Abstract: This paper analyzes the typical characteristics of large time delay, large inertia, and rapid time variation in the heating and combustion process of coke oven No. 2 in a steel plant's coking plant. It proposes methods to optimize the coke oven heating process control, reduce coking energy consumption, ensure uniform and stable coke maturation, and extend the service life of the coke oven. Computer-optimized cascade control technology is specifically applied to actual production to achieve monitoring and effective control of key parameters in coking production, ultimately realizing fully automated control of the coking plant and providing reliable and scientific decision-making basis for the enterprise. Keywords: large time delay, large inertia, fast time variation, optimized cascade control[b][align=center]The Study of the Coke Oven Optimizing Cascaded Control System YUAN Qi, PAN Lian[/align][/b] Abstract: By the analysis of the great nonlinear, inertia and quick-vary typical in the burning process of the coke oven two of a steel works Coking and Chemistry Plant. Introduce a question of how to optimize the control of heating course, reduce the energy consumption of coking, guarantee the coke is even and steady and ripe, lengthen service life of coke oven. It applied that computer contro1 technology into the actua1 production, realized the surveillance of important influencing factors and the control of accuracy and more efficiency in coking production. Eventually, it realizes automatic control in coking production. It also provides the enterprise with reliable and scientific policy-making basis. Key Words: great nonlinear, great inertia, great quick-vary, optimizing cascaded control 0 Preface The coke oven is an important thermal equipment in the metallurgical industry. It has two products: gas and coke. The coke oven provides coke for blast furnace ironmaking and steelmaking, and also provides gas for industrial furnaces and civilian use. Coke is an important raw material for metallurgy and chemical industries. Its quality and output are directly related to the stability of subsequent industrial production. At the same time, the coke oven is a major energy consumer in steel enterprises. How to save energy consumption and improve the output and quality of coke is the core issue of coke oven control and management [1]. The coke oven is an industrial furnace with a unique structure. In the total energy consumption of coking plants, the amount of gas used for coke oven heating accounts for about 70% of the total energy consumption [2]. Therefore, achieving optimal control of the coke oven heating process is of great significance for reducing coke oven fuel consumption, improving coke output and quality, extending coke oven life, reducing environmental pollution and improving working conditions. 1. Control Principle The cascade control scheme mainly focuses on the controlled parameters of the flue temperature and gas pressure, and the secondary loop. The purpose of introducing the secondary loop is to overcome the frequent fluctuations in the blast furnace gas main pipe pressure. The inner loop control mainly involves feedback control of the valve openings of coke oven gas and blast furnace gas, as well as feedback control of the damper openings of the machine-side flue and coke-side flue. It controls the pressure of the main gas pipe by adjusting the valve openings of coke oven gas and blast furnace gas, stabilizing the main gas pipe pressure at a suitable value to control the gas flow rate entering each combustion chamber. At the same time, it adjusts the damper openings of the machine-side flue and coke-side flue to stabilize the flue suction at a suitable value, ensuring that the gas entering each combustion chamber can be fully combusted [3,4]. Its control law is generally PI regulation, with a sampling control period of 15s. The main loop adopts multi-mode fuzzy control, mainly to ensure that the flue temperature is stable at the given target value, while also playing a certain stabilizing role in the fluctuations of gas pressure. A pressure-based temperature feedback automatic control system was formed, realizing the optimized control of the coke oven combustion process temperature. The block diagram of the cascade control system is shown in Figure 1 [2]. It uses two sets of detection transmitters and two controllers. The output of the first controller is used as the setting of the second controller, and the setting of the second controller is sent to the regulating valve. The first controller is called the main controller. The variable it detects and controls is called the main variable (main controlled parameter), that is, the process control index; the second controller is called the auxiliary controller. The variable it detects and controls is called the auxiliary variable (auxiliary controlled parameter), which is an auxiliary variable introduced to stabilize the main variable. The whole system includes two control loops, the main loop and the auxiliary loop. The