The Improvement of the Control of the Causticizer's Temperature and its Realization Abstract: This paper presents a fuzzy-immune-PID control algorithm after analyzing the characteristics of the causticizer's temperature. MATLAB simulation is performed and compared. This algorithm is then applied to the caustic process. The results show that this control algorithm has good stable-state precision and auto-adaptive ability. Key words: Fuzzy-control; Immune-PID control; Causticizer's temperature Abstract: This paper addresses the characteristics of causticizer temperature control in the causticizing section. A controller based on a fuzzy immune PID control algorithm is adopted and successfully applied to the actual application of an alkali recovery system in a paper mill. Compared with other existing control methods, this algorithm gives the system strong robustness and adaptability. Keywords: Fuzzy control; Immune PID; Causticizer temperature; 1. Introduction In the paper industry, alkali recovery is a very important part. Whether alkali recovery can achieve high yields depends heavily on the proper operation of the causticizing section, the finished product section of the alkali recovery workshop. The temperature inside the causticizer is a crucial parameter, directly affecting the rate and degree of causticization. Higher temperatures decrease the solubility of calcium hydroxide and increase the solubility of calcium carbonate, which is detrimental to the degree of causticization; conversely, lowering the temperature decreases the causticization rate. Furthermore, the quality of causticizer temperature control also impacts the reduction of excess ash, the washing of white mud, the reduction of residual alkali in white mud, and the reduction of total alkali in dilute white liquor. This paper employs a DCS control system based on a Siemens CPU315-2DP to collect and control on-site information in the causticization section. Simultaneously, a WinCC-based host computer is used to monitor and control the on-site data in real time. A novel fuzzy immune PID controller optimization design algorithm is proposed and applied to the design of the temperature control system in the causticization section, achieving excellent control results. 2. Process Flow Overview The green liquor from the combustion section is clarified in a green liquor clarifier and then sent to a green liquor storage tank. It is then pumped to a green liquor heater for heating before entering a digester. In the digester, the green liquor and lime undergo a digestion reaction for approximately half an hour. The emulsion then enters three causticizers connected in series for a causticization reaction lasting about 120 minutes. The causticized reactants enter a causticization liquid buffer tank, which is then pumped into a white liquor clarifier. Here, the white liquor and white mud are separated. The white liquor is sent to a white liquor storage tank and then pumped to the cooking section. The white mud is pumped to a vacuum slag washing machine (No. 1) to extract filtrate, while the remainder is sent to a semi-sediment mixing tank, mixed with green mud, and then pumped into a white mud washing machine to extract a diluted white liquor, which is sent to the combustion section. The white mud is sent to a sediment mixing tank and then to a vacuum slag washing machine (pre-mounted filter) (No. 2) to extract usable filtrate. The white mud is then transported out. The flow chart of the alkali recovery causticization process is shown in Figure 1. [b]2. Control System Design and Implementation 2.1 Hardware System Design[/b] The control system mainly performs functions such as real-time display, trend display, process alarm, PID parameter setting, and control of on-site measurement and control points. The total number of measurement and control points controlled and monitored in this section includes 13 temperature channels, 1 pressure channel, 10 liquid level channels, and 3 flow rate channels. Among these, four measurement and control points need to be controlled: the green liquor flow rate into the green liquor heater, the outlet temperature of the green liquor heater, the temperature of the No. 3 causticizer, and the liquid level of the causticization buffer tank. Therefore, the lower-level PLC system adopts a DCS control system based on the Siemens