Constant-temperature and constant-humidity air conditioning system regulation and control method and system based on artificial intelligence

Abstract

A constant-temperature and constant-humidity air conditioning system regulation and control method and system based on artificial intelligence belongs to the technical field of air conditioning intelligent control, and comprises the following steps: collecting real-time environment data through sensor networks deployed in different areas of a workshop, and preprocessing the real-time environment data; the real-time environment data comprises temperature, humidity and energy consumption; performing data compression and protocol packaging on the preprocessed real-time environment data, and uploading the data to a control center; the method comprises the following steps: acquiring running parameters of all current running air conditioners, performing prediction in a control center through a regulation and control analysis model based on the running parameters of the air conditioners, outputting a temperature prediction value and a humidity prediction value of a future time window, and when the future time window is reached, according to the acquired actual temperature, humidity and energy consumption, determining whether the air conditioners run or not. And respectively comparing the temperature prediction value and the humidity prediction value with the previously predicted temperature prediction value and the previously predicted humidity prediction value at the corresponding moment and a set energy consumption threshold value, and correcting the operation parameters of the air conditioner. Accurate regulation and control of the constant-temperature and constant-humidity air conditioning system based on artificial intelligence are achieved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent air conditioning control technology, specifically relating to a method and system for regulating a constant temperature and humidity air conditioning system based on artificial intelligence. Background Technology

[0002] With the increasing demands for precise environmental control in industrial production, constant temperature and humidity air conditioning systems play a crucial role in important settings such as power workshops, data centers, and laboratories. In power workshops, fluctuations in temperature and humidity can lead to localized overheating of electrical equipment, decreased insulation performance, or equipment failure, thereby affecting the safe and stable operation of the power system. Therefore, achieving precise control of workshop temperature and humidity in a dynamic environment has become key to ensuring equipment reliability and lifespan.

[0003] While existing technologies such as fuzzy control, PID control, swarm intelligence optimization, and data-driven prediction have improved control performance to some extent, they generally suffer from problems such as response lag, poor adaptability, insufficient accuracy, and high energy consumption, making it difficult to meet the demands for efficient and precise control under complex operating conditions. Therefore, how to achieve real-time prediction and dynamic control of temperature and humidity environments to effectively improve control accuracy and energy efficiency is an urgent problem to be solved. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for controlling a constant temperature and humidity air conditioning system based on artificial intelligence.

[0005] The present invention adopts the following technical solution.

[0006] The first aspect of this invention proposes a method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence, comprising: S1: Real-time environmental data is collected through a sensor network deployed in different areas of the workshop and preprocessed; the real-time environmental data includes temperature, humidity and energy consumption; S2: After compressing and encapsulating the preprocessed real-time environmental data, it is uploaded to the control center; S3: Obtain the operating parameters of all currently running air conditioners. At the control center, based on these operating parameters, a control analysis model is used to predict the temperature and humidity for future time windows. The control analysis model uses a state transition equation for prediction. The state transition matrix in the state transition equation obtains initial values ​​based on the heat balance equation and the humidity balance equation, and then obtains the final value using the initial values ​​and the least squares method. The control input matrix in the state transition equation calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then obtains the final value using the initial values ​​and the least squares method. S4: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold, and the air conditioner's operating parameters are corrected online.

[0007] Preferably, the regulation analysis model is specifically: The operating parameters of the air conditioner include the air supply temperature setpoint of the air conditioning unit, the fan speed of the air conditioning unit, the opening degree of the humidification valve, and the opening degree of the dehumidification valve; the operating parameters of all the air conditioners in operation are combined into an input vector; The predicted temperature and humidity of each region are set as state vectors, which are sets of temperature and humidity of different regions at the corresponding time; the actual measured temperature and humidity are used as observation vectors. Construct a state equation, wherein the predicted state vector at the next time step is equal to the current state vector multiplied by the state transition matrix, plus the control input matrix multiplied by the input vector, plus a set bias term.

[0008] Preferably, the state transition matrix is ​​specifically: The state transition matrix is ​​a horizontal concatenation of the temperature-temperature influence matrix and the temperature-humidity influence matrix, and a horizontal concatenation of the humidity-temperature influence matrix and the humidity-humidity influence matrix. The two horizontally concatenated matrices are then column-wise concatenated. The temperature-temperature influence matrix, the temperature-humidity influence matrix, the humidity-temperature influence matrix, and the humidity-humidity influence matrix are all n×n matrices, where n is the total number of regions. The initial values ​​of the elements in the temperature-humidity effect matrix and the humidity-temperature effect matrix are set values; The initial values ​​of the elements of the temperature-temperature effect matrix and the humidity-humidity effect matrix are calculated based on the heat balance equation and the humidity balance equation; The final values ​​of the temperature-temperature effect matrix, temperature-humidity effect matrix, humidity-temperature effect matrix, and humidity-humidity effect matrix are calculated using the least squares method.

