Abstract: This paper introduces the chemical dosing modification scheme, automatic loop, and PID tuning process of the Daya Bay Nuclear Power Plant, and provides some experience and PID parameters. Keywords: Nuclear power plant; Automatic chemical dosing; PID parameters 0 Introduction Similar to conventional power plants, the chemical dosing system of the Daya Bay Nuclear Power Plant adjusts the dissolved oxygen content in the secondary loop water by adding hydrazine and the pH value of the secondary loop water by adding ammonia. The purpose is to control the chemical properties of the steam-water system, prevent corrosion of thermal equipment and pipelines such as steam generators and turbines, and reduce the precipitation of impurities. In addition, this system also has the function of wet maintenance and corrosion prevention for thermal equipment. The added hydrazine and ammonia must be controlled within a reasonable range to meet the corrosion prevention requirements; too much or too little will affect the corrosion prevention performance. According to the design requirements of the Daya Bay Nuclear Power Plant, the conductivity of condensate should be controlled within the range of 12-17 S/cm, corresponding to a pH value of 9.64-9.79, and the amount of hydrazine in the condensate should be controlled between 70-90 ppb. The original system relied on chemists to adjust the stroke of the corresponding dosing pumps based on on-site inspections of conductivity and hydrazine concentration, thereby altering the injection flow rates of ammonia and hydrazine to control the pH value and hydrazine dosage in the condensate. Due to equipment aging and process requirements, it was decided to automate the secondary loop ammonia and hydrazine dosing system. 1. Modification Scheme We selected a new type of hydraulic diaphragm metering pump and asynchronous AC motor to replace the original three continuously operating dosing pumps. Of the three pumps, one is on standby, while the other two use signals from the existing conductivity and hydrazine meters, respectively sent to two PID controllers. The PID controllers perform PID calculations based on the measured values ​​and set target values, outputting adjustment commands to the frequency converter. The frequency converter outputs AC power of the corresponding frequency to drive the asynchronous AC motor, changing the motor speed and consequently altering the injection flow rate of the chemical solution, thus achieving the goal of controlling the conductivity and hydrazine dosage in the condensate. 1.1 Control Modes Due to the high requirements for reliability and availability of nuclear power plants, the modified system has three adjustment modes: manual, semi-automatic, and fully automatic, which are switched by three position changeover switches. Specifically: (1) Manual: In this mode, similar to the original traditional dosing method, the frequency converter and PID controller are bypassed. The start and stop of the dosing pump are controlled by the on-site operator. When the start button is pressed, the pump motor runs at full speed (50Hz). The injection volume of the chemical solution is controlled by adjusting the stroke of the corresponding dosing pump according to the on-site inspection parameters. (2) Semi-automatic: The PID controller is bypassed. The control of the dosing volume is achieved by adjusting the speed control potentiometer of the frequency converter. Changing the motor speed changes the corresponding chemical solution injection flow rate. (3) Automatic: The injection control of the chemical solution is completed automatically. The dosing pump is automatically started and stopped according to the set threshold. The PID controller calculates the value based on the user setting and the detection value of the on-site instrument and then controls the frequency converter to adjust the motor speed and start and stop. The corresponding chemical solution flow rate is controlled within the range required by the process. [b]1.2 Main Equipment Selection 1.2.1 Regulator[/b] Function: Completes PI control and sets control targets; sets control points for automatic pump start and stop. Model: UDC33OO Regulator (Honeywell) Details: Input/output isolation. Input Type: Various thermocouples, RTDs, linear signals; dual-loop input, digital signal input (optional); based on the characteristics of the input signal conductivity and hydrazine concentration in this project, linear signals are selected as the input type. Output Type: Current output, relay output, open-collector output, optional auxiliary current output; isolated auxiliary current output and digital input; SPDT electromagnetic relay alarm output, various alarm types can be flexibly configured; RS-422/485 MODBUS and ASCII code, DMCS communication; based on the inverter input signal and regulator output signal being a 4-20mA current signal, linear signals are selected as the input type. A/M manual/automatic switching control; features fuzzy logic control and Accutune 11 parameter adaptive algorithm with overshoot suppression. Control algorithm outputs: ON/OFF, position proportional, PID-A, PID-B, PD control with manual integral, dual-loop PID control, manual/automatic cascade control, three-position stepper control, dual output control (this project uses PID-A control algorithm output). Control accuracy: This project uses 11-bit resolution and 0.5% output accuracy across the entire range. 