CPU315-2DP, connected to the ET200M I/O station via the PROFIBUS-DP fieldbus. This not only reduces costs but also enables interconnection with other sections, achieving high-speed data transmission. The schematic diagram of the hardware system is shown in the figure below: 2.2 Fuzzy Immune PID Algorithm Design principle of the immune PID controller: Assuming the number of antigens in the kth generation is ε(k), the output of the TH cells stimulated by the antigen is TH(k), and the influence of the Ts cells on the B cells is Ts(k), then the total stimulation received by the B cells is: S(k) = TH(k) - TS(k) Where, TH(k) = k1ε(k), Ts(k) = k2f(S(k), ΔS(k))ε(k). If the amount of antigen ε(k) is taken as the deviation e(k), and the total stimulus S(k) received by the B cell is taken as the control input u(k), then ΔS(k) = Δu(k). The feedback control law is as follows: K[sub]1[/sub], K[sub]2[/sub], K[sub]3[/sub] control the reaction rate, η[sub]1[/sub], η[sub]2[/sub], η[sub]3[/sub] control the stabilizing effect, and f[sub]1[/sub](·), f[sub]2[/sub](·), f[sub]3[/sub](·) are selected nonlinear functions representing the magnitude of the cell's ability to inhibit stimulation. This paper uses three fuzzy controllers in the design, and uses fuzzy rules to approximate the nonlinear function f[sub]1[/sub]（·）, f[sub]2[/sub]（·）, f[sub]3[/sub]（·）: each input variable is fuzzified by two fuzzy sets, namely "positive" (P) and "negative" (N); the output variable is fuzzified by three fuzzy sets, namely "positive" (P), "zero" (Z) and "negative" (N). The membership functions mentioned above are all defined in the entire (-∞, +∞) interval. According to the principles of "the greater the stimulus received by the cell, the smaller the inhibitory ability" and "the smaller the stimulus received by the cell, the greater the inhibitory ability", the following fuzzy rules are adopted for the fuzzy controller: (1) If u is P and △u is P then f(u,△u) is N (1) (2) If u is P and △u is N then f(u,△u) is Z (1) (3) If u is N and △u is P then f(u,△u) is Z (1) (4) If u is N and △u is N then f(u,△u) is P (1) …… In each rule, the output of each fuzzy controller can be obtained by using Zadeh's fuzzy logic AND operation and the "centroid" defuzzification method. As can be seen from the above principle, the controller based on the immune feedback principle is actually a nonlinear PID controller. Its coefficients k'p, k'i, and k'd change with the controller output. K1, K2, and K3 are the gains. Therefore, the output of the immune PID controller is: u(k) = u(k-1) + K1[1-η1f1(u(k), Δu(k))](e(k)-e(k-1)) +K[sub]2[/sub][1-η[sub]2[/sub]f[sub]2[/sub](u(k),△u(k))]e(k)+K[sub]3[/sub][1-η[sub]3[/sub]f[sub]3[/sub](u(k),△u(k))](e(k)-2e(k-1)+e(k-2)) =u(k-1)+k′[sub]p[/sub](e(k)-e(k-1))+k′[sub]i[/sub]e(k)+k′[sub]d[/sub](e(k)-2e(k-1)+e(k-2) 2.3 Algorithm implementation In terms of algorithm implementation, for the control of green liquor flow rate, outlet temperature of green liquor heater, and liquid level of causticizing liquid buffer tank, since the process requirements are not very high, we use the standard PID control module (FB41). This function block collects field data in real time, and satisfactory control results can be achieved by correctly setting the interface parameters on the WINCC interface. For the control of causticizer temperature, we independently write a function block using the fuzzy immune PID control algorithm discussed above. The output u(k) of the immune PID controller is used to control the pneumatic valve to achieve the control of causticizer temperature. 2.4 System Simulation Let the mathematical model of the above control system object be: , where k is the model gain coefficient, T is the time constant, and τ is the time lag constant. The following simulation study focuses on the temperature of a causticizer in a paper mill. The parameters of this object are: k=3.45, T=18s, τ=30s. The simulation is performed using the M-function in the MATLAB toolbox, with a sampling time of 20s. K1, K2, and K3 are 0.6, 0.3, and 0.1, respectively. η1, η2, and η3 are 0.80, 0.60, and 0.10, respectively. The simulation curve of the system under a unit step signal and external large disturbance is shown in Figure 3. [align=center]Figure 3 Comparison curves of fuzzy immune PID, neural network PID, and conventional PID control[/align] From the above