A constant temperature and humidity air conditioner PID self-adaptive intelligent control method and system

Abstract

The application relates to the field of air conditioners, and discloses a constant-temperature and constant-humidity air conditioner PID self-adaptive intelligent control method, which comprises the following steps: S1. collecting environment parameter data and equipment operation data in real time, S2. pre-processing the collected data; S3. based on the state vector, synchronously executing a bidirectional correction mechanism; S4. through a collaborative strategy engine; S5. an actuator acting according to the control instruction to adjust the air supply temperature and humidity (or return air, indoor temperature and humidity); and S6. collecting system response data after adjustment. Through bidirectional self-adaptive correction of thresholds and parameters, collaborative decision-making and closed-loop feedback mechanisms, the application can significantly improve the control precision, response speed and operation stability of the constant-temperature and constant-humidity system, effectively suppresses overshoot oscillation, reduces energy consumption, and prolongs the service life of equipment.

Description

Technical Field

[0001] This invention relates to the field of air conditioning, and more particularly to a PID adaptive intelligent control method and system for constant temperature and humidity air conditioning. Background Technology

[0002] Constant temperature and humidity air conditioning systems are widely used in places with extremely high requirements for the precision of environmental parameter control, such as data centers, clean rooms, museums, laboratories, and pharmaceutical production workshops. Traditional control strategies mostly adopt a static threshold comparison mechanism based on the return air enthalpy, temperature, and humidity values ​​and the set enthalpy, temperature, and humidity values. That is, when the measured enthalpy, temperature, and humidity values ​​deviate from the set values ​​by more than the preset threshold, temperature and humidity adjustment actions are triggered (such as starting the compressor, adjusting the chilled water valve, adjusting the heating valve, adjusting the humidification valve, etc.).

[0003] However, this type of control method has significant technical defects: because it relies on a fixed threshold, the system is slow to react to sudden changes in environmental parameters (such as door opening, equipment start-up and shutdown, sudden increase in personnel, and changes in humidity), resulting in the system only starting to adjust after the temperature and humidity have deviated significantly; and it is prone to frequent start-up and shutdown under sensor data fluctuations or small disturbances, causing system oscillations, aggravating equipment wear and energy waste. Secondly, a single threshold cannot distinguish between different physical conditions such as "high temperature and low humidity" and "low temperature and high humidity", nor does it take into account the nonlinear response characteristics and aging attenuation of actuators (such as cold coils, reheaters, and humidifiers), resulting in a mismatch between control actions and actual needs and poor control accuracy.

[0004] In recent years, although some studies have attempted to introduce fuzzy control, neural networks or model predictive control to improve control accuracy, most of them only optimize the control algorithm without simultaneously correcting the "trigger threshold" and "execution parameters". The control logic is fragmented and no collaborative linkage mechanism between the threshold boundary and the control parameters has been established.

[0005] Therefore, there is an urgent need for an intelligent control system for constant temperature and humidity air conditioning that can synchronously and dynamically correct control thresholds and system parameters, achieve bidirectional collaborative optimization, support edge deployment and closed-loop self-evolution, so as to break through the bottlenecks of traditional solutions in terms of accuracy, stability and adaptability, and meet the needs of highly sensitive environments. Summary of the Invention

[0006] The purpose of this invention is to provide a PID adaptive intelligent control method and system for constant temperature and humidity air conditioning to solve the above problems. The specific technical solution is as follows:

[0007] A PID adaptive intelligent control method for constant temperature and humidity air conditioning includes the following steps:

[0008] S1. Real-time collection of environmental parameter data and equipment operation data, wherein the environmental parameters include return air temperature and humidity, fresh air temperature and humidity, and indoor disturbance factor, and the equipment operation data includes actuator opening degree, historical control deviation, and energy consumption data;

[0009] S2. Preprocess the collected data, including outlier removal, sliding filtering, trend feature extraction, and construct the current state vector of the system;

[0010] S3. Based on the aforementioned state vector, a bidirectional correction mechanism is executed synchronously:

[0011] A1: Threshold Correction Module: Based on the current environmental disturbance intensity, historical deviation trend, and equipment response delay, dynamically adjust the enthalpy or temperature and humidity values ​​to control the threshold band width and trigger boundary;

[0012] B1: Parameter correction module: Based on changes in actuator characteristics, temperature and humidity coupling relationship, and control performance feedback, PID control parameters or decoupling matrix coefficients are tuned in real time;

[0013] S4. Through the collaborative strategy engine, the corrected threshold boundary and control parameter combination are matched to the current system operating state to generate the final control command;

[0014] S5. The actuator operates according to the control command to adjust the supply air temperature and humidity (or return air, indoor temperature and humidity);

[0015] S6. Collect the system response data after adjustment, evaluate the control performance indicators, and feed them back to the threshold correction module and parameter correction module to achieve closed-loop self-optimization.

