An air conditioner intelligent regulation and control method and system

Abstract

This application relates to the field of air conditioning technology and provides an intelligent air conditioning control method, including: real-time collection of outdoor environmental data, indoor environmental data, and personnel status data, and fusion of the data to generate a comprehensive dataset; constructing a dynamic temperature gradient field based on the comprehensive dataset; dividing functional areas and calculating an influence relationship matrix based on the dynamic temperature gradient field; using the influence relationship matrix as the core constraint, employing a multi-objective optimization algorithm to obtain an air conditioning control parameter set; and progressively adjusting the output parameters of each individual air conditioner unit based on the air conditioning control parameter set through a switching mechanism. This application achieves dynamic global optimization of the air conditioning system in multiple objectives such as energy saving, comfort, and equipment lifespan, effectively solving problems such as control lag, excessive energy consumption, uneven comfort, and high equipment wear.

Description

Technical Field

[0001] This application belongs to the field of air conditioning technology, and in particular relates to an intelligent air conditioning control method and system. Background Technology

[0002] In today's office environment, air conditioning systems are key facilities for ensuring indoor comfort and normal equipment operation. Their control methods and energy-saving effects have attracted much attention. They usually undertake the important task of accurately controlling indoor temperature and humidity during the working hours of employees and machines.

[0003] Most existing office air conditioning systems rely on users manually switching on and off and adjusting the temperature. However, individual temperature preferences vary significantly, and different people experience indoor temperature differently, making it difficult to standardize air conditioning temperature settings. Furthermore, different people have varying levels of responsibility for energy conservation; many forget to turn off the air conditioning when leaving the office area, resulting in the phenomenon of "air conditioning running after get off work," causing the air conditioner to run continuously without human intervention and wasting a large amount of electricity. In addition, manually operated air conditioning systems cannot automatically adjust the indoor temperature to a comfortable range when the external temperature and humidity change. This is very detrimental to the health of office workers who spend long hours working at desks without much exercise, and frequent, large-scale adjustments accelerate the wear and tear on the air conditioning equipment.

[0004] Therefore, there is an urgent need for an intelligent air conditioning control method that can automatically adjust based on changes in office space demand, personnel activity, and external environmental factors such as temperature and humidity. Summary of the Invention

[0005] This application provides an intelligent air conditioning control method and system, which can solve one of the above-mentioned problems in the prior art.

[0006] In a first aspect, embodiments of this application provide an intelligent air conditioning control method, including:

[0007] Real-time collection of outdoor environmental data, indoor environmental data, and personnel status data; and fusion of the data to generate a comprehensive dataset.

[0008] Based on the comprehensive dataset, a dynamic temperature gradient field is constructed;

[0009] The functional areas are divided, and the influence relationship matrix is ​​calculated based on the dynamic temperature gradient field.

[0010] Using the aforementioned influence relationship matrix as the core constraint, a multi-objective optimization algorithm is employed to obtain the air conditioning control parameter set;

[0011] Based on the aforementioned air conditioning control parameter set, the output parameters of each individual air conditioning unit are gradually adjusted through a switching mechanism.

[0012] Furthermore, the real-time collection of outdoor environmental data, indoor environmental data, and personnel status data, and the fusion of these data to generate a comprehensive dataset, includes:

[0013] It integrates outdoor meteorological sensors, indoor environmental monitoring sensors, and personnel status detection sensors to collect data from multiple dimensions.

[0014] The multidimensional data sources are preprocessed and spatiotemporally aligned, and then fused to generate a comprehensive dataset.

[0015] Furthermore, the spatiotemporal alignment includes dynamic timestamp calibration and sampling frequency adjustment;

[0016] The dynamic timestamp calibration specifically includes: for sensors that can be synchronized by command, the time when the acquisition command is received is used as the data calibration timestamp; for sensors that cannot be synchronized by command, the data calibration timestamp is calculated using a time drift estimation model.

[0017] The sampling frequency adjustment specifically includes: using a signal resampling method to unify all data sequences to a predetermined target sampling rate.

[0018] Furthermore, constructing a dynamic temperature gradient field based on the comprehensive dataset includes:

[0019] Obtain the real-time operating parameters and physical location of all individual air conditioning units, construct a temperature diffusion model for each individual air conditioning unit, calculate the temperature distribution of each individual air conditioning unit, and generate a sub-gradient field;

[0020] The field superposition algorithm is used to vector superimpose the sub-gradient fields of all individual air conditioners with known heat sources to generate a composite temperature difference gradient field.

[0021] The predicted environmental values ​​of the composite temperature gradient field are fused with the collected data values ​​of the comprehensive dataset to dynamically adjust the composite temperature gradient field.

[0022] Furthermore, the step of fusing the predicted environmental values ​​of the composite temperature gradient field with the collected data values ​​of the comprehensive dataset to dynamically adjust the composite temperature gradient field includes:

[0023] The indoor environmental data is compared with the predicted environmental values ​​by using a data fusion algorithm, and the parameters of the temperature diffusion model are dynamically adjusted to calibrate the composite temperature gradient field.