[0009] Preferably, the initial values ​​of the temperature-temperature influence matrix and the humidity-humidity influence matrix are calculated based on the heat balance equation and the humidity balance equation, specifically: Establish the heat balance equation and the humidity balance equation, and linearize the heat balance equation and the humidity balance equation at the set operating point to obtain the linear ordinary differential equation of temperature and the linear ordinary differential equation of humidity. The heat transfer coefficient between each two regions is obtained by solving the linear ordinary differential equation of temperature based on historical temperature data at different times when the air conditioning system is in non-controlled mode. Based on historical humidity data at different times when the air conditioning system is in non-controlled mode, the humidity transfer coefficient between each two regions is solved by the linear ordinary differential equation of humidity. The initial value of the element in the i-th row and i-th column of the temperature-temperature influence matrix is ​​calculated by subtracting the sampling time interval from 1, dividing by the air heat capacity of the i-th region, and then multiplying by the sum of the heat transfer coefficients of the i-th region and all other regions; other regions refer to regions other than the i-th region. The initial value for the element in the i-th row and j-th column of the temperature-temperature influence matrix is ​​the sampling time interval divided by the air heat capacity of the i-th region and then multiplied by the heat transfer coefficients of the i-th and j-th regions. The initial value of the element in the i-th row and i-th column of the humidity-humidity influence matrix is ​​calculated by subtracting the sampling time interval from 1, dividing by the air humidity capacity of the i-th region, and then multiplying by the sum of the moisture transfer coefficients of the i-th region and all other regions. The initial value for the element in the i-th row and j-th column of the humidity-humidity influence matrix is ​​the sampling time interval divided by the air humidity capacity of the i-th region and then multiplied by the moisture transfer coefficients of the i-th and j-th regions.

[0010] Preferably, the linear ordinary differential equations for temperature and humidity are as follows: The linear ordinary differential equation for temperature is as follows: multiply the heat transfer coefficients of the i-th and j-th regions at the current moment by the temperature difference between the j-th and i-th regions at the current moment, and use this as the influence temperature of the j-th region on the i-th region at the current moment. Sum the influence temperatures of all other regions on the i-th region at the current moment, add the set heat generation rate of the internal equipment heat source in the i-th region, add the set external heat exchange parameters, multiply the sum by the sampling time interval and divide by the air heat capacity of the i-th region to get the temperature change of the region in the non-controlled mode of the air conditioning system, and add the temperature change of the region in the non-controlled mode of the air conditioning system to the temperature of the i-th region at the current moment to get the temperature of the i-th region at the next moment. The linear ordinary differential equation for humidity is as follows: Multiply the humidity transfer coefficients of the i-th and j-th regions at the current moment by the difference between the humidity of the j-th region and the i-th region at the current moment, and use this as the humidity of the j-th region on the i-th region at the current moment. Sum the humidity of all other regions on the i-th region at the current moment, add the set internal process humidity generation rate of the i-th region, add the set external airtightness humidity exchange, multiply the sum by the sampling time interval and divide by the air humidity capacity of the i-th region to get the humidity change of the region when the air conditioning system is in non-controlled mode, and add the humidity change of the region when the air conditioning system is in non-controlled mode to the humidity of the i-th region at the current moment to get the humidity of the i-th region at the next moment.

[0011] Preferably, the final values ​​of the temperature-temperature influence matrix, temperature-humidity influence matrix, humidity-temperature influence matrix, and humidity-humidity influence matrix are calculated using the least squares method, specifically: Obtain historical temperature and humidity data for each moment within a set period when the air conditioning system is in non-control mode. The error at time v is calculated by subtracting the product of the state transition matrix and the state vector at time v from the state vector at time v+1 within the set period. The objective function of the least squares method is to calculate the sum of squares at each time point and minimize it. Each time point includes time points 1, 2, 3, ..., N-1; N is the total number of time points within the set period.

[0012] Preferably, the control input matrix is ​​specifically: The control input matrix consists of a temperature control input matrix and a humidity control input matrix arranged vertically. Both the temperature control input matrix and the humidity control input matrix are n×p matrices, where p is the total number of elements in the input vector. The initial value of the element in the i-th row and m-th column of the temperature control input matrix is ​​the temperature control related term of the m-th air conditioning parameter in the i-th region, calculated according to the linear equation of cooling power; the temperature control related term of the m-th air conditioning parameter in the i-th region is the sampling time interval divided by heat capacity, multiplied by air density, multiplied by air specific heat capacity, and multiplied by the air supply volume affected by the m-th air conditioning parameter in the i-th region. The initial value of the element in the i-th row and m-th column of the humidity control input matrix is ​​the humidity control related term of the m-th air conditioning parameter in the i-th region, calculated according to the linear equation of dehumidification power. The humidity control related term is the air supply volume affected by the m-th air conditioning parameter in the i-th region, calculated by dividing the sampling time interval by the humidity capacity, multiplying by the air density, multiplying by the air specific humidity, and multiplying by the air supply volume affected by the m-th air conditioning parameter in the i-th region. The final values ​​of the temperature control input matrix and the humidity control input matrix are calculated using the least squares method.

[0013] Preferably, the linear equation for cooling power is: when only the m-th air conditioning parameter is adjusted, the temperature of the i-th region at the next moment is equal to the temperature of the i-th region at the current moment plus the temperature control related term of the m-th air conditioning parameter in the i-th region multiplied by the cooling capacity of the air conditioning output affected by the adjustment of the corresponding m-th air conditioning parameter. The linear equation for dehumidification power is: when only the m-th air conditioning parameter is adjusted, the humidity of the i-th region at the next moment is equal to the humidity of the i-th region at the current moment plus the humidity control related term of the m-th air conditioning parameter in the i-th region multiplied by the humidity content of the air conditioning output affected by the adjustment of the corresponding m-th air conditioning parameter.