[0016] As an improvement to the above technical solution, in step S3, the threshold correction module includes: matching a preset threshold range based on working condition clustering, wherein the working condition consists of temperature and humidity combination, time period label, and load level; dynamically drifting the threshold center point according to the prediction results of the enthalpy value and temperature and humidity value change trends; setting an adaptive dead zone mechanism and a hysteresis direction compensation mechanism; the dead zone width is positively correlated with the disturbance intensity and equipment delay; and adjusting the threshold trigger direction according to the historical overshoot direction.

[0017] As an improvement to the above technical solution, in step S3, the parameter correction module includes: online identification of the actuator transfer function, updating the "valve opening matching temperature / humidity change" response model, dynamically adjusting the temperature and humidity decoupling control matrix, updating the influence coefficient every preset period based on the least squares method, automatically tuning the PID parameters based on overshoot, settling time, and integral error by using fuzzy rules or reinforcement learning algorithms, establishing an equipment aging compensation factor, and adjusting the gain coefficient as the running time decays.

[0018] As an improvement to the above technical solution, in step S4, the collaborative strategy engine includes a state machine-driven mechanism that automatically switches collaborative modes according to the current system state, including:

[0019] Stable operating mode: Thresholds are relaxed, parameters are conservative;

[0020] Fast response mode: threshold tightening, high parameter gain, and feedforward;

[0021] Energy-saving optimization mode: threshold expansion, parameter smoothing, and extended integral;

[0022] Device degradation mode: Threshold compensation delay, parameter reduction response strength;

[0023] Abnormal disturbance mode: threshold pre-tightening, parameters enable robust control set.

[0024] As an improvement to the above technical solution, in step S5, the control performance indicators include: integral absolute value of temperature and humidity control deviation (ITAE), overshoot percentage, system stable response time, frequency of adjustment actions per unit time, and energy efficiency ratio. The feedback mechanism adjusts the threshold correction coefficient and parameter correction weight synchronously according to preset rules based on the degree of performance indicator degradation.

[0025] As an improvement to the above technical solution, in step S6, a feature library of historical high-quality control segments is constructed for matching similar working conditions. An online reinforcement learning algorithm is adopted to optimize the threshold boundary and control parameters by minimizing control error, energy consumption and action frequency as the reward function.

[0026] A constant temperature and humidity air conditioning system, applied to the aforementioned constant temperature and humidity air conditioning PID adaptive intelligent control method, includes a data acquisition module, a data preprocessing module, a threshold correction module, a parameter correction module, a collaborative strategy engine, an execution control module, a performance evaluation and feedback module, and a storage module. Specifically: the data acquisition module acquires real-time environmental and equipment data; the data preprocessing module performs filtering, alignment, and feature extraction; the threshold correction module dynamically adjusts enthalpy, temperature, and humidity values ​​to control trigger boundaries; the parameter correction module tunes the controller's internal parameters in real time; the collaborative strategy engine matches the current operating conditions and combines correction strategies; the execution control module outputs control commands to the actuator; the performance evaluation and feedback module quantifies the control effect and sends back optimization signals; and the storage module stores historical data, operating condition templates, and a strategy rule base.

[0027] As an improvement to the above technical solution, the system is deployed in an edge controller or building automation host and connected to sensors and actuators.

[0028] As an improvement to the above technical solution, a data bus is also included. The data acquisition module, data preprocessing module, threshold correction module, parameter correction module, collaborative strategy engine, execution control module, and performance evaluation and feedback module are all connected to the data bus. Each module shares the current system state vector through a publish / subscribe mechanism. The state vector includes real-time enthalpy value, temperature and humidity value, deviation trend, equipment response delay, disturbance level, and operating condition label. The collaborative strategy engine, as the control terminal of the data bus, schedules the control command stream according to a preset priority to ensure that the output of threshold correction and parameter correction takes effect synchronously without conflict.