auxiliary loop consists of auxiliary variable detection transmitters, auxiliary controllers, regulating valves and auxiliary processes; the main loop consists of main variable detection transmitters, main controllers, auxiliary controllers, regulating valves, auxiliary processes and main processes. [align=center]Figure 1 Block Diagram of Cascaded Control System[/align] The cascaded control scheme adopted in this system uses flue temperature and mixed gas pressure as the controlled parameters of the main and secondary loops, respectively. The purpose of introducing the secondary loop is to overcome the frequent fluctuations in the mixed gas main pipe pressure. Its control law is generally PI regulation with a control cycle of 15s. The main loop adopts multi-mode fuzzy control. The gas pressure is stabilized through the secondary loop to ensure the relative stability of the gas flow rate, and the setpoint of the gas pressure control system is adjusted at certain intervals according to the flue temperature. The flue temperature is replaced by the simulated flue temperature obtained by fitting the temperature at the top of the regenerator. There are inevitably a large number of disturbance factors in the system. According to the different locations of the disturbance factors, they can be divided into the following situations: (1) Disturbances occur in the secondary loop; When the flow rate (pressure) of coke oven gas and blast furnace gas changes, it first affects the temperature of the flue of the whole furnace. At this time, the temperature at the top of the regenerator also changes. The secondary regulator then takes automatic action to change the gas flow rate to stabilize the flue temperature. If the disturbance is small, the secondary loop can generally control it in time and will not affect the flue temperature of the main loop. If the disturbance is large, most of it can be overcome through the secondary loop, and the remaining small part is eliminated by the main loop. Therefore, if the disturbance exists in the secondary loop, the flue temperature can be restored to the set value in time. (2) Disturbances occur in the main loop; When the parameters of the coal entering the furnace change and affect the flue temperature, the main controller takes action by changing the set value of the flow rate (pressure) of the mixed gas main pipe. The existence of the secondary loop also improves the control characteristics, so the flue temperature can be restored to the set value in time. (3) The disturbance acts on both the main circuit and the secondary circuit at the same time; if the disturbance causes the main and secondary variables to change in the same direction at the same time, that is, the main and secondary variables increase or decrease at the same time, the control direction of the two regulators on the actuator is consistent, which strengthens the control effect and improves the control quality. If the main and secondary variables change in opposite directions, the control direction of the two regulators on the actuator is opposite. A small change in the valve opening can meet the control requirements and improve the control quality. For example, if the moisture content of the coal entering the furnace decreases, the temperature of the flue will increase. At this time, the pressure (flow rate) of the mixed gas can be reduced, which can reduce the temperature of the flue and thus achieve the purpose of maintaining the flue temperature stability [3,4]. 2 System Design 2.1 System Structure The block diagram of the coke oven heating optimization cascade control system adopted is shown in Figure 2 [3]. It includes the following aspects: (1) Stable coking time scheme This scheme is an optimized cascade control system combining two feedforwards, two feedbacks and one monitoring, namely: Two feedforwards: feedforward control of heat supply and feedforward control of flue gas suction, that is, the heat supply is determined according to the parameters of the coal charged into the furnace, the final temperature of the coke cake and the coking time, and the feedforward input gas flow rate and pressure are determined according to the parameters of the heating gas and the coke oven operating rate; the feedforward input flue gas suction is determined according to the gas flow rate, calorific value and target flue temperature and air coefficient. [align=center] Fig.2 Coke Oven Heating Optimizing Cascaded Control System[/align] Two feedbacks: target flue temperature feedback and furnace temperature feedback control, that is, the measured value of the flue temperature is indirectly obtained through the correlation between the top temperature of the storage tank and the flue temperature, and the heat supply is adjusted by the feedback of the deviation between the set value and the measured value; the coking index is obtained after the crude gas temperature is measured, and the furnace temperature is adjusted by the feedback of the deviation with the set value. 