[0014] Preferably, the final values ​​of the temperature control input matrix and the humidity control input matrix are calculated using the least squares method, specifically: Obtain historical temperature and humidity data of the air conditioning system at each moment within a set cycle when only the m-th air conditioning parameter is adjusted. When only the m-th air conditioning parameter is adjusted, the state vector of the air conditioning system at time v+1 within the set period is subtracted from the product of the state transition matrix and the state vector at time v, and then the product of the current control input matrix and the input vector is subtracted to obtain the error at time v. The sum of squares at each time point is calculated to minimize the sum of squares as the objective function of the least squares method. Each time point includes the 1st, 2nd, 3rd, ..., N-1th time points; N is the total number of time points within the set period. When performing the least squares method, only the elements in the m-th column of all rows of the temperature control input matrix and the elements in the m-th column of all rows of the humidity control input matrix are adjusted. Repeat all of the above steps for all air conditioning parameters to obtain the final control input matrix.

[0015] Preferably, when the future time window is reached, the actual temperature, humidity, and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and a set energy consumption threshold, respectively, to perform online correction of the air conditioner's operating parameters. Specifically: When the next moment arrives, it is determined whether the correction condition has been met. If it has, the corresponding correction is performed. If it has not been met, the actual temperature, humidity and energy consumption obtained at the next moment are used to determine whether the correction condition has been met, until the correction condition is met or the current moment reaches the total time of the set future time window. If the current moment reaches the total time of the set future time window, the steps S3-S4 are repeated from S3.

[0016] Preferably, the correction conditions and the corresponding corrections are as follows: If the difference between the actual measured temperature and the predicted temperature exceeds the set temperature difference, the correction condition is met, and the supply air temperature setting is reduced according to the set first step length. If the difference between the actual measured humidity and the predicted humidity exceeds the set maximum humidity difference, the correction condition is met and the opening of the dehumidifier valve is increased according to the set second step size. If the difference between the actual measured humidity and the predicted humidity is less than the set minimum humidity difference, the correction condition is met and the opening of the humidifier valve is increased according to the set second step size. If the actual measured energy consumption exceeds the set energy consumption threshold, the supply air temperature setting will be increased or the fan speed of the air conditioning unit will be reduced according to the set third step. If the current time reaches the total time of the set future time window, then repeat steps S3-s4 starting from S3 again.

[0017] A second aspect of the present invention provides an artificial intelligence-based control system for a constant temperature and humidity air conditioning system using the method described in the first aspect of the present invention, comprising a data acquisition module, a data transmission module, a prediction module, and a correction module, specifically: Data Acquisition Module: Collects real-time environmental data through a sensor network deployed in different areas of the workshop and performs preprocessing; the real-time environmental data includes temperature, humidity, and energy consumption. Data transmission module: Compresses and encapsulates the pre-processed real-time environmental data before uploading it to the control center; Prediction Module: This module acquires the operating parameters of all currently running air conditioners. Based on these parameters, it uses a control analysis model at the control center to predict temperature and humidity values ​​for future time windows. The control analysis model uses state transition equations for prediction. The state transition matrix in these equations obtains initial values ​​based on the heat balance and humidity balance equations, and then uses these initial values ​​and the least squares method to obtain the final values. The control input matrix in the state transition equations calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then uses these initial values ​​and the least squares method to obtain the final values. Correction module: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold to perform online correction of the air conditioner's operating parameters.

[0018] A third aspect of the present invention provides a computer device including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described in the first aspect of the present invention.

[0019] A fourth aspect of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method described in the first aspect of the present invention.

[0020] The beneficial effects of this invention are that, compared with the prior art, By combining physical mechanism models with data-driven algorithms, the system is initialized based on explicit physical laws and then optimized using historical data. This results in a model with a solid physical foundation that can adapt to the dynamic characteristics of specific workshops, significantly improving prediction accuracy. The system continuously corrects its predictions and future states based on the latest measured data. This allows the system not only to predict future environmental changes but also to immediately self-adjust when predictions deviate. From data acquisition, transmission, and preprocessing to model prediction, strategy generation, online correction, and historical data recording and model self-updating, a complete autonomous closed loop is formed. This significantly reduces reliance on human experience, lowers operational labor costs, and constructs a new generation of constant temperature and humidity environment control system integrating precise sensing, intelligent prediction, forward-looking control, energy efficiency optimization, and safety and reliability. Attached Figure Description

[0021] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0023] like Figure 1 As shown, Embodiment 1 of the present invention proposes a method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence, comprising: S1: Real-time environmental data is collected through a sensor network deployed in different areas of the workshop and preprocessed; the real-time environmental data includes temperature, humidity and energy consumption; In S1, the preprocessing is based on physical model outlier correction. Specifically, the constraint relationship between environmental parameters is established using the thermodynamic equation of the air conditioning system. When the sensor data violates the constraint relationship, Kalman filtering is used for correction. The constraints include: 1. Inter-regional gradient constraints: In the absence of strong local heat / cold sources or strong airflow disturbances, the temperature and humidity changes in adjacent regions should exhibit spatial continuity, and their gradients should be within a preset range. 2. Saturated water vapor pressure constraint: The relative humidity of the air must not exceed 100% (saturated state); 3. Air Conditioner Performance Constraints: There is a definite physical relationship between the air conditioner's outlet temperature and its set status (cooling / heating mode, set temperature, fan speed), which is obtained from the manufacturer.