[0029] As an improvement to the above technical solution, the data acquisition module, data preprocessing module, execution control module and collaborative strategy engine are deployed on the edge controller to achieve real-time closed-loop control.

[0030] The beneficial effects of this invention are as follows: This invention significantly improves the control accuracy, response speed and operational stability of constant temperature and humidity systems through bidirectional adaptive correction of thresholds and parameters, collaborative decision-making and closed-loop feedback mechanisms, effectively suppresses overshoot oscillations, reduces energy consumption, extends equipment life, and has self-learning and evolution capabilities to adapt to complex and changing working conditions.

[0031] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Of course, implementing any product or method of this application does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a system framework diagram of the present invention. Detailed Implementation

[0034] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0036] Please see Figure 1 In this embodiment of the invention, a PID adaptive intelligent control method for a constant temperature and humidity air conditioner includes the following steps:

[0037] S1. Real-time collection of environmental parameter data and equipment operation data, wherein the environmental parameters include return air temperature and humidity, fresh air temperature and humidity, and indoor disturbance factor, and the equipment operation data includes actuator opening degree, historical control deviation, and energy consumption data;

[0038] S2. Preprocess the collected data, including outlier removal, sliding filtering, trend feature extraction, and construct the current state vector of the system;

[0039] S3. Based on the aforementioned state vector, a bidirectional correction mechanism is executed synchronously:

[0040] A1: Threshold Correction Module: Based on the current environmental disturbance intensity, historical deviation trend, and equipment response delay, dynamically adjust the enthalpy or temperature and humidity values ​​to control the threshold band width and trigger boundary; the threshold correction module also supports independent dynamic adjustment of the temperature control threshold and humidity control threshold, and when a sensible heat-dominated or latent heat-dominated operating condition is detected, it switches to a single-parameter control mode with temperature or humidity as the main control variable.

[0041] B1: Parameter correction module: Based on changes in actuator characteristics, temperature and humidity coupling relationship, and control performance feedback, PID control parameters or decoupling matrix coefficients are tuned in real time;

[0042] S4. Through the collaborative strategy engine, the corrected threshold boundary and control parameter combination are matched to the current system operating state to generate the final control command;

[0043] S5. The actuator operates according to the control command to adjust the supply air temperature and humidity (or return air, indoor temperature and humidity);

[0044] S6. Collect the system response data after adjustment, evaluate the control performance indicators, and feed them back to the threshold correction module and parameter correction module to achieve closed-loop self-optimization.

[0045] In step S3, the threshold correction module includes: matching a preset threshold range based on operating condition clustering, wherein the operating condition consists of a combination of temperature and humidity, time period labels, and load level; dynamically drifting the threshold center point according to the predicted results of enthalpy and temperature and humidity value changes; setting an adaptive dead zone mechanism and a hysteresis direction compensation mechanism; the dead zone width is positively correlated with the disturbance intensity and equipment delay; and adjusting the threshold triggering direction according to the historical overshoot direction. Specifically:

[0046] Threshold partitioning based on operating condition clustering: Historical data is clustered by "temperature and humidity combination + load level + time period" using K-means or DBSCAN, and an "optimal threshold band" (such as ±0.5kJ / kg enthalpy difference allowable band) is preset for each type of operating condition. The current operating condition is matched to its category in real time, and the corresponding threshold interval is called.

[0047] Threshold drift based on deviation trend prediction: Use LSTM or exponential smoothing to predict the trend of enthalpy, temperature and humidity values ​​in the next 5-15 minutes. If the prediction will deviate rapidly, tighten the threshold in advance (e.g., from ±1.0 to ±0.3) to intervene in advance.

[0048] If the system stabilizes, the threshold is relaxed (energy-saving mode).

[0049] Regarding adaptive dead zone and hysteresis control, the dead zone width ΔH = f(environmental disturbance intensity, equipment response delay), and the hysteresis direction is dynamically adjusted based on the historical overshoot direction (e.g., if the previous humidification overshoot occurred, the current humidification threshold will be raised). More specifically:

[0050] Dynamic threshold band = baseline enthalpy value or temperature and humidity value ± [basic dead zone + α × disturbance factor + β × trend slope], where α and β are learnable weights that are optimized based on feedback from historical control effects.

