of changes in, for example, target reheating temperature, production rate and / or stock dimension and material grade, and production delay.
In general, the concept of the process control for the reheating furnace is always a challenging issue, since its development leads to an increasing complexity of the utilized control configurations. In the reheating furnace, distinct approaches to design the temperature control scheme is needed. The most common strategy is to design diagonal PID controllers. However, the performance of these controllers is not always satisfactory, because of the existence of large time varying delays and the high interaction between the control loops.
The discharge temperature of the steel stock is not only to be in the correct range but the temperature differentials within the body of the reheated steel stock is also to be minimized. It is difficult to manage precisely the reheating process since there is no way of obtaining this information through direct measurement. To solve this problem, models are used in a reheat furnace at Level 2 process control system. These are mathematical models which calculate the temperature of the steel at each node of a two-dimensional matrix so that bulk temperature as well as temperature distribution can be evaluated. These models are the heart of the Level 2 reheating furnace process control system.
The models used for the regulation of the temperature in the reheating furnace need to have the objectives of (i) reheating the steel stock as per the ‘optimum cycle’ predetermined for each type of material which is accomplished for the complete range of furnace production rates including transitions and delays, and (ii) improve the accuracy of control of the heat supply to the regulation zones using the knowledge of the actual temperatures of the steel. As a result of the models at the Level 2 control, it is possible to reduce fuel consumption, reduce scale generation giving a direct effect on yield and overall mill productivity improvement, and to precise control the target discharge temperature
In this context, in the last few decades, ‘advanced process control’ (APC) systems have received more and more attention. Among APC solutions, the adopted approach is frequently based on ‘model predictive control’ (MPC) technology. MPC strategies, exploiting the knowledge of process models, allow to convert physical control problems into model control ones, taking into account economic aspects.
From another approach, considerable studies in the mathematical modelling, based on the physical properties, have been done in the last few decades. A simplified model, by considering first principles, was first developed. However, this model did not consider the radiative heat transfer and complex phenomena, such as turbulences. Another model with less computationally complexity was introduced, which was valid only for steady state operation. Another novel mathematical model, based on the zone method of radiation analysis and combing this model with CFD has been developed. This model has been extended to a 3D approach. Although several studies have been done for the physical modelling, little attention has been paid for the black box modelling. Since unknown disturbances effect the reheating furnace distinctly, black box modelling appears to be a better approach for the reheating furnace. In some studies, the Auto-Regressive with an eXogenous term (ARX), modelling approach for the reheating furnaces, have been adopted.
In the present-day scenario, the popular control schemes, rely on the framework of the MPC strategies. These approaches, in spite of a good MPC’s performance in controlling the temperature and handling the constraints, are not always practical in case there is model mismatches and in cases there are high plant disturbances, while for a satisfactory operation, a high number of utilized tunning parameters is needed. Further, in several cases, it is difficult to take into account the coupling effects and the interactions of inputs-outputs for the reheating furnace to design the proper optimal control scheme.
A number of studies have been carried out on the modelling, simulation and control of reheating furnace since 1970s, and considerable progress has been made. But the application of these models to real time reheating furnace control is still a challenge because of the lack of appropriate model (both simple and precious) for control computer. For the sake of online optimization control, one of the studies have developed a series of discrete state space models for a variety of reheating furnaces used for the rolling of the steel. The advantages of the developed discrete state space models are (i) simple in structure, (ii) fast in computation, (iii) few memory requirements, and (iv) easy for implementation in digital computer.
In one of the studies, a non-linear predictive control model is designed for a continuous reheating furnace for steel stock. Based on a first-principles mathematical model, the controller defines local furnace temperatures so that the steel stock can reach its desired final temperatures. The controller is suitable for non-steady-state operating situations and reaching user-defined desired steel stock temperature profiles. In the control algorithm, a non-linear unconstrained dynamic optimization problem is solved by the quasi-Newton method. The design of the controller exploits the fact that the reheating furnace is a continuous production process. Long-term measurement results from an industrial application of the controller demonstrate its reliability and accuracy. Fig 3 shows non-linear predictive control model of reheating furnace.
