1 Introduction
Currently, China has very strict requirements for the quality indicators of feedwater and steam in large boilers, thus requiring continuous monitoring of boiler water quality. pH measurement mostly uses traditional PID control algorithms. However, during the reaction process, the high gain near the neutralization point makes it difficult to adjust the parameters of traditional PID controllers. Therefore, only a very small proportional gain can be used; otherwise, the system becomes unstable. Conversely, too small a proportional gain will worsen the system's dynamic characteristics. For boiler feedwater chemical dosing control devices, the dosing system has been automated , but there is no automatic dosing equipment. Dosing still requires manual dosing based on laboratory test results, which is not only labor-intensive but also poses a serious hazard to operators and causes environmental pollution due to the highly toxic and volatile substances such as ammonia and hydrazine. Therefore, two novel pH control methods are proposed: variable gain three-segment nonlinear PID and integral fuzzy control (IFC) algorithms. Digital simulation of the pH neutralization process with time delay shows that both control algorithms have strong robustness, fast response speed, and high control accuracy. In particular, the IFC algorithm can overcome the large time delay in the pH neutralization process. Through practical application in a power plant, the fully automated control of the boiler feedwater dosing and dosing system has been achieved.
2. Research on pH control methods
2.1 Conventional PID Control
PID control is a control method that uses a linear combination of the proportional (P), integral (I), and derivative (D) actions of the deviation. Figure 1 shows a conventional PID control system. Here, r is the reference input signal; PID is the controller; P is the controlled object model; d is the disturbance; e(k) is the system error; u(k) is the control quantity; and pH(k) is the output quantity of the controlled process. As can be seen from the figure, the proportional action in conventional PID control is actually a linear amplification or reduction action, which is difficult to adapt to the nonlinear characteristics of the controlled object in acid-base neutralization processes.
2.2 Variable Gain Three-Segment Nonlinear PID Control
The pH change is divided into two regions based on the inflection point: a high-gain region and two low-gain regions with different gain coefficients. The controller in the high-gain region uses a lower gain, while the controllers in the low-gain regions use different high gains to meet the system's desired performance indicators. Furthermore, to prevent integral saturation, a PID control algorithm with dead zone and output limiting is employed.
2.3 Fuzzy Control
The fuzzy control algorithm can be summarized as follows: Calculate the input variables based on the system output values obtained from the current sampling; convert the precise values of the input variables into fuzzy values; calculate the control quantity (fuzzy value) according to the fuzzy inference synthesis rules based on the input variables (fuzzy values) and fuzzy control rules; and calculate the precise control quantity from the control quantity (fuzzy value) obtained above.
3 Power Plant Boiler Feedwater Dosing Control System
A power plant has four 300 MW generator units, divided into two units: Unit 1 consists of Units 1 and 2, and Unit 2 consists of Units 3 and 4. Each unit's dosing and metering pumps handle both boiler feedwater (raw water purified through various water treatment methods to compensate for steam and water losses in the thermal power plant) and boiler water. Taking Unit 2 as an example, the dosing system employs three dosing and metering pumps: one for Units 3 and 4, switching to the standby pump when one fails. The system controls the amount of phosphate added to the boiler water by monitoring the pH value, which is required to be between 9.14 and 9.78. When the pH value of one unit falls below 9.4, the corresponding unit's dosing pump is activated. At this time, the phosphate solution in the phosphate dosing tank is pumped through pipelines (all valves on the pipelines are manual and normally open) to the dosing point on the deaerator outlet of the corresponding unit. If the dosing pump of Unit 3 malfunctions, the valve on the pipeline connected to the standby pump will be opened, and the standby pump will take over from the dosing pump of Unit 3 to dosing chemicals to the boiler water of Unit 3; the same applies to Unit 4. The addition of appropriate amounts of phosphate and sodium hydroxide to the boiler water improves its buffering capacity and helps maintain the stability of its pH value, thereby preventing scaling and corrosion of the boiler water-cooled walls.
The system introduces boiler water samples through a desuperheating and depressurization device into a phosphate meter and pH meter probe for measurement. After analog-to-digital conversion, the data is processed by the control system's PID controller to control the inverter output, which in turn controls the dosing pump speed. This allows for real-time control of the boiler water's chemical dosage, maintaining the phosphate concentration and pH within acceptable ranges. Figure 2 shows its control flowchart. The control system consists of four parts: a controller, an actuator, the controlled object, and a transmitter. The controller comprises an S7-200 PLC and corresponding control software; the actuator consists of a frequency converter , a motor , and a metering pump; the controlled object is the boiler water; and the transmitter is an analytical instrument, specifically a pH meter.
3.1 Control Flow
Figure 3 shows the boiler water chemical dosing control system for Unit 3. This system extracts 4–20 mA signals from online analytical instruments (phosphate meter, pH meter), performs window-based PID composite calculations based on operating process parameters and a defined mathematical model, and sends the intermediate results to a frequency converter to control the chemical dosing pump dosage, thus achieving automatic closed-loop regulation of the chemical dosing.
3.2 Control System Composition
This control system uses a combination of WinCC (host computer software) and a Siemens PLC. The PLC system connects to the WinCC host computer via a ProfiBus bus, as shown in Figure 4. The host monitoring section is handled by an industrial computer (WinCC). Monitoring personnel can monitor the system's operation in real time via a CRT, set or modify system operating parameters, and remotely control the system via the CRT software. The host computer processes and manages data and connects to the MIS system. The host computer can configure the controller, including setting the controller's network address and time, selecting the control algorithm, setting algorithm parameters, setting the setpoint of the control quantity, and selecting the input and output channels in the algorithm. The lower-level control section is handled by a programmable logic controller (PLC) installed in the field. It is the core of the automatic dosing control system, used to collect relevant water quality data. Because the chemical dosing system has a pure time lag, the control action may be untimely, causing overshoot or oscillation in the system. A computer can easily compensate for this time lag. An improved digital PID control algorithm and fuzzy control algorithm are adopted, enabling the controller to adjust the AC frequency converter in the field using the output control signal, thereby controlling the motor speed and regulating the dosing pump. The electrical control system is designed with both remote and local control modes to achieve manual/semi-automatic/automatic functions, with the latter two functions switched by the host computer.
4. Application of IFC algorithm in filtering
In the control system, the basic principle of the filtering program is to continuously sample five values within a period, calculate their average value, collect the current value, and calculate the difference between the collected value and the average value, Δ = Xi - X. If |Δ| > 0.2, Xi is discarded, and X = 0.2 is taken as the signal fluctuation range value set according to the actual situation. If |Δ| ≤ 0.2, X1 is popped from the stack, X2 replaces X1, X3 replaces X2, X4 replaces X3, and so on. X5 is replaced with the currently sampled X6, and then X is calculated again using these five new values and compared. By repeatedly executing this program, the filtering function can be achieved. Figure 5 shows the amplified pH value trend after using the filtering program, which shows that the filtering effect is good. Figure 6 shows the control operation interface.
5 Conclusion
Practice has proven that the PLC-based automatic chemical dosing control system can flexibly meet the online monitoring needs of various chemical dosing systems. Since its commissioning, the system has operated stably and reliably, enabling full automatic adjustment of boilers and auxiliary equipment, achieving the expected results. It has solved the problem of difficulty in ensuring stable water quality indicators under manual control, reduced the workload of operators, and received unanimous praise.