[0051] In step S3, the parameter correction module includes: online identification of the actuator transfer function, updating the "valve opening matching temperature / humidity change" response model, dynamically adjusting the temperature and humidity decoupling control matrix, updating the influence coefficient every preset period based on the least squares method, automatically tuning the PID parameters based on overshoot, settling time, and integral error by using fuzzy rules or reinforcement learning algorithms, establishing an equipment aging compensation factor, and adjusting the gain coefficient as the operating time decays. Specifically, the characteristics of the actuator are identified online, a small step test (low disturbance mode) is applied to actuators such as cold coils and humidifier valves, the "opening → temperature drop / humidity increase" transfer function is identified, the control model parameters are updated, and an equipment aging compensation factor (such as efficiency decay rate η(t)) is established.

[0052] Dynamic adjustment of temperature and humidity decoupling coefficient:

[0053] Construct a temperature-humidity influence matrix (using a 2×2 static or quasi-static gain matrix), and update the matrix online every 6 hours based on least squares to achieve precise decoupling control and avoid "excessive dehumidification when cooling" or "temperature spikes when humidifying". It is understandable that adjusting the temperature will affect the humidity, and adjusting the humidity will also affect the temperature. Therefore, this matrix is ​​used to quantify this cross-influence.

[0054] Regarding PID parameter self-tuning (based on performance feedback):

[0055] Define the control performance metrics: ITAE (time-weighted absolute error integral), overshoot OS, and settling time Ts;

[0056] If OS > 5% or Ts > 15min, then fuzzy rule adjustments to Kp and Ki are triggered (where Kp is the proportional gain, which outputs the control quantity proportionally according to the current error (set value - actual value); Ki is the integral gain, which eliminates the steady-state error of the system (i.e., small deviations that exist for a long time) based on the accumulation (integration) of the error over time): large overshoot will decrease Kp and increase Kd; slow response will increase Kp and appropriately increase Ki. For example: (fuzzy PID self-tuning: IF large deviation AND rapid deviation change THEN significantly increase Kp; IF small deviation AND continuous oscillation THEN decrease Ki and increase Kd; IF long-term steady-state error THEN slightly increase Ki). Furthermore, reinforcement learning (such as PPO) can be used to automatically explore the optimal parameter combination.

[0057] In step S4, the cooperative strategy engine includes a state machine-driven mechanism that automatically switches cooperative modes based on the current system state, including:

[0058] Stable operating mode: Thresholds are relaxed, parameters are conservative;

[0059] Fast response mode: threshold tightening, high parameter gain, and feedforward;

[0060] Energy-saving optimization mode: threshold expansion, parameter smoothing, and extended integral;

[0061] Device degradation mode: Threshold compensation delay, parameter reduction response strength;

[0062] Abnormal disturbance mode: threshold pre-tightening, parameters enable robust control set.

[0063] In step S5, the control performance indicators include: integral absolute value of temperature and humidity control deviation (ITAE), overshoot percentage, system stable response time, frequency of adjustment actions per unit time, and energy efficiency ratio. The feedback mechanism adjusts the threshold correction coefficient and parameter correction weight synchronously according to preset rules based on the degree of performance indicator degradation.

[0064] It is evident that threshold correction and parameter correction are mutually influential and evolve collaboratively.

[0065] When the threshold tightens, the control frequency increases, and the PID gain needs to be reduced simultaneously to prevent oscillation.

[0066] When equipment aging causes slow response, → the threshold should be relaxed, actions should be initiated earlier, and Kp should be increased to compensate for the delay.

[0067] When environmental disturbances are severe, → start feedforward compensation + tighten threshold + switch to robust PID parameter set.

[0068] As an improvement to the above technical solution, in step S6, a feature library of historical high-quality control segments is constructed for matching similar working conditions. An online reinforcement learning algorithm is adopted to optimize the threshold boundary and control parameters by minimizing control error, energy consumption and action frequency as the reward function.

[0069] A constant temperature and humidity air conditioning system, applied to the aforementioned constant temperature and humidity air conditioning PID adaptive intelligent control method, includes a data acquisition module, a data preprocessing module, a threshold correction module, a parameter correction module, a collaborative strategy engine, an execution control module, a performance evaluation and feedback module, and a storage module. Specifically: the data acquisition module acquires real-time environmental and equipment data; the data preprocessing module performs filtering, alignment, and feature extraction; the threshold correction module dynamically adjusts enthalpy, temperature, and humidity values ​​to control trigger boundaries; the parameter correction module tunes the controller's internal parameters in real time; the collaborative strategy engine matches the current operating conditions and combines correction strategies; the execution control module outputs control commands to the actuator; the performance evaluation and feedback module quantifies the control effect and sends back optimization signals; and the storage module stores historical data, operating condition templates, and a strategy rule base.