figure, we can see that the fuzzy immune PID algorithm proposed in this paper is not only robust, but also has a good control effect. Although the neural network PID algorithm [6] has a fast response speed, it has a large overshoot and serious oscillation. Moreover, its robustness is slightly inferior to the fuzzy immune PID algorithm proposed in this paper. The self-tuning curves of the controller coefficients k'[sub]p[/sub],k'[sub]i[/sub],k'[sub]d[/sub] during the simulation are as follows [align=center] Figure 4 Fuzzy immune PID control self-tuning curve Figure 5 Fuzzy immune PID control self-tuning curve Figure 6 Fuzzy immune PID control self-tuning curve [/align] 2.5 System Application The following figure shows the WinCC online signal history trend of the causticizer temperature collected from the production site. From the figure, we can see that the temperature of the causticizer is stable at about 100℃±1℃, achieving a good control effect. [align=center] Figure 7 WinCC online historical trend of causticizer temperature [/align] 3. Conclusion The fuzzy immune PID controller proposed in this paper has the characteristics of simple structure, strong adaptability and easy real-time control. It also makes full use of the adaptive capability of artificial immune algorithm, and does not require precise identification of the controlled object. It largely overcomes the shortcomings of traditional fuzzy controller design method which relies too much on expert experience and has a lot of subjectivity. A simulation study was carried out on the causticization temperature control object of alkali recovery. The results show that the controller can not only effectively control objects with large pure time delay, but also has good anti-interference ability and adaptive ability, and has strong robustness. It has achieved good results in actual production. References [1] Liu Jinkun. Advanced PID control and its MATLAB simulation [M]. Beijing: Electronic Industry Press, 2003, 129-134. [2] Peng Daogang, Yang Ping, Wang Zhiping, Yang Yanhua, Liu Yuling. Application of fuzzy immune PID control in main temperature control system [J]. Computer Measurement and Control, 2005, 13(3): 250-251. [3] Xing Xiaojun, Zhang Hongcai, Yan Jianguo. Optimization design and simulation of fuzzy controller based on immune principle [J]. Computer Applications, 2006, 26(5): 1113-1115. [4] Zheng Enrang, Nie Shiliang. Control System Simulation [M]. Beijing: Peking University Press, 2006, 256-261. [5] Jin Yihui. Process Control [M]. Beijing: Tsinghua University Press, 1999. [6] Li Yan, Li Minghui, Wang Mengxiao. Self-tuning PID control algorithm based on BP network and its application in causticizing section [J]. China Pulp & Paper, 2005, 24(4): 42-44. [7] Qin Yuping, Ding Yongkui, Liu Shangui. Technical transformation of causticizing section in alkali recovery workshop [J]. Pulp & Paper Science and Technology, 2002, 21(5): 60-61. [8] Xie Keming, Guo Hongbo, Xie Gang, et al. Artificial immune algorithm and its application [J]. Computer Science and Engineering, 2005(18): 77-80. [9] Jiao Licheng, Du Haifeng. Progress and prospect of artificial immune system [J]. Journal of Electronics, 2003, 31(10): 1540-1548. [10] Takahashi K, Yamada T. Application of an immune feedback mechanism to control systems [J]. JSME Int J Seraesc, 1998: 41(2): 184-191. About the author: Li Guodong, male, born in September 1982. Originally from Tai'an, Shandong, he is currently a 2005 graduate student majoring in control theory and control engineering at Shaanxi University of Science and Technology, with a research direction of industrial automation and intelligent control. Mr. Wang Mengxiao is a professor, doctoral supervisor, visiting scholar in Japan, national expert with outstanding contributions, and vice chairman of the China Paper Industry Association. His main research areas are vertical and horizontal research on automation of the pulp and paper process. His main research areas include process optimization control, Enterprise Resource Planning (ERP), and Computer Integrated Process Systems (CIPS) for pulp and paper processes. He has published 4 monographs and over 40 academic papers. This article was supported by the Shaanxi Provincial Department of Education Fund Project (Fund No.: 07JK192). Contact information: Tel: (0)13571005729 Email: [email protected]