[0024] When calibrating the composite temperature gradient field, dynamic environmental data is used as a boundary condition to drive the refresh of the composite temperature gradient field, wherein the dynamic environmental data includes the opening and closing status of doors and windows, personnel status data, and outdoor environmental data.

[0025] Furthermore, the division of functional regions, combined with the dynamic temperature gradient field, and the calculation of the influence relationship matrix include:

[0026] Based on the dynamic temperature gradient field, personnel status data, and indoor space layout, multiple functional areas are dynamically divided, each of which is assigned a unique identifier and a preset temperature and humidity target value.

[0027] Based on the dynamic temperature gradient field, the thermodynamic influence relationship between the functional regions is analyzed, and the temperature gradient intensity and influence direction factor are determined.

[0028] Based on the indoor space layout, the airflow paths between the functional areas are determined, and a path smoothness coefficient is generated.

[0029] The matrix elements of the influence relationship matrix are calculated by combining the temperature gradient intensity, the influence direction factor, and the path accessibility coefficient.

[0030] Furthermore, the step of determining the airflow paths between the functional areas based on the indoor spatial layout and generating a path smoothness coefficient includes:

[0031] Based on the dynamic temperature gradient field, an airflow path from one functional area to another is determined, and the airflow path is a virtual pipe connecting the key points of the two functional areas.

[0032] Calculate the projected area of ​​the virtual pipe in the indoor space, generate a local projected area, and based on the indoor space layout, determine and calculate the sum of the projected areas of each obstruction in the virtual pipe, and generate an obstruction projected area.

[0033] The path accessibility coefficient is calculated based on the local projected area and the obstructed projected area.

[0034] Furthermore, the process of obtaining the air conditioning control parameter set using the aforementioned influence relationship matrix as the core constraint and a multi-objective optimization algorithm includes:

[0035] A multi-objective cost function is constructed with the goals of minimizing total air conditioning energy consumption, maximizing equipment operational stability, and maximizing collective comfort.

[0036] A distributed model predictive control algorithm is used to simulate the evolution of indoor state under different combinations of air conditioning parameters within a preset time period. The algorithm is then iteratively solved using the multi-objective cost function to output a set of air conditioning control parameters constrained by the influence relationship matrix.

[0037] Furthermore, the step of progressively adjusting the output parameters of each individual air conditioner unit based on the air conditioning control parameter set through a switching mechanism includes:

[0038] Obtain the current operating parameter set of each air conditioner unit, and calculate the absolute difference of each parameter of each air conditioner unit in combination with the air conditioner control parameter set;

[0039] Based on the absolute difference, a transition path is generated using distributed linear interpolation. The transition path is a sequence of parameter outputs from the current time to the intermediate time point between the current time and the preset time.

[0040] Based on the parameter output sequence, the output parameters of each air conditioning unit are gradually adjusted;

[0041] During the adjustment process, environmental parameters and the operating status of each air conditioner are monitored in real time, and the transition path is dynamically updated based on the monitoring results.

[0042] Secondly, embodiments of this application provide an intelligent air conditioning control system, comprising:

[0043] The first processing module is used to collect outdoor environmental data, indoor environmental data, and personnel status data in real time, and to merge the data to generate a comprehensive dataset.

[0044] The second processing module is used to construct a dynamic temperature gradient field based on the comprehensive dataset.

[0045] The third processing module is used to divide functional areas and calculate the influence relationship matrix in conjunction with the dynamic temperature gradient field.

[0046] The fourth processing module is used to obtain the air conditioning control parameter set by employing a multi-objective optimization algorithm with the aforementioned influence relationship matrix as the core constraint.

[0047] The fifth processing module is used to gradually adjust the output parameters of each individual air conditioner unit based on the set of air conditioner control parameters through a switching mechanism.

[0048] The beneficial effects of the embodiments in this application compared with the prior art are:

[0049] This application discloses an intelligent air conditioning control method that synchronously collects and fuses multi-source heterogeneous data, including outdoor environmental data, indoor environmental data, and personnel status data. This solves the problems of sensor clock asynchrony and data format heterogeneity, providing a data foundation for subsequent data analysis. Simultaneously, the construction of a dynamic temperature gradient field achieves a leap from single-point perception to full-field three-dimensional perception, providing a scientific basis for the dynamic division of functional areas and the calculation of the influence relationship matrix. The dynamic division of functional areas refines the granularity of air conditioning control, adapting to environmental changes and providing targets for personalized control. The influence relationship matrix quantifies the thermal influence of different zones, ensuring that multi-objective optimization is performed within the scope of physical laws. Furthermore, comfort modeling transforms subjective comfort into objective indicators, and the generated collective comfort index makes the optimization process more human-centered. Under physical constraints, the multi-objective optimization maximizes the benefits of each objective, achieving dynamic global optimization of the air conditioning system across multiple objectives such as energy saving, comfort, and equipment lifespan. This effectively solves problems such as control lag, excessive energy consumption, uneven comfort, and high equipment wear. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating an intelligent air conditioning control method according to an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the structure of an intelligent air conditioning control system according to an embodiment of the present invention;

[0053] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0054] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0055] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0056] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0057] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0058] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0059] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0060] Please see Figure 1 As shown, the present invention is an intelligent air conditioning control method, comprising the following steps:

[0061] S100: Real-time collection of outdoor environmental data, indoor environmental data, and personnel status data; and fusion of the data to generate a comprehensive dataset.