[0070] Specifically, data preprocessing is deployed on an edge controller or PLC, the threshold correction engine is deployed on an embedded Linux controller, parameter self-tuning is deployed on a PLC function block, and the cooperative strategy state machine is set by the program. Preferably, the system is deployed in an edge controller or building automation host and connected to sensors and actuators.

[0071] In some embodiments, a data bus is also included. The data acquisition module, data preprocessing module, threshold correction module, parameter correction module, collaborative strategy engine, execution control module, and performance evaluation and feedback module are all connected to the data bus. Each module shares the current system state vector through a publish / subscribe mechanism. The state vector includes real-time enthalpy value, temperature and humidity value, deviation trend, equipment response delay, disturbance level, and operating condition label. The collaborative strategy engine, as the control terminal of the data bus, schedules the control command stream according to a preset priority to ensure that the outputs of threshold correction and parameter correction take effect synchronously without conflict. The data acquisition module, data preprocessing module, execution control module, and collaborative strategy engine are deployed on the edge controller to realize real-time closed-loop control.

[0072] By constructing a two-way correction mechanism that combines dynamic threshold drift with online parameter tuning, and integrating environmental perception, historical learning, and execution feedback, the constant temperature and humidity system can be upgraded from passive response to active prediction and adaptive adjustment.

[0073] Understandable:

[0074] Threshold correction determines "when to initiate adjustment and the boundary of adjustment intensity," and belongs to the "switching logic" of control triggering;

[0075] Parameter correction determines "how to adjust and control the adjustment amount", which belongs to the "execution strategy" of control action;

[0076] Example scenario: Sudden environmental change, such as a door suddenly opening and hot, humid air rushing in.

[0077] A1: Threshold Correction Module:

[0078] If a rapid increase in enthalpy, temperature, and humidity values ​​is detected, along with a sharp increase in the rate of change of deviation, the threshold band will be tightened immediately (from ±1.0 to ±0.3) to trigger control in advance; at the same time, the "fast response mode" status flag will be activated.

[0079] A2: Parameter Correction Module:

[0080] Receive the "Fast Response Mode" flag → switch to high-gain PID parameter set (Kp↑, Ki↑); start feedforward compensation (pre-open the chilled water valve according to the sudden change in fresh air temperature and humidity); temporarily strengthen the "cooling takes precedence over dehumidification" weight in the decoupling matrix.

[0081] A3: Execution result:

[0082] The system can activate strong cooling and dehumidification within 30 seconds and stabilize within 2 minutes, avoiding the temperature and humidity spikes caused by the 5-8 minute delay in response of traditional controls.

[0083] A4: Feedback Learning

[0084] This adjustment resulted in a 2% overshoot. In the next scenario, the threshold will be tightened by 10%, and Kp will be reduced by 5%.

[0085] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. A PID adaptive intelligent control method for a constant temperature and humidity air conditioner, characterized in that, Includes the following steps: S1. Real-time collection of environmental parameter data and equipment operation data, wherein the environmental parameters include return air temperature and humidity, fresh air temperature and humidity, and indoor disturbance factor, and the equipment operation data includes actuator opening degree, historical control deviation, and energy consumption data; S2. Preprocess the collected data, including outlier removal, sliding filtering, trend feature extraction, and construct the current state vector of the system; S3. Based on the aforementioned state vector, a bidirectional correction mechanism is executed synchronously: A1: Threshold Correction Module: Based on the current environmental disturbance intensity, historical deviation trend, and equipment response delay, dynamically adjust the enthalpy or temperature and humidity values ​​to control the threshold band width and trigger boundary;    B1: Parameter correction module: Based on changes in actuator characteristics, temperature and humidity coupling relationship, and control performance feedback, PID control parameters or decoupling matrix coefficients are tuned in real time; S4. Through the collaborative strategy engine, the corrected threshold boundary and control parameter combination are matched to the current system operating state to generate the final control command; S5. The actuator operates according to the control command, adjusting the supply air temperature and humidity or return air temperature and humidity, and the indoor temperature and humidity; S6. Collect system response data after adjustment, evaluate control performance indicators, and feed them back to the threshold correction module and parameter correction module; In step S3, the threshold correction module includes: matching a preset threshold range based on working condition clustering, wherein the working condition consists of temperature and humidity combination, time period label, and load level; dynamically drifting the threshold center point according to the prediction result of the enthalpy value or temperature and humidity value change trend; setting an adaptive dead zone mechanism and a hysteresis direction compensation mechanism; the dead zone width is positively correlated with the disturbance intensity and equipment delay; and adjusting the threshold trigger direction according to the historical overshoot direction. In step S3, the parameter correction module includes: online identification of the actuator transfer function, updating the "valve opening matching temperature / humidity change" response model, dynamically adjusting the temperature and humidity decoupling control matrix, updating the influence coefficient every preset period based on the least squares method, automatically tuning the PID parameters based on overshoot, settling time, and integral error by using fuzzy rules or reinforcement learning algorithms, establishing an equipment aging compensation factor, and adjusting the gain coefficient as the running time decays.

