Self-adaptive composite high-precision pressure control algorithm

Abstract

The invention discloses a self-adaptive composite high-precision pressure control algorithm and an implementation system, and belongs to the technical field of pressure control. According to the algorithm, a composite framework of adaptive segmentation PID + feedforward-feedback composite compensation + fuzzy logic dynamic correction is adopted, PID parameters are adaptively adjusted through range segmentation and deviation, and dynamic response and steady-state precision are balanced; the gas source pressure and temperature interference are pre-judged through feedforward compensation, and overshoot is suppressed through Kalman filtering noise reduction and integral separation in feedback compensation; pID parameters are corrected in real time through fuzzy logic, and load sudden change is adapted. The system for realizing the algorithm comprises an upper computer, an STM32H743 main control unit, a high-precision sensor, an actuating mechanism driving module and a proportional valve. The steady-state error is less than or equal to + / -0.001 MPa in the range of 0-1MPa, the step response time is less than or equal to 3ms, the anti-interference capability is high, parameters do not need to be manually adjusted, and the method is suitable for high-precision pressure control scenes such as semiconductor manufacturing, precise instruments, aerospace and the like.

Description

Technical Field

[0001] This invention relates to the field of pressure control technology, and in particular to a high-precision pressure control algorithm and its supporting implementation system that integrates adaptive piecewise PID, feedforward-feedback compensation and fuzzy logic correction. It is suitable for scenarios with stringent requirements for pressure control accuracy, dynamic response speed and anti-interference capability, such as semiconductor manufacturing, precision instruments, aerospace testing, medical equipment and other fields. Background Technology

[0002] Pressure control technology is one of the core supporting technologies in industrial production, precision instruments, aerospace and other fields. Its control accuracy and stability directly affect product quality or equipment operational reliability. With the development of technology, more and more scenarios require nanometer-level pressure regulation accuracy and millisecond-level dynamic response. For example, the pressure control accuracy of process gas in semiconductor wafer etching process needs to reach ±0.001MPa, and the pressure response time of aerospace hydraulic system testing equipment needs to be ≤5ms.

[0003] Existing pressure control technologies mainly rely on traditional PID control algorithms, but they have the following significant drawbacks: Insufficient steady-state accuracy: Traditional PID algorithms use fixed parameters, which are prone to overshoot and oscillation in high-precision scenarios with small ranges, making it difficult to balance dynamic response and steady-state accuracy; Weak anti-interference ability: External interference such as ambient temperature drift and gas source pressure fluctuation will cause static error accumulation, and the accuracy will be significantly reduced after long-term operation; Poor adaptive ability: When the load of the controlled object (such as the volume of the sealed cavity or the resistance of the pipeline) changes suddenly, the PID algorithm with fixed parameters cannot adjust in time, resulting in deterioration of control performance; Noise sensitivity: The signals collected by the pressure sensor are susceptible to electromagnetic interference and mechanical vibration, which directly affects the accuracy of deviation calculation and further reduces control precision.