[0062] In this application, multi-source heterogeneous data, including outdoor environmental data, indoor environmental data, and personnel status data, are collected and fused synchronously, which solves the problems of sensor clock asynchrony and data format heterogeneity, and provides a data foundation for subsequent data analysis.

[0063] In some embodiments, step S100 above includes:

[0064] It integrates outdoor meteorological sensors, indoor environmental monitoring sensors, and personnel status detection sensors to collect data from multiple dimensions.

[0065] The multidimensional data sources are preprocessed and spatiotemporally aligned, and then fused to generate a comprehensive dataset.

[0066] In this embodiment, outdoor meteorological sensors are used to collect outdoor environmental data, specifically including outdoor temperature, outdoor humidity, wind speed, etc., which can be collected through outdoor weather stations, etc. Indoor environmental monitoring sensors are used to collect indoor environmental data, specifically including temperature and humidity sensors, CO2 sensors, door and window sensors, etc. Among them, temperature and humidity sensors are used to collect indoor temperature and indoor humidity, CO2 sensors are used to collect indoor carbon dioxide concentration data to measure indoor air quality, and door and window sensors are used to collect whether office doors and windows are closed. Personnel status detection sensors include human body micro-motion sensors and smart cameras, etc. Among them, human body micro-motion sensors are sensors that collect the presence status of human bodies and can obtain in real time whether there is human activity in a specific area, and smart cameras are used to collect the number of people active in a specific area in real time.

[0067] In some embodiments, since different sensors come from different manufacturers and their data formats and transmission methods may differ, various types of sensors are connected to a unified server, and a standardized interface protocol is established to ensure that different types of sensors can seamlessly interface and interact with the unified server. Specifically, the unified server connects various types of sensors through multiple adapter interfaces, such as RS-485, Modbus, Zigbee, LoRa, MQTT, etc., to generate a heterogeneous sensor network. By developing or configuring corresponding driver adapters for each type of sensor, the proprietary protocols and data formats of different manufacturers are converted into internally unified standardized data objects, which serve as the basis for subsequent data processing.

[0068] In this embodiment, a unified server sends acquisition scheduling instructions to each sensor. Within one acquisition cycle, the acquisition data from each sensor is sent to the unified server. During this process, the sampling rates of different sensors may be different, and the sensor clocks may be out of sync, which may lead to data alignment deviations. Therefore, it is necessary to perform spatiotemporal alignment on the data. At the same time, it is also necessary to perform preprocessing operations such as data cleaning and filtering on the acquired data to reduce the impact of data noise and sudden interference and improve the reliability of the data.

[0069] In some embodiments, the spatiotemporal alignment includes dynamic timestamp calibration and sampling frequency adjustment;

[0070] The dynamic timestamp calibration specifically includes: for sensors that can be synchronized by command, the time when the acquisition command is received is used as the data calibration timestamp; for sensors that cannot be synchronized by command, the data calibration timestamp is calculated using a time drift estimation model.

[0071] The sampling frequency adjustment specifically includes: using a signal resampling method to unify all data sequences to a predetermined target sampling rate.

[0072] In this embodiment, all sensors collect data uniformly based on the acquisition instructions of a unified server. However, in practical applications, some sensors cannot directly respond to the acquisition instructions, there is a delay in data transmission, and the local clock of the sensor may be offset from the main clock of the server. Therefore, it is necessary to estimate the time offset through a specific model to calibrate the timestamp.

[0073] Specifically, for sensors that can be synchronized by commands, they can receive and respond to acquisition commands, so the time of command reception is used as the data calibration timestamp. For sensors that cannot be synchronized by commands, it is assumed that their data arrives at the unified server after a transmission delay of Δt. At this time, the unified server records the arrival timestamp T_arrive and estimates the offset Δoffset between the sensor's local clock and the server's master clock using a time drift estimation model. Specifically, a high-precision and high-stability clock source is selected as a reference, such as GPS timing, to ensure its time accuracy and stability. Then, a comparison period is selected, and every 24 hours, the data is recorded at each comparison time t. i (i=1,2,3,⋯) The sensor records the current time T of its local clock. sensor,i Meanwhile, the unified server records the current time T of its master clock by synchronizing with the reference clock source. server,i For each comparison time t i Calculate the time difference ΔT between the sensor's local clock and the server's master clock. i =T sensor,i -T server,i Furthermore, the drift rate is calculated through linear fitting: ΔT(t) = k × t + b, where k is the clock drift rate and b is the initial offset. The least squares method is then used to analyze the time difference data (ΔT) at multiple comparison points. i ,t i A linear fit is performed to obtain estimates of the clock drift rate k and b, assuming the local time when the sensor acquires data is T. sensor The server master clock time is T. server Then T sensor -T server =k×t delay+b, since the sensor data arrives at the unified server after a transmission delay of Δt, and the arrival timestamp is T_arrive, then T_arrive-T sensor =Δt, further, according to the time drift estimation model T server =T_arrive - Δt - Δoffset, and thus, by converting the time relationship during data transmission, we can obtain Δoffset = k × t. total +b, where t total The total time interval from data acquisition by the sensor to data reception by the server and completion of clock comparison estimation is given by the calibration timestamp of the data point = T_arrive - Δt + Δoffset.

