1. Introduction Currently, most compound fertilizer production lines are manually operated, resulting in poor working conditions, high labor intensity, and inaccurate ingredient mixing, often leading to unstable product quality due to significant human factors. Jinzhou Special Fertilizer Plant originally had a 100,000-ton/year NPK compound fertilizer production line, which was upgraded to a 300,000-ton/year production capacity in 2000. To further improve production and management, the plant installed a complete computer control system for compound fertilizer production. Production Process and Key Control Points: The production of NPK compound fertilizer is divided into three sections: phosphoric acid and sulfuric acid, conversion and compound fertilizer production, and granulation. The first production stage primarily produces phosphoric acid and concentrated sulfuric acid, including phosphoric acid extraction and sulfuric acid roasting. In the conversion and compound fertilizer stage, sulfuric acid and potassium chloride are first added to the reactor in proportion, heated to 110℃ and kept at that temperature for a fixed time before entering the collection tank. Phosphoric acid is then added to the collection tank, mixed, and pumped into a tubular reactor via a mixed acid pump. Finally, neutralized ammonia is added to the tubular reactor to produce the finished slurry. In the granulation stage, the finished slurry is sprayed into a granulator for granulation, and then dried to become the final compound fertilizer. The control loops in production mainly include a ratio control system for various raw materials, a temperature control system for the reactor and granulator, and a spray flow control system. Switching controls mainly include the start/stop of various mineral slurry material frequency converters and the opening/closing of various reactor valves. Some equipment requires sequential control and interlocking control. Monitoring parameters include temperature, pressure, flow rate, liquid level, and other parameters. 2. Control System Composition The control system adopts a three-layer model: lower-level machine + upper-level machine + server. The lower-level control system consists of two Eurotherm E2500 mini DCS control units and one Advantech ADAM5510 PLC; the upper-level control system comprises two industrial PCs, serving as hot backups for each other; the server is a standard PC. The specific structure is shown in Figure 1. Figure 1: Structure of the Computer Control System for Compound Fertilizer Production. The E2500 supports multiple communication protocols such as Modbus, Profibus, and OPC, offering rich control functions. The ADAM5510 is an Advantech PC-based product, supporting the Modbus protocol and programmed in C language for convenience and flexibility. The upper-level control system and the server run the iFIX runtime and development versions of industrial configuration software, respectively. The three computers are interconnected via Ethernet. The connection between the upper-level and lower-level control systems uses a 485 network with the Modbus RTU communication protocol at a communication rate of 19200bps. The ADAM5510 is installed in the field, and an RS485 repeater is added to the system. 3. Main Functions and Implementation of the System The control of the production line is mainly completed by the lower-level computer, while the upper-level computer provides a human-machine interface for each work section, performing functions such as data display, storage, and statistics. The server backs up the operation and data, and analyzes and manages historical data. 3.1 Main Functions and Implementation of the Upper-Level Computer On the upper-level computer, the monitoring interface for each work section can be accessed through the main interface. Other work sections can be accessed from within each work section's monitoring interface. The relevant process flow is displayed in the work section's monitoring interface, and process parameters are displayed in the corresponding positions. To accommodate the habits of on-site personnel, the system has a parameter list interface, displaying all system monitoring parameters in tabular form. The main control loops and monitoring parameters of each section are as follows: (1) Phosphoric acid and sulfuric acid section • Control loop: ratio control of extraction slurry flow and extraction sulfuric acid flow; single-loop PID control of flushing water flow; • Monitoring parameters: start and stop of extraction slurry frequency converter and extraction sulfuric acid frequency converter; sulfuric acid peroxy-roasting temperature, acid tank level, etc.; (2) Conversion and compound fertilizer section • Control loop: three-ratio control system of sulfuric acid flow, potassium chloride flow and phosphoric acid flow; neutralized mixed acid flow and neutralized ammonia flow ratio control system; • Monitoring parameters: start and stop of sulfuric acid frequency converter, potassium chloride frequency converter and phosphoric acid frequency converter; hydrochloric acid flow, water flow, mixed acid tank level, neutralized ammonia pressure; (3) Granulation section • Control loop: single-loop