[0004] To address the aforementioned issues, some technical solutions have attempted to introduce fuzzy PID or feedforward control, but these still have shortcomings: using only fuzzy PID without considering the predictive compensation for external disturbances results in limited anti-interference capability; using only feedforward control lacks dynamic correction through closed-loop feedback, making it difficult to cope with parameter drift under complex operating conditions. Therefore, there is an urgent need for a pressure control scheme that integrates the advantages of multiple algorithms and possesses high precision, fast response, strong anti-interference capabilities, and adaptability. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing pressure control technologies and provide an adaptive composite high-precision pressure control algorithm and implementation system, which achieves nanometer-level pressure regulation accuracy and millisecond-level dynamic response, while improving the system's anti-interference and adaptive capabilities to meet the stringent requirements of precision manufacturing, semiconductor equipment and other scenarios.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An adaptive composite high-precision pressure control algorithm is characterized by employing a composite architecture of "main control + auxiliary compensation + dynamic correction," including an adaptive segmented PID main control module, a feedforward-feedback composite compensation module, and a fuzzy logic dynamic correction module. The specific implementation is as follows: Adaptive segmented PID main control module: Range segmentation strategy: Based on the range of the target pressure Pt, the pressure is divided into three range segments: low, medium, and high. The division of each range segment can be adjusted according to the actual application scenario. An example division is as follows: Low range: 0-0.2MPa; Medium range: 0.2-0.8MPa; High-range measurement: 0.8-1MPa; Each measurement range is pre-stored with initial PID parameters (Kp, Ki, Kd) obtained through experimental optimization. The initial parameters meet the core requirements of the corresponding measurement range: low measurement range prioritizes steady-state accuracy, medium measurement range balances response speed and stability, and high measurement range prioritizes dynamic response. Adaptive parameter adjustment logic: Real-time calculation of pressure deviation e = target pressure Pt - actual pressure Pm, and deviation change rate ec = de / dt; dynamically switching PID parameter combinations based on the absolute value of deviation e. When |e|>5% Pt, use the parameter combination of "large Kp, small Ki, medium Kd" to prioritize improving dynamic response speed and quickly reduce deviation; When 1% Pt ≤ |e| ≤ 5% Pt, a parameter combination of "medium Kp, medium Ki, and small Kd" is used to balance response speed and steady-state stability and avoid overshoot. When |e| < 1% Pt, a parameter combination of "small Kp, large Ki, small Kd" is used to prioritize suppressing static errors and ensure steady-state accuracy; parameter self-tuning mechanism: based on the Ziegler-Nichols critical proportionality method, two self-tuning trigger conditions are set: Timed trigger: Parameter optimization is automatically performed every 1 hour to update the PID parameters for each range segment; Emergency Trigger: When the pressure deviation |e| of 5 consecutive samples exceeds the allowable error threshold (e.g., ±0.001MPa), normal control is immediately suspended, parameter optimization is performed, and control is resumed after completion.

[0007] Feedforward-feedback composite compensation module Feedforward control: To address the two main external disturbances—gas source pressure fluctuations and ambient temperature drift—a predictive compensation model is established to output compensation amounts in advance to counteract the impact of the disturbances. The compensation formula is as follows: U_feedforward = K1 × ( dP_source / dt ) + K2 × (T _current - T_reference ) in: U_feedforward is the feedforward compensation control variable; K1 is the gas source pressure compensation coefficient, calibrated experimentally (example value: 0.02~0.05); dP_source / dt is the gas source pressure change rate (MPa / s). K2 is the temperature compensation coefficient, calibrated experimentally (example values: 0.003~0.008); T_current is the current ambient temperature (°C). T_reference is the standard reference temperature (25℃).

[0008] The Kalman filter algorithm is used to reduce the noise of the actual pressure signal Pm acquired by the pressure sensor, eliminating sensor noise and electrical interference. Magnetic interference, the filtering equation is as follows: State prediction equation: X(k|k-1) = A×X(k-1|k-1) + B×U(k-1) State update equation: X (k|k) = X (k|k-1) + K (k)×[Z (k) - H×X (k|k-1)] where X is the state vector, A is the state transition matrix, B is the control input matrix, U is the control quantity, Z is the observation value, H is the observation matrix, and K (k) is the Kalman gain; An integral separation mechanism is introduced: when |e|>3% Pt, the integral action is paused (Ki=0) to avoid overshoot caused by integral saturation; when |e|≤3% Pt, the integral action is resumed to eliminate static error through integration.