[0074] In this embodiment, the sampling rates of different sensors may be different. Therefore, a signal resampling method is used to unify the data of all sensors to the specified target sampling rate, thereby ensuring that the data of all sensors are mapped at each time point. When the original sampling rate is lower than the target sampling rate, an upsampling + interpolation method is used to insert zero values ​​between adjacent data points in the original data sequence to increase the number of data points. The data point value at the zero-value insertion position is estimated by linear interpolation. In some embodiments, the interpolated data point value can also be estimated by polynomial interpolation and spline interpolation. When the original sampling rate is higher than the target sampling rate, a downsampling + decimation method is used to select data points from the filtered data sequence according to certain rules to reduce the number of data points.

[0075] In this embodiment, after the sensor data is aligned in time and space as described above, each calibration timestamp t corresponds to a feature vector, representing the data of each sensor at that calibration timestamp, as shown below: V(t)=[T_out(t),H_out(t),...,T_in_1(t),H_in_1(t),CO2(t),Occupancy(t),...]. Then, the feature vectors of all time points are arranged in chronological order to generate a comprehensive dataset with spatiotemporal alignment.

[0076] S200. Based on the comprehensive dataset, construct a dynamic temperature gradient field;

[0077] In the prior art, multiple air conditioning units are usually arranged in an indoor office environment to ensure temperature control in each area. This application abstracts the indoor office environment into an optimizable dynamic temperature gradient field, and regulates the indoor environment of each air conditioning unit around this dynamic temperature gradient field.

[0078] In this application, the construction of dynamic temperature gradient field realizes a leap from single-point perception to full-field three-dimensional perception, providing a scientific basis for the dynamic division of functional areas and the calculation of influence relationship matrix. The dynamic division of functional areas enables the air conditioning control granularity to be refined, adapting to environmental changes and providing targets for personalized control. The influence relationship matrix quantifies the thermal influence of the intervals, ensuring that the multi-objective optimization solution is carried out within the scope of physical laws.

[0079] In some embodiments, step S200 above includes:

[0080] Obtain the real-time operating parameters and physical location of all individual air conditioning units, construct a temperature diffusion model for each individual air conditioning unit, calculate the temperature distribution of each individual air conditioning unit, and generate a sub-gradient field;

[0081] The field superposition algorithm is used to vector superimpose the sub-gradient fields of all individual air conditioners with known heat sources to generate a composite temperature difference gradient field.

[0082] The predicted environmental values ​​of the composite temperature gradient field are fused with the collected data values ​​of the comprehensive dataset to dynamically adjust the composite temperature gradient field.

[0083] In this embodiment, the real-time operating parameters of all individual air conditioners are obtained, specifically the set temperature, wind speed, wind direction, and operating mode. Combined with the indoor space layout, the physical location coordinates of each individual air conditioner in the room are determined. Thus, a temperature diffusion model is constructed for each individual air conditioner. This temperature diffusion model treats the individual air conditioner as a point source and calculates the theoretical temperature value and air velocity direction of each point within its airflow organization coverage area based on its real-time operating parameters and physical location coordinates, thereby generating a sub-gradient field describing the independent effect of the air conditioner.

[0084] Specifically, for the temperature difference ΔT between a point (x, y, z) in space and the air outlet (x0, y0, z0) of any individual air conditioner unit, the function f(U jet ,T jet ,T room The calculation is performed using geometry, where the function f is obtained through fitting experimental data or theoretical derivation, and is used to predict the temperature change ΔT at a point (x,y,z) in space due to the effect of the air conditioning jet, based on the above input parameters, while U... jet T represents the jet velocity at the air conditioner vent, that is, the speed at which air is ejected from the vent, which affects the momentum and diffusion capacity of the jet. jet This indicates the outlet air temperature of the air conditioner, i.e., the temperature of the air supplied by the air conditioner, which affects the local temperature field. (T) roomThe indoor ambient temperature represents the temperature when the air conditioner is off or unaffected by it. "Geometry" represents geometric parameters, including the shape, size, and direction of the air outlet. Specifically, the above formula calculates the temperature difference at the air outlet of a single air conditioner unit using a multivariate function f. It couples the jet velocity, outlet temperature, indoor ambient temperature, and fine geometric parameters as independent variables. Compared to existing models that only consider distance attenuation or simple linear interpolation, this significantly improves the physical fidelity and accuracy of predicting complex jet temperature fields.

