Abstract : This paper introduces an intelligent zirconia oxygen analyzer based on fieldbus, and elaborates on the structure of the entire system and the hardware and software design of the temperature control system. The system adopts an expert controller based on a combination of expert intelligence and PID, and uses the expert system knowledge base to output and correct the PID parameters to change the PID control mode. The experiment shows that the accuracy, reliability, maintainability and testability of the temperature control system of the fieldbus intelligent zirconia oxygen analyzer are improved. Keywords: fieldbus; intelligent oxygen analyzer; temperature control; expert system 0 Introduction With the development of fieldbus technology, traditional analog instruments have gradually given way to intelligent digital instruments and have digital communication functions [1-2]. Based on the development of fieldbus intelligent instrument technology, a fieldbus-based intelligent zirconia oxygen analyzer was designed. Its temperature control system adopts the expert PID control principle, which improves the heating speed and accuracy. 1 Structure of the Fieldbus Intelligent Oxygen Analyzer The intelligent oxygen analyzer based on CAN bus uses the single-chip microcomputer C8051F040 as the central controller. The number of peripheral circuits and interface circuits for system expansion is small, the system reliability and stability are high, and the system function expansion and hardware and software upgrades are relatively convenient. The system's hardware structure is shown in Figure 1. The peripheral hardware circuitry mainly consists of six parts: system calibration, data acquisition, temperature control, calendar clock, LCD display with touchscreen, and CAN bus interface. The LCD display with touchscreen provides a powerful human-machine interface, allowing for the display and modification of relevant signals and adjustable parameters. This system uses a regulated power supply, which has advantages such as a wide applicable voltage range and strong anti-interference capability. The host is a microcontroller-based intelligent instrument; all calculations, processing, and control are completed by software. Modular components are used for the input conversion of oxygen potential and temperature signals, and the conversion of current output. These components are characterized by high reliability and high accuracy. Due to the high integration of the components used, the overall structure is simple, reliability is improved, and use, maintenance, and repair are convenient. Oxygen potential and temperature signals are converted into 0-5V signals by their respective processing modules, and then converted into digital quantities by a multiplexer and A/I converter. The microcontroller calculates the oxygen content according to the Nernst formula. The system has a PID temperature regulation function and controls the heating furnace through a solid-state relay. The system also has a configuration switch, enabling the instrument to operate in different modes. 2 Hardware Design of Temperature Control System The Cygnal 51 series microcontroller C8051F040 is a mixed-signal system-level microcontroller integrated on a single chip. Within a single chip, it integrates the analog and digital peripherals and other functional components required to form an intelligent node for microcontroller data acquisition or control, representing the current development direction of 8-bit microcontroller control systems. The chip has one 12-bit and one 8-bit multi-channel ADC, two 12-bit DACs, two voltage comparators, one voltage reference, one 32kB FLASH memory, a high-speed CIP-51 core that is fully compatible with the MCS-51 instruction set, with a peak speed of up to 25MIPS, and also has a hardware-implemented UART serial interface and a CAN controller that fully supports CAN2.0A and CAN2.0B [3]. Pins 18 and 19 of the C8051F040 are AIN0.0 and AIN0.1 pins, respectively. Since the C8051F040 has an on-chip ADC module, the temperature signal can be directly input to AIN0.0 and AIN0.1 pins after external filtering and amplification. The AMUX operates in single-ended input mode. A thermocouple is used as the temperature sensor. Its voltage, after signal amplification, is sent to the A/D terminal of the C8051F040. After conversion, it is compared with the given temperature value. The value at that moment is calculated using a PID control algorithm and pulse width modulation. After opto-isolation and power amplification, the heating power is controlled by controlling the on/off time of a high-power AC solid-state relay (zero-crossing type), thus achieving temperature control. The cold junction temperature sensing element uses an integrated temperature sensor AD590. The measured temperature is detected by the AD590 temperature sensor, amplified, and directly sent to the AIN0.1 input of the C8051F040 microcontroller. [b]3 Software Design of Temperature Control System 3.1 Analysis of Temperature Control Process[/b] The minimum cycle of heating on/off is 10ms, the shortest heating pulse length is 10ms, and the PID control output is the number of heating pulses. The larger the error, the larger the heating pulse value; the smaller the error, the smaller the heating pulse number. To ensure that the PID output has a certain