[0009] Fuzzy logic dynamic correction module: Input fuzzification: Pressure deviation e and deviation change rate ec are used as fuzzy inputs. The fuzzy subsets of each input are defined as: {negative large (NL), negative medium (NM), negative small (NS), zero (ZO), positive small (PS), positive medium (PM), positive large (PL)}, and the precise values ​​are converted into fuzzy values ​​through the triangular membership function. Output fuzzification: The PID parameter correction values ​​ΔKp, ΔKi, and ΔKd are used as fuzzy output values, and the fuzzy subsets of each output value are defined as follows: {Negative Large (NL), Negative Medium (NM), Negative Small (NS), Zero (ZO), Positive Small (PS), Positive Medium (PM), Positive Large (PL)}; Fuzzy rule base: 49 core fuzzy rules are established (7×7 input combinations), with example rules as follows: If e=PL and ec=NL, then ΔKp=NM, ΔKi=ZO, and ΔKd=PS; If e=PM and ec=NS, then Δ Kp=NS, Δ Ki=ZO, Δ Kd=ZO; If e=ZO and ec=ZO, then Δ Kp=ZO, Δ Ki=PS, Δ Kd=ZO; If e=NS and ec=PM, then ΔKp=PS, ΔKi=ZO, and ΔKd=NS; If e=NL and ec=PL, then ΔKp=PM, ΔKi=ZO, and ΔKd=NM; Defuzzification: The centroid method is used to convert the fuzzy output into precise parameter correction values ​​ΔKp, ΔKi, and ΔKd, and the PID parameters of the current range are corrected in real time to adapt to sudden load changes.

[0010] Algorithm workflow: Initialization Phase: After the system is powered on, the host computer sends the target pressure Pt, the main control unit reads Pt and determines its range, and loads the corresponding quantity. The initial PID parameters for the process segment are set, the initial temperature compensation coefficient K2 and the air source pressure compensation coefficient K1 are initialized, and the allowable error threshold (such as ±0.001MPa) is set. Signal acquisition and preprocessing: The actual pressure Pm of the controlled object is acquired through a high-precision pressure sensor, the ambient temperature T_current is acquired through a temperature sensor, and the gas source pressure P_source is acquired through a gas source pressure sensor; Kalman filtering is performed on Pm to reduce noise, and the pressure deviation e and the deviation change rate ec are calculated. Feedforward compensation calculation: Calculate the gas source pressure change rate dP_source / dt based on the collected P_source, and combine it with the difference between T_current and T_reference, and substitute it into the feedforward compensation formula to calculate the feedforward control quantity U_feedforward; Fuzzy logic correction: Fuzzify e and ec, query the fuzzy rule base, and obtain the PID parameter correction values ​​ΔKp, ΔKi, and ΔKd by defuzzification, and update the current PID parameters; Adaptive PID calculation: Based on the absolute value of e, select the corresponding PID parameter combination and calculate the PID control quantity U_pid; Total control output: Total control output U_total = U_pid + U_feedforward. The digital signal is converted into an analog signal by a D / A converter and output to the actuator (proportional valve / servo valve) to regulate the air source output pressure. Closed-loop feedback and iteration: Real-time monitoring of the deviation between the actual pressure Pm and the target pressure Pt. If |e| ≤ the allowable error threshold, maintain the current control parameters and continue monitoring; if |e| > the allowable error threshold, return to step 2 and re-execute the signal acquisition and control process. Parameter self-tuning trigger: If the timed trigger or emergency trigger conditions are met, the normal control process is paused, Ziegler-Nichols parameter optimization is performed, the initial PID parameters for each range segment are updated, and then the process returns to step 2 to continue closed-loop control. Attached Figure Description

[0011] Figure 1 This is a block diagram of the system hardware components.

[0012] Figure 2 This is a flowchart of the algorithm's workflow.

Claims

1. The algorithm is characterized in that, The algorithm adopts a composite architecture of "main control + auxiliary compensation + dynamic correction", including an adaptive segmented PID main control module, a feedforward-feedback composite compensation module and a fuzzy logic dynamic correction module. The algorithm's workflow includes initialization, signal acquisition and preprocessing, feedforward compensation calculation, fuzzy logic correction, adaptive PID calculation, total control output, closed-loop feedback and iteration, and parameter self-tuning steps.

