Abstract: As industrial control objects become increasingly complex, traditional PID control can no longer meet the requirements. This paper introduces a fuzzy control system based on Lonworks fieldbus and presents the three-layer structure of the system. Taking a temperature controller in a chemical production process as an example, the design method of a smart node utilizing fuzzy control is also introduced. The control effect is significantly improved compared with traditional PID control. Keywords : Lonworks-bus, fuzzy control, intelligent node, temperature 1. Introduction In chemical production, temperature is usually an important control parameter. For some complex chemical production processes with precise process requirements, it is difficult to overcome the influence of process disturbances by using traditional PID control. For example, if a set of effective fuzzy control strategies is formulated based on the control experience for specific process conditions, precise and stable temperature control can be achieved. Fieldbus is a fully distributed intelligent, bidirectional serial digital communication link that directly communicates the measurement, control and execution equipment and higher-level automation control equipment in the production site. It is an open control system. Among them, LON (Local Operating Networks) bus is a comprehensive local operating network launched by ECHELON in 1991. It is widely used in metallurgy, chemical industry, power and building automation to realize the comprehensive networked field measurement and control of the system. If fuzzy control is combined with LON fieldbus to establish a fuzzy control system based on Lonworks technology, it can implement field-level fuzzy control and realize the upper computer control of complex fuzzy algorithms. This paper takes a chemical reaction production process in a chemical plant as an example [3] and introduces a temperature fuzzy control system based on Lonworks technology. The design method of the intelligent node is described in detail. The chemical reaction production process is as follows: First, several chemical raw materials are mixed in a certain proportion to form a mixture, and then another chemical raw material A is added to it to produce the required product through a chemical reaction. In order to ensure the quality and yield of the product, the optimal temperature for process control is T℃. Since it is an exothermic reaction, the reaction temperature is controlled by adjusting the flow rate of chilled water. This temperature control system has characteristics such as nonlinearity, time-varying, noise interference, and pure time delay, which are difficult to describe with a precise mathematical model. Therefore, the traditional PID control method is difficult to achieve good control results. 2. Design of temperature fuzzy control system 2.1 Structure of fuzzy control system Based on the requirements and characteristics of process control, a large amount of historical curve data such as temperature, raw material A addition rate and chilled water temperature during the production process were analyzed, and the operating experience of skilled operators was summarized and organized. Finally, the temperature fuzzy control system of "three inputs - single output" was determined. Input variables: (1) Reaction temperature: t, unit: ℃ (2) Change in reaction temperature: Δt: t(n) - t(n-1), unit: ℃. In the formula: t(n) is the reaction temperature at the current nth sampling time, t(n-1) is the reaction temperature at the previous sampling time, and the sampling period is set to 5s. (3) Raw material A addition rate: v, unit: kg/h Output variable: opening degree of chilled water flow regulating valve: u 2.2 Fuzzy subsets of each fuzzy variable ① The basic domain of reaction temperature t is [(t-t0), (t+t0)], and the domain of its fuzzy subset T is [-4, 4], where t0 is the maximum temperature deviation that may be reached in production; ② The basic domain of reaction temperature change Δt is [-3℃, 3℃], and the domain of its fuzzy subset ΔT is [-3, 3]; ③ The basic domain of raw material A addition rate v is [0, 1200kg/h], and the domain of its fuzzy subset V is [-2, 2]; ④ The basic domain of chilled water valve opening degree u is [0, 100%]. The corresponding fuzzy relationships are shown in Tables 1, 2, and 3. Among them, t1, t2, t3, and t4 are possible temperature deviations in the control, and t0>t4>t3>t2>tl. The precise value of u will be given directly in the control rules. Table 1 Fuzzy relationship between reaction temperature t and its fuzzy subset T Table 2 Fuzzy relationship between reaction temperature change Δt and its fuzzy subset ΔT Table 3 Fuzzy relationship between the rate of addition of raw material A v and its fuzzy subset V This system has established a total of 60 fuzzy control rules. According to the control rules, the following fuzzy control lookup table [3] is finally obtained, as shown in Table 4. Among them, UF is the intermediate value introduced to determine the valve opening u when considering the rate of addition of raw material A v. Its relationship with the fuzzy subset V of the rate of addition of raw material A is shown in Table 5. Table 4 Fuzzy Control Quantity u (%) Lookup Table 5 Relationship between UF and Fuzzy Subset V of Raw Material A Addition Rate [align=center] Figure 1 System Structure Diagram[/align] [align=center] Figure 2 Temperature Control Node Structure Diagram[/align] 3. System Structure Design The system structure is shown in Figure 1, consisting of three main parts: host computer, LON bus, and intelligent nodes. The host computer is primarily responsible for the installation, maintenance, and management of the LON network, and can monitor the temperature in real time. It also establishes a measurement value database, archives and tables the data for querying and printing. The host computer can also act as a Web server connected to the Internet for remote monitoring. Intelligent nodes mainly include temperature control nodes and temperature measurement nodes. The temperature measurement node measures the temperature and linearizes nonlinear values ​​to ensure sufficient accuracy and linearity, and periodically sends the corresponding digital values ​​of the temperature to the LON bus for processing by the host computer. The temperature control node uses a fuzzy control algorithm to control the cold water valve and is responsible for displaying and uploading the temperature sampling values ​​at each sampling point. 