adjustable range, the sampling period is crucial. Too small a period will result in a narrow range of control quantity; however, it cannot be too large either, otherwise it will reduce the control accuracy. Considering all aspects and experimental tests, the sampling period is set to 2s. Thus, the maximum pulse quantity of the PID output (within one control cycle) is 200. The heating process of the system is divided into three parts: First, in the initial stage of heating or when the temperature difference is large, full-power heating is adopted to rapidly heat up the heating element; Second, when the temperature changes drastically, the output is turned off, allowing the temperature to transition to the next stage under inertia. Because the heating element lags significantly, when the temperature rises too quickly, it will lead to control failure or severe overshoot in the next stage; Third, when the temperature deviation and temperature change are within a certain range, the PID parameters are adjusted online according to expert intelligence to achieve rapid and accurate control. 3.2 Principle of Expert PID Controller Temperature control systems have characteristics such as nonlinearity, strong coupling, time-varying, and time delay. Conventional PID control is difficult to achieve both high precision and speed [4-5]. This paper proposes a composite control method combining expert intelligence and PID, fully utilizing expert experience and the quantitative adjustment characteristics of PID control in the control process. The bus-based intelligent oxygen analyzer requires a system temperature control range of 680-760℃, with temperature fluctuations not exceeding 1℃ after stabilization. To achieve fast and accurate temperature control, an expert system PID control mode is established based on the characteristics of the controlled object, as shown in Figure 2. The inference structure of this expert controller adopts a data-driven forward inference strategy. The production rule adopts the form IFe(n)ANDe(n)TNout(n). The key to the expert system is the establishment and determination of expert knowledge. Knowledge acquisition comes from the long-term summaries of process engineers and from knowledge and analysis in the control field. By combining the expert system and the PID controller, the PID parameters are corrected using the expert system knowledge base, and the PID control mode is changed to achieve the best PID control effect. Based on the object characteristics and design requirements, 99 control rules were designed, and the adjustment methods and parameters under the rules were pre-stored in the controller. The expert control rules determine the control mode and whether the proportional coefficient KP, integral gain KI, and derivative gain KD need to be modified based on the current deviation e(n) and its rate of change Δe(n). During the control process, the controller does not need to tune the PID parameters according to the system identification results or a certain objective function, but adjusts the basic PID parameters according to the current state, i.e.: J = P, I, D, where KP, KI, and KD are the basic proportional, improved integral, and derivative gains, respectively; and and are the correction coefficients for the proportional, integral, and derivative terms under the current adjustment state, respectively. 3.3 Software Flow The system software uses C51 language and is compiled and linked in the Silicon Laboratories integrated development environment. The instrument's temperature range is 0～1000℃, the temperature control is 700℃±1℃, the temperature control setpoint is 680～760℃, and it is continuously adjustable. When the temperature exceeds 800℃, the thermocouple breakage protection function is activated. Data acquisition uses the method of averaging multiple measurements to avoid measurement errors. 4 Conclusion The intelligent zirconia oxygen analyzer based on fieldbus designed in this paper has the advantages of high automation level and simple structure. It has a significant effect on improving the production efficiency of thermal equipment such as power plant boilers and coke ovens. The experiment shows that the temperature control system has the advantages of wide measurement range, high operating temperature, reliable operation, timely and accurate measurement, while overcoming the disadvantages of previous instruments such as instability and easy damage. With appropriate modifications, this method can also be applied to other temperature control systems. References : [1] Li Chunwen. Design of fieldbus control system [J]. Refining and Chemical Industry, 2003, (4): 32-34. [2] Li Jian, Xu Liping, Ren Dezhi. Design of hydraulic intelligent instrument based on fieldbus [J]. Machine Tool and Hydraulics, 2003, (2): 217-218. [3] Tong Changfei. Development of C8051F series single-chip microcomputer and C language programming [M]. Beijing: Beijing University of Aeronautics and Astronautics Press, 2005. [4] Sun Zhiying, Tong Zhensheng, Zhao Wensheng, et al. Fuzzy self-tuning PID superheated steam temperature control system [J]. Journal of North China Electric Power University, 2001, (10): 33-38. [5] Huang Yu, Wang Dongfeng, Han Pu. Fuzzy self-tuning PID control and its application in superheated steam temperature system [J]. Electric Power Science and Engineering, 2004, (3): 37-40.