[0024] S2: After compressing and encapsulating the preprocessed real-time environmental data, it is uploaded to the control center; The data compression and protocol encapsulation are specifically as follows: Environmental parameters are collected using sensors that support RS485 interfaces. The sensors are connected to the edge gateway via an RS485 bus in a multi-point daisy chain. The edge gateway acts as the master station and collects data from each sensor through a polling mechanism. After receiving all Modbus RTU data, the edge gateway summarizes the data, compresses and formats it, and removes outliers. The processed data from the edge gateway is then uploaded to the control center via Modbus TCP.

[0025] S3: Obtain the operating parameters of all currently running air conditioners. At the control center, based on these operating parameters, a control analysis model is used to predict the temperature and humidity for future time windows. The control analysis model uses a state transition equation for prediction. The state transition matrix in the state transition equation obtains initial values ​​based on the heat balance equation and the humidity balance equation, and then obtains the final value using the initial values ​​and the least squares method. The control input matrix in the state transition equation calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then obtains the final value using the initial values ​​and the least squares method. Specifically, the operation process of the regulation analysis model in S3 includes: S301: Obtain the settings parameters for all currently running air conditioners; S302: Model the workshop environment as a linear dynamic system and perform rolling predictions: S302-1: Set the temperature and humidity of each area to be predicted as a state vector. ,

[0026] in, Let the temperature of regions 1, 2, ..., n at time k. The humidity of regions 1, 2, ..., n at time k, where time k is the current time and n is the total number of regions.

[0027] Set the real-time measured temperature and humidity as the observation vector z. k ; Constructing the state equations:

[0028] in The state transition matrix describes the state vector of the system from the current time step when there is no control. State vector to the next time step The dynamic evolution of the state; with embedded thermal balance constraints and humidity balance constraints; To control the input matrix, describe Regarding system status The impact; The input vector consists of air conditioning parameters.

[0029] Construct the observation equation: ; in, The observation matrix maps the system's state to observed values. ; To observe the noise, it follows a Gaussian distribution; S302-2: Predict temperature and humidity for the next N time steps, and execute a loop: Initialization: Based on the observations at the current time k and Initialize the filter; Given the air conditioning parameters under the corresponding time-to-time control strategy, , ,..., ; Forward prediction is performed based on the state equation:

[0030] A series of { , ,..., } represents the predicted temperature and humidity values ​​for the future time window; M is the total number of moments in the future time window. When the system reaches the next time step k+1, it acquires new measured data z. k+1 ; Compare predicted values ​​using filters With the new observation z k+1 The differences are corrected using online correction methods.

[0031] Specifically, the construction process of the state transition matrix A is as follows: A01: Establish a continuous-time physical model and formulate differential equations for each region based on thermal equilibrium and moisture equilibrium; The heat balance equation simplifies to:

[0032] in, Let be the air heat capacity of region i. Let be the heat transfer coefficient between region i and region j; The heat generation rate of the heat source for equipment within the region; For external heat exchange, the parameters are fixed and set according to the actual conditions of the area; , Let be the temperatures of region i and region j, respectively; For all regions except i; Similarly, the wet balance equation simplifies to:

[0033] in, Let i be the humidity capacity of region i. The moisture transfer coefficient between region i and region j The rate of moisture generation from the process within the region; The parameters are fixed and set according to the actual conditions of the area to ensure airtight moisture exchange with the outside environment. A02: The above nonlinear continuous-time model is linearized around the set operating points (average temperature, average humidity) to obtain a set of linear ordinary differential equations; based on the historical temperature data at different times when the air conditioning system is in non-controlled mode, the humidity transfer coefficient between each two regions is solved by the linear ordinary differential equation of temperature; based on the historical humidity data at different times when the air conditioning system is in non-controlled mode, the humidity transfer coefficient between each two regions is solved by the linear ordinary differential equation of humidity. continuous-time derivative Convert to discrete-time difference The sampling time interval is (Set to 1 minute or 5 minutes, consistent with the system control cycle); The discretized equation is:

[0034]

[0035] in, , The temperatures of region j and region i at time k are respectively. Let be the temperature of region i at time k+1. , Let be the humidity of region j and region i at the k-th time, respectively. Let be the humidity of region i at time k+1. This formula shows... How to depend on the temperature of all regions at the current moment, the dependency relationship is... Awaiting decision; and display How to depend on the humidity of all areas at the current moment, the dependency relationship is... Awaiting decision; A03: State vector x k The dimension is 2n (n temperatures, n humidity levels); therefore, matrix A is a 2n×2n square matrix. Will Divide into 4 n×n square matrices;

