Embedded Intelligent Control System For Fused Magnesium Furnace

Fused magnesia with features such as a high melting point, good anti-oxidation, and strong insulating properties, are mainly used in the production of various types of magnesium refractory materials. Fused magnesia is produced, in most cases, by a unique three-phase AC fused magnesium furnace (FMF).The FMF adjusts the distances between the three-phase electrodes and the surface of molten pool appropriately to control the values of the currents running in the three-phase electrodes and therefore the intensity of the electrical arc. The raw materials, namely the magnesite, are melted by the heat released by the arc in the FMF and then coagulated into the final products. The power consumed in the process is the largest cost item in the smelting process, it accounts for more than 60 percent of the total cost. Therefore, the control objective of FMF is to assure that the energy consumption per ton is as small as possible while meeting the product quality constraints. The energy consumption per ton is directly related to the values and fluctuations of the currents running in the three-phase electrodes. Therefore, controlling the currents in a stable and optimal manner in the smelting process is critical in meeting the energy consumption reduction target.The normal smelting process includes three operating conditions:heating and melting, feeding, and exhausting. And it has complex and different dynamic characteristics in different operating conditions. More specifically, the characteristics that lead to important control implications include:1) The operation layer model with energy consumption per ton as the output and the three-phase electrode current as the input exhibits features such as strong nonlinearities, multivariable couplings, unresolved melting mechanism, and difficulties in modeling. Moreover, its dynamic characteristics are affected by changes in its boundary conditions such as particle size and impurity constituent of raw materials, and smelting conditions.2) Frequent changes in the raw material particle size and impurity constituent will cause the arc resistance between the lower end of the electrode and the surface of the molten pool to vary, and thus the smelting currents fluctuate. Consequently, abnormal conditions, such as semi-molten, over-heating, abnormal feeding and abnormal exhausting can arise if the setpoints of electrode currents are not properly adjusted on time.3) Energy consumption per ton cannot be measured online, it can be obtained only by testing the product after the completion of the smelting process.4) The control layer model with the three-phase electrode current as the output and the speed and direction for three-phase drive motor as the input has strongly nonlinear in the feeding and exhausting conditions.In order to optimize the operation of FMF, the control system needs to be able to realize the current setting control, current track and logic control, fault diagnosis, and process monitoring. Because the FMF has harsh environments such as high dust, high temperature, strong electromagnetic interference and so on, traditional programmable logic controller (PLC) together with computer control system is hard to guarantee safe and reliable operations.Therefore it needs to develop an integrated intelligent control system which can achieve the four functions mentioned above in an accurate and reliable manner.Before the work reported in this thesis, the operation optimization of FMF was difficult to achieve by using the existing control methods, and the smelting process was still operated by the manual setting control, thereby leading to compromised product output, increased energy consumption per ton and frequent occurrence of abnormal working conditions. This thesis, supported by the National 973 Program "The integrated operational control methods for complex processes of manufacturing systems (2009CB320601)", describes a comprehensive study of the embedded intelligent control systems of FMF. The main results are listed as follows:1. To minimize the energy consumption per ton, an intelligent operational feedback control of FMF is proposed, which consists of the current closed-loop setting control and switching control method of current setpoints tracking. The current closed-loop control algorithm consists of the case-based reasoning (CBR) current presetting module, prediction model of energy consumption per ton, PI-based current pre-setpoints feedforward, and feedback compensation modules. The switching control algorithm, designed for current setpoints tracking, consists of PID-based current control module for the heating and melting conditions, a rule-based reasoning (RBR) current control modules for feeding and exhausting conditions, and the corresponding switching mechanisms for the current control modules.2. To be able to identify and deal with abnormal conditions of smelting process, this thesis presents a data-driven abnormal condition identification and self-healing control method. The proposed method identifies the abnormal conditions based on rule-based reasoning, and the self-healing control is developed using case-based reasoning to correct the current setpoints based on the identification results. The outputs of the control loop track the corrected setpoints, thereby forcing the process to recover from the abnormal conditions.3. An intelligent embedded control system of FMF is developed to achieve the four functionalities mentioned previously. Hardware platform is designed, which is based on the embedded PC/104 bus, including a CPU board and signal input-output modules, and a signal isolation protection circuit board for the harsh smelting environments. A software platform is developed by using RTAI-Linux dual core real-time embedded operating system. The proposed intelligent operational feedback control method, abnormal condition identification and self-healing control method are used in the development of intelligent control software of FMF, and it also can realize logic control and process monitoring. The intelligent control software of FMF classifies different programs into real-time and non-real-time, for which a real-time hardware abstraction layer (RTHAL) and a control middleware are used to manage and call for two types of programs.4. The industrial experiment and application research of embedded intelligent control system of FMF was carried out in actual smelting process. For closed-loop setting control algorithm and data-driven abnormal condition identification and self-healing control algorithm, the experiments of current setting were carried out with the changes in the smelting conditions and boundary conditions. The experimental results show that the algorithm described above can obtain appropriate current setpoints in a timely manner. Compared with the manual settings, energy consumption per ton is reduced by 3 percent, the product output is increased by 1.2 percent, and occurrences of abnormal conditions are reduced by more than 50 percent. For. the switching control algorithm, the experiments of current control were carried out. Comparative experimental results show that the control effects of using the switching control algorithm can significantly reduce the current volatility. Compared with the manual control methods, the ratio of current tracking error exceeding the permissible range is decreased by 29.8 percent. Actual industrial application operation results show that the embedded intelligent control system of FMF decreases the energy consumption per ton by 6.2 percent, increases the product output by 2.9 percent, while ensuring the integrity and stability and optimal operation of the smelting process. For other similar production processes with features of high energy consumption and high pollution, the proposed embedded intelligent control system of FMF and its development process offer great reference and experience in optimizing setpoints and integrating the production process to achieve sophisticated control functions.