Fig 3 Non-linear predictive control model of reheating furnace
With the arrival of more affordable computing power, the use of model-based control in the reheating furnaces has become widespread. Presently three types of furnace models are available. First types of models are the computational fluid dynamics (CFD) models, which are a class of models based on physical laws related to the fluid flow, mixing, combustion, and heat transfer which are solved on a densely discretized computational domain of the real furnace detailed geometry. These models have relatively high accuracy but at the expense of being computationally intensive, because of the movement of the stocks inside the furnace as well as the charging and discharging process, the transport phenomena in the reheating furnace are periodically transient. Hence, the CFD used for modelling provide more suitable results at steady state, while missing to represent the transient behaviour of the process. Fig 3 shows outline of a reheating furnace under steady state operation.
Fig 3 Outline of a reheating furnace under steady state operation
One of the studies simulated the movement of the stock in a virtual way and reported that the furnace geometry including stock does not change, and the thermal energy contained in a stock is moved from one position to the next. To overcome the issue of high processing time in CFD simulation, another study developed a method for the simulation of reheating furnaces in steady-state, in which the stocks are modelled as a highly viscous fluid. Although lower calculation time can be achieved, compared to transient or iterative approaches, relatively this method only applies to periodic transient reheating operation. Hence, CFD models are unsuitable for simulating the thermal behaviour of transient steel reheating in real or near real-time with varying stock geometry and non-uniform batch scheduling.
The second types of models are semi-empirical models, which do not rigidly adhere to specific physical laws but which rely on measured data as model inputs. They aim to achieve extremely fast simulations but have to compromise on accuracy. These models are frequently used for supervisory temperature control, which relies mainly on a limited number of thermocouple measurements installed along the furnace roof and hearth. Given that these limited thermocouple measurements cannot fully represent the temperature map over large control zones within the furnace, and that their responses are not always representative of the temperature in a control zone, this approach can cause inconsistencies in temperature regulation. As a result, the supervisory temperature control has to compromise on inherent inaccuracies in both measurements and modelling assumptions. The situation can be further exacerbated by unsteady heating demands since highly dynamic thermal behaviours exist within the furnace which can only be captured by thermocouple measurements with time delay.
The third types of models are, ‘black-box’ models which do not involve any specific physical law but normally contain sets of adaptive weights, i.e., numerical parameters which are tuned by a learning algorithm with training data. These models are capable of approximating non-linear functions of their inputs, and include the class of statistical models known as Artificial Neural Networks (ANNs). However, in practice the application of this kind of models is limited by the availability of training data. Alternatively, mathematical models based on the zone-method are widely used for the reheating furnaces. The advantages of a zone model lie in its ability to represent radiation heat transfer accurately within high temperature furnace enclosures with far less discretization requirements for the computational domain. The fluid flow, mixing and heat release from combustion of fuel can be derived separately from other sources hence making the overall computation process much more efficient. Zone based models (so-called zone models) belong to this category.
One of the studies investigated the heating characteristics of stocks in a walking beam reheating furnace using a simplified zone model. The study used an empirical correlation to calculate the convective coefficient on the stock’s surfaces, thereby avoiding the need to calculate for enthalpy exchange due to flows of combustion products. For this, a simple model was used to estimate radiation flux on the stock’s lateral faces rather than calculating any interactive radiation between all faces of the stocks. Another study has developed a three-dimensional furnace model to simulate the thermal performances of a large-scale reheating furnace.
The modelling approach takes into account the net radiation interchanges between the top and bottom firing sections of the furnace and also allows for enthalpy exchange because of the flows of combustion products between these sections. One of the studies has combined the advantages of the classical zone method of radiation analysis and conventional CFD in a robust manner which has overcome the challenges of incorporating three-dimensional flow field within the zone model and considering interactive radiation between stocks. The model has been developed based on first-principles and verified by the trial data from the same reheating furnace as modelled. With all model parameters fixed, simulations have been repeated with different timestep setting. Results have suggested that the virtual furnace model is able to predict the overall thermal behaviour of the furnace with reasonable accuracy. Even with consumer-level PC hardware, the developed model has shown a fairly promising computational efficiency, around 170 times faster than the actual run time of large-scale reheating furnaces, so that it is then successfully incorporated into population based genetic algorithm for multi-objective optimization of reheating furnace operations. This study has also suggested that the developed model is capable of capturing the non-linear dynamic of the furnace and has great potential to be incorporated directly into dedicated furnace control algorithms.