4. Intelligent Node Hardware Design The intelligent node uses the Neuron chip MC143150 with external memory. It has three on-chip CPUs: a network CPU, an application CPU, and a media access CPU. These are connected to the I/O port driver circuit, timers, on-chip memory, and network communication interface via an 8-bit data bus and a 16-bit address. The chip has 11 programmable I/O port objects. Different pin configurations provide flexible interfaces for external hardware, enabling the implementation of different I/O objects. 4.1 Temperature Control Node Design The temperature control node mainly includes: the Neuron chip MC143150, external program memory, D/A converter, actuator, display circuit, and bus transceiver, as shown in Figure 2. The D/A converter uses MAX7228, the display circuit consists of MAX7219 and related driver circuits, and the actuator includes AD694 and related peripheral circuits. The actuator is a rotary electric actuator used for cold water valve control. The knowledge base data used for fuzzy inference is stored in the E2PROM memory of the neuron chip, with a set of initial values. During system operation, new control parameters can be obtained from the host computer via the LON bus, thereby updating the control parameters. 4.2 Temperature Measurement Node Design The structure of the temperature measurement node is shown in Figure 3. It includes a neuron chip MC143150, program memory, temperature sensor, optocoupler MOC3020, A/D conversion circuit ADC0809, FTT-10A transceiver, etc. [align=center] Figure 3 Temperature Measurement Node Structure Diagram[/align] 5. Intelligent Node Software Design The node application program is written in Neuron C language. Neuron C is a dedicated language for neuron chips, an extension of ANSI C, and adds some powerful functions, such as network variable types and event scheduling statements. The task scheduling of the neuron chip is event-driven. When a given condition becomes true, a piece of code associated with that condition is executed. The software design of this intelligent node includes a main program, an A/D conversion program, a D/A conversion program, a display subroutine, a control algorithm subroutine, etc. The following is a partial source code example of implementing fuzzy control by looking up a table[2]: signed short fc(float-type*input1) { …… // Set local variables if (mcc==1) sp=sp1; pe=e; // Record the previous state of the deviation fl_sub(input1, &sp, &e); // Calculate the deviation get e fl_neg(&range_e, &f1); // Limit the deviation if (fl_it(&e, &f1)==TRUE) ce=f1; else if (fl_gt(&ce, &range_ce)==TRUE) ce=range_ce; fl_mul(&e, &f1_6, &f1); // Perform range transformation on the deviation fl_div(&f1, &range_e, &f1); fl_add(&f1, &f1_6, &f1); fl_round(&f1, &f2); // Round the transformed error rol = low-byte(1ro1); fl_mul(&ce, &f1_6, &f1); // Perform range transformation on the deviation change fl_div(&f1, &f1_6, &f1); fl_add(&f1, &f1_6, &f1); fl_round(&f1, &f2); // Round the transformed deviation change lcow = fl_to_ulong(&f2); cow = low_byte(1cow); table_u = table[ro1][cow]; …… fl_from_ulong(ltable_u, &f1); // Perform range transformation on the lookup result fl_mul(&f1, &range_dtu, &f2); // Look up the fuzzy control table fl_div(&f2, &f1_6, &f1); …… return f_out;// Return the output control increment [align=center] Figure 4 Temperature Curve Comparison[/align] After the system was put into operation, it achieved better control effect than the traditional single-loop PID control method, as shown in the temperature curve comparison in Figure 4. It can be seen from the figure that the fuzzy control has a short transition time and a small overshoot, which meets the requirements of process production. 6. Conclusion Fuzzy control technology has been widely used in industrial processes, household appliances and other fields in China, but the network application of fuzzy control technology is still rare. This paper combines Lonworks technology with fuzzy control technology and realizes real-time measurement and control through a host computer, achieving good control effect in practical applications. The system can also make full use of host resources, so that the fuzzy control algorithm is located at the upper layer, thereby binding multiple device nodes to facilitate the construction of different fuzzy controllers. The control parameters can be input by the user through the human-machine interface, which is highly versatile and flexible and convenient to operate, providing an effective way for the integration of fuzzy control generators and field equipment. The author's innovation is that the Lonworks technology is combined with the fuzzy control technology, and the design method and example of the intelligent node are given. The network monitoring is realized through the computer, and the temperature measurement and control can be implemented remotely. The control effect is good in the application. References [1] Fu Xiaofeng et al. Research on the application of Lonworks technology and fuzzy PID control in central air conditioning system [J]. Electrical Drive Automation, 2005, 27 (2): 23-26 [2] Zhong Liyuan, Pang Xiaohong. Implementation method of fuzzy controller based on Lonworks fieldbus [J]. Computer Simulation, 2005, 22 (10): 155-158 [3] Feng Xiaojun, Zhao Xin. Application of fuzzy control technology for temperature in chemical production [J]. Jianghan Petroleum Technology, 2005, 15 (2): 60-62 [4] Li Jun'e, Li Lilan. Application of LonWorks bus in boiler fuzzy control system [J]. Microcomputer Information, 2006, 1-1: 31-32