[0085] In one possible embodiment, ΔT=f(U jet ,T jet ,T room ,geometry)=(T jet -T room The formula is calculated as: )×g(r / r0,z / d0,Re, other geometric parameters), where r represents the radial distance from a point (x,y,z) in space to the jet axis, z represents the axial distance from a point (x,y,z) in space to the air conditioning unit, the jet axis represents the direction and center position of the air conditioning unit's flow, specifically the center coordinates of the air outlet of the air conditioning unit, r0 represents the characteristic radius of the air outlet, and d0 represents the characteristic diameter of the air outlet. r / r0 and z / d0 are used to normalize jet problems of different sizes and velocities. Re represents the flow factor, used to characterize the flow state, which affects the turbulence characteristics and diffusion capacity of the jet. It is worth noting that this embodiment introduces dimensionless parameters r / r0, z / d0, and the flow factor Re in the formula, enabling the final model to deeply reflect the physical essence of jet motion. It also normalizes jet problems of different sizes and velocities under the same theoretical framework, thereby improving scientific rigor and significantly reducing computational complexity while maintaining accuracy.

[0086] Furthermore, the coverage area of ​​the airflow organization of the individual air conditioning unit is calculated, that is, the farthest distance that its jet can affect. This is specifically estimated based on the air outlet velocity and the area of ​​the air vent. ,in, T represents the maximum distance of the jet, k represents an empirical coefficient that depends on the jet type and spatial conditions. jet T represents the air outlet temperature of the air conditioner. room Indicates the indoor ambient background temperature, U jet This represents the jet velocity at the air conditioner's outlet, and A0 represents the outlet area of ​​the individual air conditioner unit. Specifically, the higher the outlet velocity and the larger the outlet area, the farther the jet can reach. Specifically, the above formula will use the temperature difference... The formula combines jet velocity and air outlet area to comprehensively estimate the maximum influence distance of the jet. Compared with the existing technology that simply specifies a fixed range based on experience, this formula can more accurately and dynamically define the jet "sphere of influence" of each air conditioner.

[0087] Furthermore, regarding the speed of the individual air conditioning units, ,in U represents the quantization of velocity decay with distance z. jet Let d0 represent the jet velocity at the air conditioner outlet, d0 represent the characteristic diameter of the outlet, and n represent the attenuation coefficient. The velocity direction of an individual air conditioner unit at a point (x, y, z) in space is calculated using the jet axis direction and the relative position of that point, thus obtaining the velocity vector of the individual air conditioner unit. This vector is then combined with the temperature difference ΔT to form a sub-gradient field. For each point in the sub-gradient field, T = T0... room +ΔT yields the actual temperature at that point, while the direction and rate of temperature change at that point are obtained by calculating the temperature difference between surrounding points. The temperature gradient reflects the trend of temperature change in space, while the airflow velocity vector, including velocity magnitude and direction, describes the motion state of the airflow at that point. Specifically, the velocity magnitude formula U(z) explicitly describes the attenuation law of the jet axis velocity with axial distance. Combined with the temperature difference ΔT, it forms a unified "sub-gradient field" containing both temperature scalar and velocity vector, thus describing the dual impact of a single air conditioner on the space thermal and dynamic environments.

[0088] Specifically, the sub-gradient field clearly describes the thermal and airflow effects of the individual air conditioner on each point in the surrounding space under independent operation, providing a high-quality data foundation for subsequent full-field superposition and fusion. At the same time, by superimposing and fusing the sub-gradient fields of multiple air conditioners, the overall temperature distribution and airflow of the indoor space can be simulated when multiple air conditioners are running simultaneously.

[0089] In this embodiment, a field superposition algorithm is used to vector superimpose the sub-gradient fields generated by all air conditioners. At the same time, the thermal radiation fields generated by known fixed heat sources in the space, such as server racks, based on their rated power and location are also superimposed. This generates a composite temperature gradient field for the entire space. The composite temperature gradient field includes the theoretical temperature value, temperature change gradient, and preliminary estimate of airflow velocity at any point in the space.

[0090] Specifically, considering a fixed heat source, the temperature distribution of the thermal radiation field generated at any point r=(x,y,z) in space can be obtained by integrating... The temperature contribution of each volumetric element dV' is calculated. Specifically, each volumetric element dV' is considered as a point source, and its contribution decays with distance. This represents the attenuation coefficient related to room ventilation conditions, while the exponential term... Reflecting the suppression of heat diffusion by room ventilation, r'=(x',y',z') is the position vector of a certain volume element dV' within the heat source volume, through... The distance q(r') is measured from a volumetric element dV′ within a heat source to any point r=(x,y,z) in space. q(r') represents the power density of a fixed heat source, and its specific distribution is determined through numerical integration or experiments in practical applications. Indicates the thermal conductivity of air. This represents the ambient temperature at point r=(x,y,z) after the sub-gradient field vectors are superimposed. It is worth noting that the above formula treats a fixed heat source as a volumetric heat source with a continuous spatial power density distribution q(r'), and calculates its thermal radiation field throughout the space through volume integrals. This is fundamentally different from the existing technology that simplifies the heat source into a point source or a uniform surface source. This method can quantify the complex and asymmetric three-dimensional temperature field impact of non-uniform heat sources, such as those only generating heat at the rear of a server, or large extended heat sources, such as continuous production lines, on their surrounding space, thus solving the problem of excessive error. Simultaneously, the exponential term... The introduction of this method quantifies the inhibitory effect of room ventilation conditions on heat diffusion. When ventilation is good, hot air is quickly carried away, and the heat becomes more localized; when ventilation is poor, heat is more likely to accumulate in the space and spread further, thus improving the accuracy of prediction.

