Abstract: An AI intelligent regulator, flow controller, and pressure transmitter were used to measure and control parameters such as temperature, flow rate, and pressure in a small-scale device, with good results. Adding an industrial control computer to form a small-scale distributed control system with the AI ​​intelligent regulator enabled decentralized control and centralized management, further improving the system's functionality and reliability. Keywords: Small-scale device, AI intelligent regulator, industrial control computer, temperature control Introduction The small-scale device is mainly used for oil analysis and catalyst evaluation, determining the optimal operating conditions for the process to guide actual production. Two sets of devices were configured as needed. Each set mainly consists of an oil inlet pump, oil circuit, gas circuit, mixer, reactor, and oil-gas separator. The reactor has a cylindrical structure and is heated by a six-section electric heater (2KW per section), plus 2KW for pipeline insulation, for a total heating power of 28KW. The main parameters measured and controlled during the experiment were temperature, flow rate, pressure, and oil inlet rate. The reactor wall temperature and pipeline insulation control have a total of 14 points, ranging from room temperature to 800℃, with an accuracy of ±1℃. The reactor internal temperature detection has 12 points, the flow control has 2 points, and the pressure detection has 4 points. The control system has 32 input points and 16 output points. There are also 4 points for the start and stop control of 4 metering pumps. The main control parameter is the reactor temperature. The structure of the control system is shown in Figure (1). [IMG=Control system structure diagram]/uploadpic/THESIS/2007/12/2007121814244751688K.jpg[/IMG] Temperature control is achieved by AI708 intelligent regulator. It has the advantages of small size, high reliability, and strong anti-interference ability, and has a variety of control modes to choose from. It has RS232/485 interface and can form a small (DCS) distributed control system. The instrument input adopts K-type thermocouple, the output adopts SSR drive module, and the temperature control element adopts 20A SSR solid-state relay. The reactor's internal temperature is monitored using a WP-80 16-point temperature monitoring instrument, which can input thermocouple signals of various graduation numbers, achieve automatic cold junction compensation, and has a self-stabilizing zero function. Flow control uses a D07-12M/ZM controller, with flow display and setting achieved by a D08-1D/ZM display instrument. This instrument has an external setting function, allowing setpoints to be provided by external devices such as computers. Pressure monitoring uses four pressure transmitters with digital display capabilities. To improve the reliability of the control system, automatic circuit breakers and fast-acting fuses are added to the power supply circuits of each instrument. A dedicated instrument control cabinet is designed, housing all instruments and related components, including the start/stop buttons for the metering pumps, contactors, relays, etc. The input and output signals of various instruments are connected to the field devices through common terminals, giving the control system a complete and integrated design. Control System Commissioning and Operation Since the system has numerous temperature control loops and is a key control parameter, high accuracy in constant temperature control is required. Excessive temperature overshoot is not permitted during the experiment. Therefore, a scheme was adopted to control the reactor wall temperature and detect the internal temperature. This reduces the pure time lag of the controlled object, thereby achieving the required temperature control accuracy. The AI708 intelligent regulator provides five adjustment modes. During debugging, the self-tuning parameter function was used first. After 1-2 self-tuning processes, most control loops met the requirements. For a few control loops that did not meet the requirements, AI artificial intelligence adjustment (ctrl=1) was used. The parameters that needed to be tuned were M50, P, t, and ctl. Since the characteristics of the heaters in each section of the reactor are basically the same, the parameter range selected for individual control loops was M50=680~760, P=25~33, t=200~300, ctl=4. The constant temperature control accuracy was less than ±1℃, and the maximum overshoot was less than 3℃, achieving satisfactory results. For the debugging of gas flow control and pressure detection instruments, as long as the debugging is carried out according to the instruction manual, it is easy to meet the actual requirements. Since all instruments and their components are assembled in the control cabinet, interference issues must be fully considered during installation. Therefore, the power lines and signal lines of the system are arranged separately to avoid mutual interference. Furthermore, considering the grounding issue of the control system, the grounding of various instruments is connected to the control cabinet shell and grounded in a unified manner with the grounding system on site, ensuring the reliability of the system and improving the anti-interference capability. Improvement and Function of the Control System The control system composed of instruments has been put into use on site for more than a year, and all functional indicators meet the requirements, and the application results are satisfactory. However, the operation of various instruments and the recording and statistics of various test data still require repeated participation of operators during the test process, and various data reports can only be completed manually. In order to further improve the automation level of the device and realize centralized management of various functional indicators, computer monitoring function is added on the basis of the original instrument control system to make the control system more complete. The improved control system is shown in Figure (2). [IMG=Improved Control System]/uploadpic/THESIS/2007/12/2007121814245480164H.jpg[/IMG] The computer selected is Advantech IPC610/PⅢ industrial control computer, which is connected to the AI ​​intelligent regulator through RS232/485 conversion interface to form a small (DCS) distributed control system. The various parameters of the AI ​​intelligent regulator can be set and modified by the industrial control computer. The detection of internal temperature and pressure signals and the setting of flow control are realized by using the PCL812PG template. The control system program is designed using the industrial configuration software KingSCADA 5.1 and runs in the WINDOWS98 environment. The main functions of the control system are as follows: (1) Process flow display and real-time control and parameter display of each measurement and control point; (2) Parameter setting and modification. Various measurement and control parameters, high and low limit alarms, real-time data storage time, timed printing time, etc. can be set and modified by the keyboard through the human-machine interface; (3) Data reports, including various forms of statistical reports such as shift reports, daily reports, and monthly reports, can be displayed or printed; (4) Data storage and historical curve printing. The main measurement and control parameters can be stored on the disk at regular intervals, and data for 3 consecutive months can be saved. The corresponding historical curves can be displayed or printed according to the stored data; (5) Over-limit alarm. When the parameter with the alarm limit set exceeds the limit, the parameter in the corresponding position on the screen flashes and gives an audible prompt, while printing the alarm parameter and time. Conclusion The advantages of using an AI intelligent controller and industrial control unit to form a control system are high reliability, low failure rate, high measurement and control accuracy, and suitability for long-term continuous operation in industrial settings. Due to the adoption of a distributed control system (DCS) architecture, the detection and control of small devices are performed by instruments, while process management is achieved by the industrial control computer. When the industrial control computer fails, only the computer management, data display, and recording functions are affected; the field instruments can still perform normal detection and control. When a single instrument fails, only that instrument needs to be addressed, without affecting other control loops, greatly improving the reliability of the control system.