1.2.2 Frequency Converter Function: Changes the speed of the dosing pump according to the regulator's instructions to change the dosing rate; automatically starts/stops the dosing pump according to the regulator's or manual instructions. Model: ACS140 (ABB) Control Accuracy: Sensitively tracks given signals, average accuracy greater than 1%; Response: Fast response speed, average delay less than 9ms; Power Supply Voltage: Three-phase 380V (±10%); Frequency: 50Hz; Protection Functions: Overcurrent, overvoltage, undervoltage, overheating, I/O terminal short circuit, grounding, output short circuit, input phase loss, motor overload, and stall protection. 1.2.3 Metering Pump Function: Completes the dosing target based on the inverter's speed control and start/stop control. Model: RB090 (MROY series hydraulic diaphragm metering pump, Milton Roy, USA) Performance Characteristics: Hydraulically driven diaphragm without plunger packing, maintenance-free high-precision inlet and outlet one-way check valves, built-in adjustable pressure relief valve, dual diaphragm/with rupture alarm function, optional PTFE material diaphragm, life not less than 20,000h. Single/double head structure. The flow rate can be adjusted independently, manually, electrically, by frequency conversion, or pneumatically. Flow rate can be adjusted while the pump is running or stopped. Adjustment range: 100:10, steady-state accuracy: ±1%. Maximum suction lift: 3m water column. Medium temperature: plastic up to 60℃, metal up to 90℃. Multiple pump head materials are available (PVC, PVDF, 316SS alloy, etc.). This project uses a 316SS alloy pump head due to the corrosion requirements of the chemical reagent medium. 2. Automatic Adjustment Circuit The principle of the automatic adjustment control circuit is shown in Figure 1. Each of the two adjustment circuits uses the high and low alarm nodes of the regulator to control the pump's start and stop. When the measured value is lower than the set minimum point, the dosing pump starts and holds; when the measured value is higher than the maximum set value, the dosing pump stops. The threshold control provides both automatic start/stop functionality and high/low threshold protection to prevent excessive system overshoot. Since the standby pump is used for both ammonia and hydrazine dosing, a switch for the standby pump is provided. When the switch is in the "standby" position... The ammonia addition regulator and the hydrazine addition regulator control their respective dosing pumps; when the switch is in the "ammonia addition" position, the ammonia addition regulator controls the standby pump to add ammonia, and the hydrazine addition regulator controls the hydrazine dosing pump; when the switch is in the "hydrazine addition" position, the hydrazine addition regulator controls the standby pump to add hydrazine, and the ammonia addition regulator controls the ammonia dosing pump. [b]3 Setting and Adjusting Parameters 3.1 Setting Control Thresholds[/b] According to the design requirements of the Daya Bay Nuclear Power Plant, the conductivity of condensate should be controlled between 12 and 17 μs/cm, and the amount of hydrazine in the condensate should be controlled between 70 and 90 ppb. The conductivity reacts quickly to ammonia addition, but considering that there is still a certain lag, the conductivity start point is set at 12.5 μs/cm, and the stop point is set at 16.5 μs/cm. Through experiments, this setting can ensure that the conductivity is controlled between 12 and 17 μs/cm. Due to the distance between the dosing point and the measurement point, the hydrazine measurement response is slow, resulting in a significant lag in the hydrazine measurement signal at the Daya Bay Nuclear Power Plant. It takes approximately 12 minutes for the hydrazine signal to decrease after dosing stops. Once the signal begins to decrease, even with immediate dosing at the fastest possible speed, the decrease is rapid, easily causing the hydrazine value to exceed the acceptable range. To avoid excessively low hydrazine values ​​during pump switching, it is essential to ensure the pump is operational immediately after switching. Therefore, the minimum starting point must be set higher than the