1. Monitoring: Monitoring of oxygen content in the flue gas (or α feedback), that is, the deviation between the air coefficient calculated from the measured oxygen content of the exhaust gas and its set value, and the set value of the flue gas suction is adjusted by feedback. The furnace temperature control adopts cascade control and the suction control adopts the set value follow-up control scheme. (2) Control scheme when coking time changes This scheme is a scheme that combines optimized cascade control with expert system. It only runs when the coking time changes, and is implemented step by step according to the change of coking time during operation. It has the characteristics of simple operation, stability and reliability. When the specified coking time and furnace temperature are reached, it enters the stable coking time. (3) Temperature detection system collects the temperature of the top space of the heat storage chamber or the temperature of the cross hole to the temperature of the fire channel. Using the vertical fire channel soft measurement model, the average furnace temperature of the entire coke oven is obtained according to the temperature of the top of the heat storage chamber or the temperature of the cross hole. (4) Establish relevant mathematical models and use the model self-learning system to automatically correct the models. The control mathematical model mainly includes the feedforward heating mathematical model, the target fire channel temperature calculation model, the flue gas suction feedforward control model, and the coking end time judgment model. 2.2 Optimization control model of target flue temperature The target flue temperature is the target value of the average value of the flue temperature on the machine and coke sides. It is a major process indicator to ensure the maturity of coke cake within the specified coking time. Reasonably determining the target flue temperature not only affects the quality of coke, but is also crucial for energy saving and consumption reduction [5,6]. If the target value is set too high, the energy consumption per unit product will rise rapidly and it is easy to cause the phenomenon of "coke scraping", which is time-consuming and laborious. If it is set too low, the coke cannot mature. According to statistical analysis, for every 10°C increase in flue temperature, the coking heat consumption increases by 95kJ/kg. Therefore, the determination of the target flue temperature is an important link in energy saving and consumption reduction of coke oven. The factors affecting the target flue temperature are mainly coking time and coke cake center temperature, as well as the bulk density, moisture content, carbonization chamber width, and furnace wall thickness of the coal charged into the furnace. The practical model of the target flue temperature can be obtained by numerically solving the heat transfer equation. The mathematical model mainly includes the following: (1) Feedforward heat supply model This type of model is the most commonly used in coke oven control abroad. The key to the feedforward heating control model is to determine the coking heat consumption. According to the coke oven heat balance theory Q[sub]in[/sub]=Q[sub]out[/sub], the heat transfer equation between the furnace wall and the coal can be summarized as: (2) Target flue temperature model For a newly commissioned coke oven, the coking time needs to be gradually shortened from start-up to production. This process takes at least 1-2 months. It is very important to establish a reasonable heating system for future production by measuring the coke cake center temperature at different coking times. Based on the regression analysis of the measured data and theoretical data of the coke cake center temperature at different coking times, the coke cake center temperature model and the target flue temperature mathematical model can be obtained as follows: Machine side coke cake center temperature model: (3) Flue suction control model Suction control can ensure the reasonable combustion of heating gas, because the quality of combustion depends on the mixing ratio of air and gas, which is called the excess air coefficient (α)[6]. The gas volume is set after adjustment by the feedback control loop, while the air volume is determined by the accumulator suction and damper opening. With the damper constant, the air volume can be controlled by the suction. The gas volume and the controlled mixing ratio (α) are generally kept constant using a setpoint follow-up control scheme. Because furnace temperature changes slowly when using gas for heating, each adjustment of flow rate (pressure) takes approximately 4 hours to noticeably reflect the temperature change. Therefore, the adjustment range of the gas and flue gas suction should not be too large each time. Through analysis of field data, a relationship model between gas pressure and flue gas suction under optimal combustion conditions can be found, and the flue gas suction can be directly adjusted based on gas pressure fluctuations. This ensures the