[0036] in, This is a temperature-temperature effect matrix, representing the effect of temperature on the temperature at the next moment. This is a temperature-humidity influence matrix, representing the effect of humidity on temperature at the next moment (usually weak, and can be initialized to zero or a value less than a certain set threshold). This is a humidity-temperature effect matrix, representing the effect of temperature on humidity at the next moment (usually weak, and can be initialized to zero or a value less than a certain set threshold). This is the humidity-humidity effect matrix, representing the effect of humidity on humidity at the next moment. , The diagonal elements represent the region's inherent inertia; for temperature,

[0037]

[0038] in, for The initial value of the element in the i-th row and i-th column should be close to 1 but slightly less than 1 due to thermal inertia. for The initial value of the element in the i-th row and i-th column is also close to 1 but slightly less than 1; , The off-diagonal elements represent the coupling of regions; for temperature,

[0039]

[0040] in, for The initial value of the element in the i-th row and j-th column is determined according to... Settings (for) (Adjacent regions have a small positive coefficient, while non-adjacent regions have a coefficient of 0). for The initial value of the element in the i-th row and j-th column; A04: The physical model provides initial structure and parameter estimates, but accurate coefficients need to be learned from historical data through system identification; During a set period of operation in non-controlled mode of the air conditioning system, a sequence of state vectors undergoing purely natural evolution is collected. , ,..., Ensure that the data for the set period covers different operating conditions (such as weekdays / rest days, daytime / nighttime) to capture comprehensive system dynamics.

[0041] The problem is transformed into a linear regression problem to find the final matrix. , making

[0042] in, , These are connected to the non-control mode state within the set period; To set the total number of moments within a given period; Specifically, the control input matrix The build process is as follows: Control input matrix It quantitatively describes the control input vector. (i.e., air conditioning control strategy) on system state The direct impact of (i.e., the temperature and humidity of the workshop at the next moment).

[0043] Therefore, it is necessary to construct a control input matrix that can accurately quantify the impact of air conditioning control actions on the temperature and humidity of various areas in the workshop. B01: Define the input vector Clearly define the adjustable air conditioning parameters; This includes the supply air temperature setpoint, the air conditioning unit's fan speed, the humidification valve opening, and the dehumidification valve opening; analyze How does each element affect the energy and mass balance of its service area through the laws of physics? For the supply air temperature setpoint, when the supply air temperature decreases, the amount of cooling delivered to the room by the air conditioner increases. For the fan speed of the air conditioning unit, increasing the fan speed increases the supply air volume. It changed the cold air delivery capacity. This changes the inter-region coupling coefficient in matrix B; and the cooling power of region i. It can be modeled as: ; in, air density, The specific heat capacity of air, This refers to the air supply volume; Let be the lumped parameter temperature of region i, which is numerically equal to the measured value of the sensors deployed in that region. And in the model, it is used as a single variable representing the overall thermal state of the region. This represents the change in cooling capacity output by the air conditioner caused by the change in the m-th parameter. for The m-th air conditioning parameter in the data; The same modeling approach is used for humidity.

[0044] B02: Input vector u k The dimension is p, and there are p control variables; therefore, the control input matrix B is a 2n×p matrix. B is two n×p matrices:

[0045] B T Temperature control input matrix, the effect of control input on temperature, B H This is the humidity control input matrix, which controls the impact of the input on humidity. Matrix B is highly sparse. A control input (such as the setpoint of an air conditioner) typically only affects its air supply area and adjacent areas, without directly affecting distant areas. Initially, the coefficients for unaffected areas are set to 0.

[0046] For air conditioning input serving region i Its impact is mainly reflected in the corresponding matrix. The m-th parameter Above; and the corresponding matrix The m-th parameter superior; Based on the physical model in B01, the cooling power is linearly discretized.

[0047] Among them, with Control-related items are , for The initial estimate is a negative coefficient; Humidity matrix B H Similarly, the formula is:

[0048] Among them, with Control-related items are , for The initial estimate is a negative coefficient; B03: Physical initialization provides the order of magnitude, but the precise gain needs to be identified through experiments and data; Keep the system stable in other aspects (e.g., turn off other sources of interference), then obtain the historical temperature and humidity data of the air conditioning system at each moment within the set period when only the m-th air conditioning parameter is adjusted (e.g., lower the supply air temperature setpoint of air conditioner No. 1 by 1°C), and maintain this change. Obtain all state vectors within the set period.

[0049] Under a step input, the response of the state equation mainly reflects the influence of B (because A dominates the natural dynamics, while the step input forcibly changes the system). Analyze the relationship between changes in the state and changes in the control input; Ignoring the transient details of A, we re-identify the entire transient response data using the least squares method; By adjusting only the elements in the m-th column of all rows of the temperature control input matrix and the elements in the m-th column of all rows of the humidity control input matrix, the control input matrix after adjusting the elements in the m-th column is calculated using a regression model. :

[0050] in, To set the state vectors at time v+1 and time v d within the cycle when only the m-th air conditioning parameter is adjusted, Let v+1 be the input vector at time v; change the air conditioning parameters until all air conditioning parameters from the 1st to the pth time have been adjusted to obtain the final control input matrix; S4: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold, and the air conditioner's operating parameters are corrected online.