[0091] In some embodiments, fusing the predicted environmental values ​​of the composite temperature gradient field with the collected data values ​​of the integrated dataset to dynamically adjust the composite temperature gradient field includes:

[0092] The indoor environmental data is compared with the predicted environmental values ​​by using a data fusion algorithm, and the parameters of the temperature diffusion model are dynamically adjusted to calibrate the composite temperature gradient field.

[0093] When calibrating the composite temperature gradient field, dynamic environmental data is used as a boundary condition to drive the refresh of the composite temperature gradient field, wherein the dynamic environmental data includes the opening and closing status of doors and windows, personnel status data, and outdoor environmental data.

[0094] In this embodiment, indoor environmental data is compared with the predicted environmental values ​​of the composite temperature gradient field at the corresponding locations. Specifically, by calculating the Kalman gain, key parameters in the temperature diffusion model, such as the flow factor Re or the empirical coefficient k, are dynamically adjusted so that the output prediction value of the gradient field model continuously converges to the actual measurement value. This process effectively solves the problem of "computational instability" caused by model simplification and nonlinear propagation.

[0095] Furthermore, when calibrating the composite temperature gradient field, the opening and closing status of doors and windows, the location and number of people, and outdoor meteorological data are injected into the temperature diffusion model as boundary conditions, driving the sub-gradient field to be dynamically updated according to environmental conditions, thereby refreshing the composite temperature gradient field.

[0096] In some embodiments, an adaptive hierarchical mesh technique is used to discretize the indoor space and analyze the current gradient field state in real time. Furthermore, in areas with drastic temperature gradient changes and complex heat sources, such as directly in front of air conditioning vents, densely populated work areas, and around server racks, the computational mesh is automatically densified to capture more refined temperature changes. In areas with uniform temperature and no significant thermal disturbance, such as open corridors and deserted corners, a sparse mesh is used to save computational resources.

[0097] S300. Divide the functional areas and calculate the influence relationship matrix based on the dynamic temperature gradient field.

[0098] In some embodiments, step S300 above includes:

[0099] Based on the dynamic temperature gradient field, personnel status data, and indoor space layout, multiple functional areas are dynamically divided, each of which is assigned a unique identifier and a preset temperature and humidity target value.

[0100] Based on the dynamic temperature gradient field, the thermodynamic influence relationship between the functional regions is analyzed, and the temperature gradient intensity and influence direction factor are determined.

[0101] Based on the indoor space layout, the airflow paths between the functional areas are determined, and a path smoothness coefficient is generated.

[0102] The matrix elements of the influence relationship matrix are calculated by combining the temperature gradient intensity, the influence direction factor, and the path accessibility coefficient.

[0103] In this embodiment, the space is dynamically divided into virtual functional zones based on the isotherm distribution of the dynamic temperature gradient field, personnel status data, and the indoor space layout. Specifically, based on the dynamic temperature gradient field, the space is identified as a "plateau zone" with sparse isotherms and relatively uniform temperature, and a "gradient zone" with dense isotherms and drastic temperature changes. At the same time, a heat map of personnel activity is generated through personnel status data. Combined with the indoor space layout, a clustering algorithm is used to output multiple functional areas. Each functional area is recalculated and adjusted according to changes in personnel and heat exchange.