flue gas suction adapts to the gas pressure. The oxygen content of the exhaust gas in the flue gas is measured and corrected to maintain a certain excess air coefficient, ensuring optimal combustion. In practical implementation, because the flow rate (pressure) difference between the turbine and coke sides is significant when using mixed gas for heating, the system can correct the predetermined furnace temperature characteristic curve by judging the deviation between the actual net coking time and the target net coking time. In actual operation at Shuigang Steel, due to limitations, it's impossible to determine the coking end time based on a model. Instead, the coking end time is generally determined by two characteristics of the crude gas before and after the coking end time. The first is the change in crude gas color from yellow to bluish-white; the second is the significant rise and subsequent drop in crude gas temperature before the fire extinguishes. In summary, to optimize coke oven heating, the control system needs to include calculation units such as a feedforward heat supply model, a target flue temperature model, a branch flue suction control model, and a coking end time determination model—that is, a control model. There are multiple ways to obtain the control model; currently, it is mainly obtained through on-site testing and statistical processing or regression analysis of a large amount of production data. 2.3 Real-time Trend Screen Before system commissioning, the flue temperature fluctuated significantly, with the fluctuation range basically stable within ±25℃. After commissioning, the overall temperature is relatively stable, with most fluctuations stabilizing within ±10℃, and the maximum fluctuation being ±15℃. Therefore, the flue temperature has been well controlled. Simultaneously, automatic control of the mixed gas pressure on the coke oven side and the branch flue suction has been achieved. The pressure control deviation of the mixed gas on the coke oven side is mostly within ±50Pa. Meanwhile, the suction control deviation of the flue gas on the coke oven side is mostly within ±5Pa. The real-time trend images of 12 main parameters are shown below: [align=center]Fig.3 Real Time Tendency Picture[/align] 3 Conclusion The new No. 2 coke oven in a steel plant is a typical complex system with large inertia, nonlinearity, and time-varying characteristics. It is subject to many interference factors, with a transition time of 6-10 hours. Furthermore, its process characteristic parameters are significantly affected by factors such as coal loading, coal properties, and moisture content. Traditional control methods are insufficient to achieve good control results. Therefore, an optimized cascade control system for coke oven heating, namely the OCC (Optimizing Cascaded Control) system, was adopted. This system combines cascade control and feedback control methods found in complex control systems and has been successfully applied in actual production. The actual production practice has demonstrated that this control scheme is a suitable system for coke oven control. It can not only adjust the amount of heat supplied to the coke oven in a timely manner but also has strong anti-interference capabilities. Of course, it is not only feasible in theory, but also performs well in actual systems. Generally, the operation effect is analyzed by comparing the gas consumption, uniformity coefficient, and flue temperature changes before and after the commissioning of the cascade control system to assess the degree of improvement in the coke oven operation system and energy saving after a period of operation. (1) The optimization of the standard flue temperature reduced the coke cake center temperature by 5°C, which not only saved the amount of heating gas, but also, according to the NOx generation in the coking process, reduced coke cake center temperature can greatly reduce NOx emissions. (2) The adoption of the OCC process improved the coke oven operating environment and reduced the labor intensity of workers. (3) Due to the reduction in coking energy consumption, CO2 emissions were reduced, which is beneficial to environmental protection. References [1] Yao Zhaozhang. Coking Science (Revised Edition) [M]. Beijing: Metallurgical Industry Press, 1995. [2] Zheng Dongming; Yan Wenfu. Review and application of automatic control system for coke oven heating [J]. Journal of East China Metallurgical Institute, Vol. 16, No. 3, 1999.7. [3] Yan Wenfu; Zheng Mingdong et al. Research and application of mathematical model for cascade control optimization of coke oven heating. Journal of Anhui University of Technology, Vol. 299-302, 2003.4. [4] Er Lianjie. Automatic control system [M]. Beijing: Aviation Industry Press, 1994. [5] Wang Yanchao; Tian Hui. Optimization design of reliable real-time industrial control network. Microcomputer Information, No. 3-2, 2008. [6] Bao Liwei; He Min; Shen Zhiqiang. Computer optimization system for large coke oven. Metallurgical Automation, Vol. 18, No. 1, 1994.1.