2. The adaptive composite high-precision pressure control algorithm according to claim 1, characterized in that, The adaptive piecewise PID main control module is implemented as follows: (1) Divide the target pressure Pt into three ranges: low, medium, and high. Pre-store the optimized initial PID parameters for each range. (2) Calculate the pressure deviation e=Pt-Pm and the deviation change rate ec=de / dt in real time, and dynamically switch the PID parameter combination according to the magnitude of |e|: When |e| > 5% Pt, use the parameter combination of "large Kp, small Ki, medium Kd"; When 1% Pt ≤ |e| ≤ 5% Pt, use the parameter combination of "medium Kp, medium Ki, small Kd"; When |e| < 1% Pt, use the parameter combination of "small Kp, large Ki, small Kd"; (3) Based on the Ziegler-Nichols method, set up two parameter self-tuning mechanisms: timed triggering and emergency triggering.

3. The adaptive composite high-precision pressure control algorithm according to claim 1, characterized in that, The feedforward-feedback composite compensation module is implemented as follows: (1) Feedforward control establishes a compensation model based on the gas source pressure change rate and the difference in ambient temperature. The compensation formula is: U_feedforward = K1×(dP_source / dt) + K2×(T_current - T_reference), where K1 is the gas source pressure compensation coefficient, K2 is the temperature compensation coefficient, and T_reference is 25℃. (2) Feedback compensation uses Kalman filtering algorithm to reduce the noise of pressure signal and introduces integral separation mechanism. When |e| > 3% Pt, the integral action is paused and when |e| ≤ 3% Pt, the integral action is resumed.

4. The adaptive composite high-precision pressure control algorithm according to claim 1, characterized in that, The implementation method of the fuzzy logic dynamic correction module is as follows: (1) The fuzzy subsets of input quantities e and ec are both {NL, NM, NS, ZO, PS, PM, PL}, which are fuzzified by the triangular membership function; (2) The fuzzy subsets of the output quantities ΔKp, ΔKi, and ΔKd are all {NL, NM, NS, ZO, PS, PM, PL}; (3) Establish 49 core fuzzy rules, and obtain accurate correction values ​​by defuzzification using the centroid method, and update PID parameters in real time.

5. The adaptive composite high-precision pressure control algorithm according to claim 2, characterized in that, The range of measurement segments is defined as follows: low range 0-0.2MPa, medium range 0.2-0.8MPa, and high range 0.8-1MPa, or the threshold can be adjusted according to the actual application scenario.

6. The adaptive composite high-precision pressure control algorithm according to claim 2, characterized in that, The timed triggering period is 1 hour, and the emergency triggering condition is that the pressure deviation |e| of 5 consecutive samples exceeds the allowable error threshold.

7. A high-precision pressure control system implementing the algorithm of any one of claims 1-6, characterized in that, It includes a host computer, a main control unit, a signal acquisition module, an algorithm processing module, an actuator drive module, and an actuator. These modules are sequentially connected for communication. (1) The main control unit adopts STM32H743 high-performance MCU with an operation frequency ≥480MHz; (2) The signal acquisition module includes a high-precision pressure sensor (accuracy ≤ ±0.0005MPa), a temperature sensor (accuracy ≤ ±0.1℃) and a gas source pressure sensor (accuracy ≤ ±0.001MPa). (3) The actuator drive module includes an 18-bit resolution D / A converter and a power amplifier circuit; (4) The actuator is a proportional valve or servo valve with a response time of ≤5ms.

8. The high-precision pressure control system according to claim 7, characterized in that, The high-precision pressure sensor is a Honeywell ASDX series, the temperature sensor is a PT1000 platinum resistance thermometer, the gas source pressure sensor is a Sensirion SDP810, the D / A converter is an AD5791, and the actuator is a Bosch Rexroth 4WRA6 proportional valve.

9. The high-precision pressure control system according to claim 7, characterized in that, The host computer communicates with the main control unit via RS485 bus or Ethernet to send out target pressure Pt and display actual pressure Pm, control parameters, deviations and other data.

10. The adaptive composite high-precision pressure control algorithm according to any one of claims 1-6, characterized in that, The algorithm has a steady-state pressure error of ≤ ±0.001MPa, a step response time of ≤3ms, and a pressure fluctuation of ≤ ±0.0008MPa under operating conditions with an ambient temperature of -10℃ to 60℃ and a gas source pressure fluctuation of ±5%.

Images