2. The PID adaptive intelligent control method for a constant temperature and humidity air conditioner according to claim 1, characterized in that: In step S4, the cooperative strategy engine includes a state machine-driven mechanism that automatically switches cooperative modes based on the current system state, including: Stable operating mode: Thresholds are relaxed, parameters are conservative; Fast response mode: threshold tightening, high parameter gain, and feedforward; Energy-saving optimization mode: threshold expansion, parameter smoothing, and extended integral; Device degradation mode: Threshold compensation delay, parameter reduction response strength; Abnormal disturbance mode: threshold pre-tightening, parameters enable robust control set.

3. The PID adaptive intelligent control method for a constant temperature and humidity air conditioner according to claim 1, characterized in that: In step S5, the control performance indicators include: absolute integral value of temperature and humidity control deviation (ITAE), overshoot percentage, system stable response time, frequency of adjustment actions per unit time, and energy efficiency ratio. The feedback mechanism adjusts the threshold correction coefficient and parameter correction weight synchronously according to preset rules based on the degree of performance indicator degradation.

4. The PID adaptive intelligent control method for a constant temperature and humidity air conditioner according to claim 3, characterized in that, In step S6, a feature library of historical high-quality control segments is constructed for matching similar operating conditions. An online reinforcement learning algorithm is adopted, with the minimum control error, energy consumption and action frequency as the reward function, to jointly optimize the threshold boundary and control parameters.

5. A constant temperature and humidity air conditioning system, applied to the constant temperature and humidity air conditioning PID adaptive intelligent control method as described in claim 1, characterized in that, The system includes a data acquisition module, a data preprocessing module, a threshold correction module, a parameter correction module, a collaborative strategy engine, an execution control module, a performance evaluation and feedback module, and a storage module. Specifically: the data acquisition module acquires real-time environmental and equipment data; the data preprocessing module performs filtering, alignment, and feature extraction; the threshold correction module dynamically adjusts enthalpy or temperature and humidity values ​​to control trigger boundaries; the parameter correction module tunes the controller's internal parameters in real time; the collaborative strategy engine matches the current operating conditions and combines correction strategies; the execution control module outputs control commands to the actuators; the performance evaluation and feedback module quantifies the control effect and sends back optimization signals; and the storage module saves historical data, operating condition templates, and a strategy rule base.

6. The constant temperature and humidity air conditioning system according to claim 5, characterized in that: The system is deployed in an edge controller or building automation host and connects to sensors and actuators.

7. The constant temperature and humidity air conditioning system according to claim 5, characterized in that: It also includes a data bus, and the data acquisition module, data preprocessing module, threshold correction module, parameter correction module, collaborative strategy engine, execution control module, and performance evaluation and feedback module are all connected to the data bus. Each module shares the current system state vector through a publish / subscribe mechanism. The state vector includes real-time enthalpy value, temperature and humidity value, deviation trend, equipment response delay, disturbance level, and operating condition label. The collaborative strategy engine acts as the control terminal of the data bus and schedules the control command stream according to a preset priority.

8. The constant temperature and humidity air conditioning system according to claim 5, characterized in that: The data acquisition module, data preprocessing module, execution control module, and collaborative strategy engine are deployed on the edge controller.

Images