PID control of spraying flow A, B, C; single-loop PID control of temperature; • Monitoring parameters: head temperature A, B, C; tail temperature A, B, C; start and stop of finished material and return material belt frequency converter; belt material gravity and speed signal; motor speed. In the corresponding section monitoring interface, each control loop has a corresponding PID operation panel, allowing for PID parameter setting and manual/automatic switching, similar to operating conventional instruments. For ratio control systems, ratio settings are convenient. The ratio control system also includes ratio disconnection/connection functions. The two host computers serve as hot backups for each other, operating on different interfaces during monitoring, greatly expanding the monitoring scope for operators. 3.2 Main Functions and Implementation of the Lower-Level Computer The lower-level computer specifically implements loop control and parameter monitoring. The main process flow is monitored by two sets of E2500 systems, while ADAM5510 monitors the finished material and return conveyor belts. The E2500 control device is feature-rich and flexibly configurable. All parameters have corresponding Modbus addresses, allowing direct reading and writing by the host computer. The E2500 control device has a standard PID control algorithm, which can be directly used for single-loop and ratio control systems. For reactor temperature control and other aspects of the system, where the system exhibits strong nonlinearity with changing operating conditions, adaptive PID control is employed. Multiple sets of PID parameters are set simultaneously, and the system automatically switches operation under corresponding operating conditions to ensure control quality. Some parameters of the E2500 have Modbus address source addresses. This feature allows for convenient online switching of system connection relationships. For example, when manually operating or debugging a ratio control system before commissioning, it's necessary to disconnect the ratio connection and temporarily convert the setpoint of the corresponding PID control to a local setpoint. This can be easily achieved by running iFix script commands to modify the setpoint source address. In the monitoring of finished materials and returned materials, the main focus is on collecting gravity and velocity signals to calculate instantaneous and cumulative flow. The system utilizes the ADAM5510's COM1 port to extend a simple keyboard and display, facilitating local parameter setting, tare control, calibration, and other operations. 3.3 Reports and System Data Management There are various methods for implementing reports. The system uses VB6.0 to import EXCEL objects into the application, adding custom macros to the pre-defined report templates. Unnecessary items in the EXCEL menu bar and toolbar are disabled, leaving only page setup, print preview, print, and exit options, making it convenient for operators. Daily reports are printed and stored at 8:00 AM sharp. These reports include instantaneous flow rates, cumulative flow rates, return material quantities, finished product quantities, and historical records of temperatures and pressures at various points. Historical records are generated every half hour, summarizing data from the past 24 hours to produce the daily report. On the first day of each month, the total amount for the previous month and the average daily consumption of various materials are calculated to generate a monthly production report. Historical data design, acquisition, and display utilize iFIX's HTA, HTC, and HTD, respectively, generating a file every 24 hours and storing it on the hard drive. Through trend displays, historical reports, and data analysis, production conditions are evaluated, and the production process is continuously optimized. 4. System Reliability Design The system incorporates redundancy design at both the power supply and host computer levels, as well as the E2500 control unit. Two industrial control computers operate, with one running and the other as a hot backup, mutually monitoring each other through iFix EDA programming. Both industrial control computers are connected to the E2500 via RS-485 communication interfaces. The E2500 system employs dual-CPU module redundancy technology. Both CPU modules use the same application program, and one CPU module periodically communicates with the other. If no response is received within a timeout period, the system automatically switches to operational mode. All equipment on the production line has on-site manual switches, allowing for easy switching from the control system to manual operation without affecting the entire production line. This facilitates equipment maintenance and enhances system reliability. 5. Conclusion The computer control system for compound fertilizer production at Jinzhou Special Fertilizer Plant was fully operational in 2000. Its stable and reliable operation has significantly improved production efficiency and elevated factory production and management to a new level. Given the large amount of data collected from the production line, in-depth data mining and analysis presents a significant new challenge.