[0104] Specifically, the dynamic temperature gradient field is converted into a three-dimensional mesh vector field. Each mesh cell contains a temperature value, a direction vector of the temperature gradient, and an amplitude value. Its coordinate system is aligned with the physical space in the indoor spatial layout. The indoor spatial layout is a three-dimensional model that includes geometric information such as walls and columns, as well as material properties used to determine the degree of airflow obstruction and functional semantic information labeled with "workstation," "meeting room," etc. At the same time, a human activity heat map is generated by human micro-motion sensors and smart cameras deployed in the space. This heat map is a two-dimensional probability distribution map; the higher the value, the denser the population in that area. By mapping the aforementioned dynamic temperature gradient field, personnel heat map, and indoor spatial layout data to the same world coordinate system, it is ensured that each grid cell has corresponding temperature, gradient, personnel density, and spatial attribute values. Furthermore, for each three-dimensional grid cell in space, environmental uniformity characteristics, personnel density characteristics, and spatial connectivity characteristics are calculated. For environmental uniformity characteristics, the variance of temperature within a certain neighborhood around the cell, such as a 3x3x3 grid, is calculated. Regions with small variance are designated as "plateau areas," and regions with large variance are designated as "gradient areas." For personnel density characteristics, the personnel density corresponding to that cell is directly taken. Density values ​​are used as features. Simultaneously, the total density of the surrounding area can be calculated to characterize influence. For spatial connectivity features, the field of view and airflow accessibility of the spatial region where the cell is located are obtained, i.e., the path accessibility coefficient. After normalizing the above features, they are concatenated into a unified multi-dimensional feature vector representing the attributes of the grid cell. Furthermore, an unsupervised machine learning clustering algorithm is used to group the feature vectors of all grid cells in the space. In a preferred embodiment, the DBSCAN algorithm is used to group the grid cells. DBSCAN does not require pre-specifying the number of partitions and can automatically discover grid cells of arbitrary shapes. Clustering can separate noise points in uninhabited and abnormally temperatured corners, improving the adaptability of grouping. During the clustering process, the parameter settings are adaptively set according to the overall size of the space and the grid resolution. After clustering is completed, each set of assigned grid cells forms a functional partition. It is worth noting that the boundaries of each functional partition may be tortuous and irregular, generated entirely by data, rather than simple rectangles. In addition, each functional partition obtained by clustering is assigned a globally unique identifier, and a functional type label and a corresponding preset temperature and humidity target value are set according to the characteristics of the partition.

[0105] In one embodiment, if the average value of the population density feature in a partition is very high and the functional semantic information from the indoor space layout includes "meeting room", then it is labeled as "meeting area" and given a lower set temperature target, such as 24±0.5°C. If a partition has good connectivity features and low population density, it may be labeled as "public circulation area" and the set temperature may be higher, such as 27±1°C.

[0106] Furthermore, for the functional zones within the space, an N-order square matrix A is defined, where the matrix element A_ij represents the degree of influence of a unit temperature change in zone j on the temperature of zone i. Specifically, the thermal influence relationship between functional zones is quantified through a dynamic temperature gradient field, and the airflow influence relationship between functional zones is determined based on the indoor spatial layout. Specifically, A_ij = f(G, D, P), where G represents the intensity of the temperature gradient from j to i, and D represents the influence direction factor. In cooling mode, if the gradient direction is from j to i, i.e., the temperature of zone j is lower than that of zone i, it means that cold air can flow from j to i, having a cooling effect on i. In this case, A_ij is... Negative values ​​indicate the influence of the gradient vector; conversely, positive values ​​indicate the influence of the gradient vector. Specifically, the sign of the influence direction factor is determined by the direction of the gradient vector. P represents the path accessibility coefficient, which is between 0 and 1, where 0 represents complete blockage and 1 represents complete accessibility. f is a function used to calculate A_ij. In one embodiment, it is a linear weighted combination; in other embodiments, it is a nonlinear model trained by machine learning. It is worth noting that A_ij calculated by function f is a dynamically changing real number with positive and negative values. The absolute value of A_ij is positively correlated with the product of G and P, and is used to represent the strength of the influence. The positive or negative sign ** indicates the type of influence, i.e., cooling or heating.

[0107] In some embodiments, determining the airflow paths between the functional areas based on the indoor space layout and generating a path smoothness coefficient includes:

[0108] Based on the dynamic temperature gradient field, an airflow path from one functional area to another is determined, and the airflow path is a virtual pipe connecting the key points of the two functional areas.

[0109] Calculate the projected area of ​​the virtual pipe in the indoor space, generate a local projected area, and based on the indoor space layout, determine and calculate the sum of the projected areas of each obstruction in the virtual pipe, and generate an obstruction projected area.

[0110] The path accessibility coefficient is calculated based on the local projected area and the obstructed projected area.

[0111] In this embodiment, the interior space layout includes geometric information such as walls and columns, as well as material properties used to determine the degree of obstruction to airflow. These are incorporated to calculate the path unobstructedness coefficient between functional areas.

[0112] In this embodiment, due to the complex structure in the space, the air flow path between the two functional areas is simplified into a virtual pipeline connecting the key points of the two functional areas. The key points can be the centers or representative points of the functional areas, which helps subsequent calculations and analyses.

[0113] Specifically, in one embodiment, the virtual pipeline is defined as a three-dimensional cylinder with a circular cross-section. Its axis is a straight line or a fitted curve connecting two points. Its cross-sectional area is the projected area of the virtual pipeline in the indoor space, and its specific value is dynamically determined according to the air flow intensity. Based on the outlet air speed and air volume of each air conditioner unit, the effective diffusion radius of the air flow on this path is estimated according to the jet theory, and then S_path is calculated. Specifically, the larger the air volume, the larger the S_path value. Further, an intersection operation is performed between the virtual pipeline and the model of the indoor space layout to identify all obstacle entities intersecting with the virtual pipeline, such as walls, columns, furniture, cabinets, etc. For each identified obstacle, its accurate three-dimensional geometric shape and material properties are extracted from the model of the indoor space layout for calculating the accurate projected area and the air flow permeability coefficient of the material. Specifically, in one embodiment, there is a pre-set material permeability coefficient comparison table obtained through experimental tests. This table defines the air flow permeability coefficient C_material of each material, and its value ranges from 0 to 1. 0 means the material completely blocks air, such as a solid concrete wall, thick glass, etc.; 0 < C_material < 1 means the material partially penetrates, such as a perforated brick, a wooden partition with certain gaps, etc.; and C_material = 1 means the material has no impact on the air flow, and it only serves as a virtual boundary.