[0051] Specifically, the online calibration method includes: When the future time window is reached, the actual temperature, humidity, and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold to perform online correction of the air conditioner's operating parameters. Specifically: When the next moment arrives, it is determined whether the correction condition has been met. If it has, the corresponding correction is performed. If it has not been met, the actual temperature, humidity and energy consumption obtained at the next moment are used to determine whether the correction condition has been met, until the correction condition is met or the current moment reaches the total time of the set future time window. If the current moment reaches the total time of the set future time window, the steps S3-S4 are repeated from S3. The correction conditions and corresponding corrections are as follows: If the difference between the actual measured temperature and the predicted temperature exceeds the set temperature difference, the correction condition is met, and the supply air temperature setting is reduced according to the set first step length. If the difference between the actual measured humidity and the predicted humidity exceeds the set maximum humidity difference, the correction condition is met and the opening of the dehumidifier valve is increased according to the set second step size. If the difference between the actual measured humidity and the predicted humidity is less than the set minimum humidity difference, the correction condition is met and the opening of the humidifier valve is increased according to the set second step size. If the actual measured energy consumption exceeds the set energy consumption threshold, the supply air temperature setting will be increased or the fan speed of the air conditioning unit will be reduced according to the set third step. Based on the fine-tuning rules, a new set of fine-tuned air conditioning operating parameters is generated as a correction strategy.

[0052] If the current time reaches the total time of the set future time window, then repeat steps S3-s4 starting from S3 again.

[0053] All calibrated air conditioning parameters must be compared with the preset safe operating boundaries and process requirement boundaries in the system. These boundaries include: equipment physical limits (such as maximum / minimum supply air temperature, compressor maximum frequency), strict temperature and humidity ranges required by the process, and minimum start-up and shutdown intervals (to prevent losses caused by frequent start-ups and shutdowns). If the correction strategy violates any constraints, it is corrected to the nearest boundary value. After verification, a safe and feasible online correction strategy is output.

[0054] S404: The validated correction strategy is issued to the workshop air conditioning system for execution, overriding previous control instructions. The entire process data (including the cause of deviation, correction actions, and correction results) from the initial strategy to actual feedback, correction operation, and post-correction effect is recorded in the historical database. This data will be used to: periodically feed back to the prediction model in S3 and the optimization model in S4 for training and refining the models, reducing the deviation between future predictions and actual results; and optimize the fine-tuning step size and rules in S403, making the online correction process increasingly accurate and efficient.

[0055] Embodiment 2 of the present invention proposes an artificial intelligence-based control system for a constant temperature and humidity air conditioning system using the method described in Embodiment 1 of the present invention, comprising a data acquisition module, a data transmission module, a prediction module, and a correction module, specifically as follows: Data Acquisition Module: Collects real-time environmental data through a sensor network deployed in different areas of the workshop and performs preprocessing; the real-time environmental data includes temperature, humidity, and energy consumption. Data transmission module: Compresses and encapsulates the pre-processed real-time environmental data before uploading it to the control center; Prediction Module: This module acquires the operating parameters of all currently running air conditioners. Based on these parameters, it uses a control analysis model at the control center to predict temperature and humidity values ​​for future time windows. The control analysis model uses state transition equations for prediction. The state transition matrix in these equations obtains initial values ​​based on the heat balance and humidity balance equations, and then uses these initial values ​​and the least squares method to obtain the final values. The control input matrix in the state transition equations calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then uses these initial values ​​and the least squares method to obtain the final values. Correction module: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold to perform online correction of the air conditioner's operating parameters.

[0056] Embodiment 3 of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in Embodiment 1 of the present invention.

[0057] Embodiment 4 of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method described in Embodiment 1 of the present invention.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence, characterized in that, include: S1: Real-time environmental data is collected through a sensor network deployed in different areas of the workshop and preprocessed; the real-time environmental data includes temperature, humidity and energy consumption; S2: After compressing and encapsulating the preprocessed real-time environmental data, it is uploaded to the control center; S3: Obtain the operating parameters of all currently running air conditioners. At the control center, based on these operating parameters, a control analysis model is used to predict the temperature and humidity for future time windows. The control analysis model uses a state transition equation for prediction. The state transition matrix in the state transition equation obtains initial values ​​based on the heat balance equation and the humidity balance equation, and then obtains the final value using the initial values ​​and the least squares method. The control input matrix in the state transition equation calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then obtains the final value using the initial values ​​and the least squares method. S4: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold, and the air conditioner's operating parameters are corrected online.

2. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 1, characterized in that: The regulation and analysis model is specifically as follows: The operating parameters of the air conditioner include the air supply temperature setpoint of the air conditioning unit, the fan speed of the air conditioning unit, the opening degree of the humidification valve, and the opening degree of the dehumidification valve; the operating parameters of all the air conditioners in operation are combined into an input vector; The predicted temperature and humidity of each region are set as state vectors, which are sets of temperature and humidity of different regions at the corresponding time; the actual measured temperature and humidity are used as observation vectors. Construct a state equation, wherein the predicted state vector at the next time step is equal to the current state vector multiplied by the state transition matrix, plus the control input matrix multiplied by the input vector, plus a set bias term.

3. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 2, characterized in that: The state transition matrix is ​​specifically as follows: The state transition matrix is ​​a horizontal concatenation of the temperature-temperature influence matrix and the temperature-humidity influence matrix, and a horizontal concatenation of the humidity-temperature influence matrix and the humidity-humidity influence matrix. The two horizontally concatenated matrices are then column-wise concatenated. The temperature-temperature influence matrix, the temperature-humidity influence matrix, the humidity-temperature influence matrix, and the humidity-humidity influence matrix are all n×n matrices, where n is the total number of regions. The initial values ​​of the elements in the temperature-humidity effect matrix and the humidity-temperature effect matrix are set values; The initial values ​​of the elements of the temperature-temperature effect matrix and the humidity-humidity effect matrix are calculated based on the heat balance equation and the humidity balance equation; The final values ​​of the temperature-temperature effect matrix, temperature-humidity effect matrix, humidity-temperature effect matrix, and humidity-humidity effect matrix are calculated using the least squares method.

4. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 3, characterized in that: The initial values ​​of the temperature-temperature effect matrix and humidity-humidity effect matrix are calculated based on the heat balance equation and the humidity balance equation, specifically: Establish the heat balance equation and the humidity balance equation, and linearize the heat balance equation and the humidity balance equation at the set operating point to obtain the linear ordinary differential equation of temperature and the linear ordinary differential equation of humidity. The heat transfer coefficient between each two regions is obtained by solving the linear ordinary differential equation of temperature based on historical temperature data at different times when the air conditioning system is in non-controlled mode. Based on historical humidity data at different times when the air conditioning system is in non-controlled mode, the humidity transfer coefficient between each two regions is solved by the linear ordinary differential equation of humidity. The initial value of the element in the i-th row and i-th column of the temperature-temperature influence matrix is ​​calculated by subtracting the sampling time interval from 1, dividing by the air heat capacity of the i-th region, and then multiplying by the sum of the heat transfer coefficients of the i-th region and all other regions; other regions refer to regions other than the i-th region. The initial value for the element in the i-th row and j-th column of the temperature-temperature influence matrix is ​​the sampling time interval divided by the air heat capacity of the i-th region and then multiplied by the heat transfer coefficients of the i-th and j-th regions. The initial value of the element in the i-th row and i-th column of the humidity-humidity influence matrix is ​​calculated by subtracting the sampling time interval from 1, dividing by the air humidity capacity of the i-th region, and then multiplying by the sum of the moisture transfer coefficients of the i-th region and all other regions. The initial value for the element in the i-th row and j-th column of the humidity-humidity influence matrix is ​​the sampling time interval divided by the air humidity capacity of the i-th region and then multiplied by the moisture transfer coefficients of the i-th and j-th regions.

5. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 4, characterized in that: The linear ordinary differential equations for temperature and humidity are as follows: The linear ordinary differential equation for temperature is as follows: multiply the heat transfer coefficients of the i-th and j-th regions at the current moment by the temperature difference between the j-th and i-th regions at the current moment, and use this as the influence temperature of the j-th region on the i-th region at the current moment. Sum the influence temperatures of all other regions on the i-th region at the current moment, add the set heat generation rate of the internal equipment heat source in the i-th region, add the set external heat exchange parameters, multiply the sum by the sampling time interval and divide by the air heat capacity of the i-th region to get the temperature change of the region in the non-controlled mode of the air conditioning system, and add the temperature change of the region in the non-controlled mode of the air conditioning system to the temperature of the i-th region at the current moment to get the temperature of the i-th region at the next moment. The linear ordinary differential equation for humidity is as follows: Multiply the humidity transfer coefficients of the i-th and j-th regions at the current moment by the difference between the humidity of the j-th region and the i-th region at the current moment, and use this as the humidity of the j-th region on the i-th region at the current moment. Sum the humidity of all other regions on the i-th region at the current moment, add the set internal process humidity generation rate of the i-th region, add the set external airtightness humidity exchange, multiply the sum by the sampling time interval and divide by the air humidity capacity of the i-th region to get the humidity change of the region when the air conditioning system is in non-controlled mode, and add the humidity change of the region when the air conditioning system is in non-controlled mode to the humidity of the i-th region at the current moment to get the humidity of the i-th region at the next moment.

6. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 5, characterized in that: The final values ​​of the temperature-temperature influence matrix, temperature-humidity influence matrix, humidity-temperature influence matrix, and humidity-humidity influence matrix are calculated using the least squares method, specifically as follows: Obtain historical temperature and humidity data for each moment within a set period when the air conditioning system is in non-control mode. The error at time v is calculated by subtracting the product of the state transition matrix and the state vector at time v from the state vector at time v+1 within the set period. The objective function of the least squares method is to calculate the sum of squares at each time point and minimize it. Each time point includes time points 1, 2, 3, ..., N-1; N is the total number of time points within the set period.

7. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 6, characterized in that: The control input matrix is ​​specifically as follows: The control input matrix consists of a temperature control input matrix and a humidity control input matrix arranged vertically. Both the temperature control input matrix and the humidity control input matrix are n×p matrices, where p is the total number of elements in the input vector. The initial value of the element in the i-th row and m-th column of the temperature control input matrix is ​​the temperature control related term of the m-th air conditioning parameter in the i-th region, calculated according to the linear equation of cooling power; the temperature control related term of the m-th air conditioning parameter in the i-th region is the sampling time interval divided by heat capacity, multiplied by air density, multiplied by air specific heat capacity, and multiplied by the air supply volume affected by the m-th air conditioning parameter in the i-th region. The initial value of the element in the i-th row and m-th column of the humidity control input matrix is ​​the humidity control related term of the m-th air conditioning parameter in the i-th region, calculated according to the linear equation of dehumidification power. The humidity control related term is the air supply volume affected by the m-th air conditioning parameter in the i-th region, calculated by dividing the sampling time interval by the humidity capacity, multiplying by the air density, multiplying by the air specific humidity, and multiplying by the air supply volume affected by the m-th air conditioning parameter in the i-th region. The final values ​​of the temperature control input matrix and the humidity control input matrix are calculated using the least squares method.

8. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 7, characterized in that: The linear equation for cooling power is: when only the m-th air conditioning parameter is adjusted, the temperature of the i-th region at the next moment is equal to the temperature of the i-th region at the current moment plus the temperature control related term of the m-th air conditioning parameter in the i-th region multiplied by the cooling capacity of the air conditioning output affected by the adjustment of the corresponding m-th air conditioning parameter. The linear equation for dehumidification power is: when only the m-th air conditioning parameter is adjusted, the humidity of the i-th region at the next moment is equal to the humidity of the i-th region at the current moment plus the humidity control related term of the m-th air conditioning parameter in the i-th region multiplied by the humidity content of the air conditioning output affected by the adjustment of the corresponding m-th air conditioning parameter.

9. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 8, characterized in that: The final values ​​of the temperature control input matrix and humidity control input matrix are calculated using the least squares method, specifically: Obtain historical temperature and humidity data of the air conditioning system at each moment within a set cycle when only the m-th air conditioning parameter is adjusted. When only the m-th air conditioning parameter is adjusted, the state vector of the air conditioning system at time v+1 within the set period is subtracted from the product of the state transition matrix and the state vector at time v, and then the product of the current control input matrix and the input vector is subtracted to obtain the error at time v. The sum of squares at each time point is calculated to minimize the sum of squares as the objective function of the least squares method. Each time point includes the 1st, 2nd, 3rd, ..., N-1th time points; N is the total number of time points within the set period. When performing the least squares method, only the elements in the m-th column of all rows of the temperature control input matrix and the elements in the m-th column of all rows of the humidity control input matrix are adjusted. Repeat all of the above steps for all air conditioning parameters to obtain the final control input matrix.

10. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 9, characterized in that: When the future time window is reached, the actual temperature, humidity, and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold to perform online correction of the air conditioner's operating parameters. Specifically: When the next moment arrives, it is determined whether the correction condition has been met. If it has, the corresponding correction is performed. If it has not been met, the actual temperature, humidity and energy consumption obtained at the next moment are used to determine whether the correction condition has been met, until the correction condition is met or the current moment reaches the total time of the set future time window. If the current moment reaches the total time of the set future time window, the steps S3-S4 are repeated from S3.

11. The method for controlling a constant temperature and humidity air conditioning system based on artificial intelligence according to claim 10, characterized in that: The correction conditions and corresponding corrections are as follows: If the difference between the actual measured temperature and the predicted temperature exceeds the set temperature difference, the correction condition is met, and the supply air temperature setting is reduced according to the set first step length. If the difference between the actual measured humidity and the predicted humidity exceeds the set maximum humidity difference, the correction condition is met and the opening of the dehumidifier valve is increased according to the set second step size. If the difference between the actual measured humidity and the predicted humidity is less than the set minimum humidity difference, the correction condition is met and the opening of the humidifier valve is increased according to the set second step size. If the actual measured energy consumption exceeds the set energy consumption threshold, the supply air temperature setting will be increased or the fan speed of the air conditioning unit will be reduced according to the set third step. If the current time reaches the total time of the set future time window, then repeat steps S3-s4 starting from S3 again.

12. A temperature and humidity control system based on artificial intelligence using the method of any one of claims 1-11, comprising a data acquisition module, a data transmission module, a prediction module, and a correction module, characterized in that: Data Acquisition Module: Collects real-time environmental data through a sensor network deployed in different areas of the workshop and performs preprocessing; the real-time environmental data includes temperature, humidity, and energy consumption. Data transmission module: Compresses and encapsulates the pre-processed real-time environmental data before uploading it to the control center; Prediction Module: This module acquires the operating parameters of all currently running air conditioners. Based on these parameters, it uses a control analysis model at the control center to predict temperature and humidity values ​​for future time windows. The control analysis model uses state transition equations for prediction. The state transition matrix in these equations obtains initial values ​​based on the heat balance and humidity balance equations, and then uses these initial values ​​and the least squares method to obtain the final values. The control input matrix in the state transition equations calculates initial values ​​based on the linear equations of cooling power and dehumidification power, and then uses these initial values ​​and the least squares method to obtain the final values. Correction module: When the future time window is reached, the actual temperature, humidity and energy consumption are compared with the previously predicted temperature and humidity values ​​for the corresponding time and the set energy consumption threshold to perform online correction of the air conditioner's operating parameters.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

Images