normal set value. The normal target value for hydrazine at the Daya Bay Nuclear Power Plant is 80 PPb, and we set the minimum starting point to 82 PPb. Through this method, experiments showed that during switching, due to the time required for valve replacement and other operations, the minimum hydrazine value was 75 PPb, which meets the requirements. The maximum value serves only a protective function and is set to 90 ppb. [b]3.2 PID Parameter Tuning 3.2.1 Ammonia Dosing Regulation[/b] In the experiment, the critical proportional gain was initially used. This method assumes the boundary stability of the regulating system. The critical proportional gain is first determined using a simple proportional gain test, and then the regulating parameters are tuned. The method is as follows: ① Set the integral time to maximum (or off), set the derivative time to minimum (or off), set the proportional band to a large value, and put it into closed-loop operation; ② If stable, reduce the proportional band until oscillation begins; ③ If it is a decaying oscillation, reduce the proportional band; ④ If it is an increasing oscillation, increase the proportional band to adjust the system to a small, constant amplitude oscillation. The proportional band at this time is the critical proportional band (B); record the curve of the adjusted parameter fluctuating with time and calculate its fluctuation period (critical period T). B=100, T=5min. PI control is used here. From the empirical formula, we get: P=2.2 × B=220; Ti=0.8 × T=4min; After obtaining the preliminary PI value, a trial-and-error method is used to obtain a more ideal parameter set. P and T are changed respectively, within the range of ±10% to ±20%, and 25 sets of PI are formed by 5 Pi and Ti. This is repeated to find a satisfactory PI value. Experiments show that when P is 64 and Ti is 4, The conductivity fluctuation tends to stabilize within one cycle, showing good performance. Curve recording over 12 hours shows that after using this set of adjustment parameters, the conductivity fluctuation is within ±0.1 μs/cm. 3.2.2 Hydrazine Adjustment Due to the existence of significant inertia or hysteresis in the system, the change in the error suppression function always lags behind the change in the error. The solution is to make the change in the error suppression function "lead," that is, when the error approaches zero, the error suppression function should be zero. In other words, simply introducing a "proportional" term in the controller is often insufficient; the proportional term only amplifies the magnitude of the error. Adding a "derivative" term, however, can predict the trend of error change, thus enabling the error suppression control function to be equal to zero, or even negative, in advance, thereby avoiding severe overshoot of the controlled variable. Therefore, for controlled objects with significant inertia or hysteresis, the PD controller can improve the dynamic characteristics of the system during the adjustment process. Considering the time constraints of debugging, a trial-and-error method was not feasible, so we adopted the regulator's self-tuning function. After a 4-hour self-tuning process, relatively ideal parameters were obtained: P=1.663, I=4.53, D=18.29. This also confirms the previous analysis that required the addition of differentials. Curve recording over 12 hours showed that with this set of adjustment parameters, the hydrazine fluctuation was within ±1 ppb. 4. Conclusion Although automatic dosing technology is already used in many thermal power plants, its application in nuclear power plants still requires certain characteristics. First, the reliability requirements for design and equipment selection are high. Therefore, based on research and experience, a manual function was retained to prevent dosing failure in case of control system malfunction. Second, a complete test process needs to be designed and strictly implemented (this is also required by the Daya Bay Nuclear Power Plant). The main test contents include: pipeline pressure resistance test; pump output test, pump output pressure test, and pump temperature rise test; Motor cabinet insulation test, motor temperature rise test, motor impedance test, motor current test, motor speed test; control cabinet insulation test, control cabinet no-load test, control cabinet load test, control cabinet detailed functional test. All tests and assessments form a complete functional and quality assessment report. Finally, for the hydrazine control system, due to its severe hysteresis, the adjustment of PID parameters is difficult. We hope that our analysis and results will be of some help to our peers. References [1] Chen Yujuan, Liu Dongbo. Stress analysis and calculation of instrument tubes in nuclear power plants. Chongqing: Electrical Engineering Technology, 2004 (6): 52-53.