[0114] Further, for each obstacle, calculate its projected area in the direction perpendicular to the pipeline axis to generate the initial obstacle projected area S_project. Here, the pipeline axis refers to the main direction of the air flow in the virtual pipeline, and this direction is represented by a directed vector d, usually pointing from the key point of the source partition to the key point of the target partition. Then, the "direction perpendicular to the pipeline axis" refers to all directions perpendicular to the vector d. By calculating the projections of the projected areas of each obstacle in the direction perpendicular to the pipeline axis, the blocking degree of the obstacle to the air flow when the air flow is moving is measured. In this calculation process, the material properties of each obstacle are introduced to calculate the effective blocking area S_effective of the obstacle. Its calculation formula is: S_effective = S_project × (1 - C_material). Further still, in the virtual pipeline, sum up the effective blocking areas of all obstacles intersecting with the pipeline axis to calculate the obtained blocking projected area S_block_effective.

[0115] Furthermore, the path smoothness coefficient is specifically calculated as P=max(0,1-(S_block_effective / S_path)). It can be understood that the path smoothness coefficient is a continuous value between 0 and 1, which accurately quantifies the smoothness of the airflow path from zone j to zone i. This value will be used as a core parameter for subsequent calculation of the element A_ij in the influence relationship matrix.

[0116] S400. Using the aforementioned influence relationship matrix as the core constraint, a multi-objective optimization algorithm is employed to obtain the air conditioning control parameter set.

[0117] In this application, the comfort modeling transforms subjective comfort into objective indicators, and the generated collective comfort indicators make the optimization process more humane. Under physical constraints, the multi-objective optimization solution maximizes the benefits of each objective, realizing the dynamic global optimum of the air conditioning system in multiple objectives such as energy saving, comfort, and equipment lifespan. This effectively solves problems such as control lag, excessive energy consumption, uneven comfort, and large equipment wear.

[0118] In some embodiments, step S400 above includes:

[0119] A multi-objective cost function is constructed with the goals of minimizing total air conditioning energy consumption, maximizing equipment operational stability, and maximizing collective comfort.

[0120] A distributed model predictive control algorithm is used to simulate the evolution of indoor state under different combinations of air conditioning parameters within a preset time period. The algorithm is then iteratively solved using the multi-objective cost function to output a set of air conditioning control parameters constrained by the influence relationship matrix.

[0121] In this embodiment, a multi-objective cost function is constructed and solved using a distributed optimization algorithm. The final result is a globally optimal set of air conditioning control parameters that achieves multiple objectives such as energy saving, comfort, and equipment lifespan. Specifically, J = w_energy × J_energy + w_comfort × J_comfort + w_device × J_device, where J_energy represents the total energy consumption index, specifically the sum of the power consumption of all individual air conditioners within a preset time; J_comfort represents the collective comfort index, dynamically calculated through the output of the collective comfort model; J_device represents the equipment operational stability index, typically measured through air conditioning settings such as temperature and fan speed variations or the equipment's start-stop frequency; and w_energy, w_comfort, and w_device are weighting coefficients used to adjust the relative importance of different indicators.

[0122] In this embodiment, during the iterative solution of the multi-objective cost function, the influence relationship matrix is ​​used as a constraint. Specifically, the thermal coupling relationship represented by the influence relationship matrix A is used as an equality or inequality constraint for the optimization problem. For example, it can be expressed as: T_i(t+1)=A*T_set_j(t)+..., that is, the temperature of zone i at the next moment is affected by the set temperatures of all zones j at the current moment. In this way, the optimization process of solving the air conditioning control parameter set of the multi-objective cost function is restricted to this framework, making the control of the air conditioning system more reasonable and effective, thereby improving the accuracy and effectiveness of the air conditioning system control.

[0123] In this embodiment, the distributed model predictive control algorithm decomposes the above optimization problem into several smaller sub-problems, such as one sub-problem for each functional area or each air conditioner. The global optimal solution is approximated through iterative negotiation between sub-problems, thereby reducing computational complexity. Furthermore, the distributed model predictive control algorithm operates within a preset control cycle. In each control cycle, the distributed model predictive control algorithm solves for a series of control commands in the future time domain based on the current total energy consumption index, equipment operation stability index, and collective comfort index, i.e., air conditioner setting parameters. In the next cycle, the status of each index is updated, and the algorithm re-predicts and optimizes, thereby realizing iterative algorithm operation to continuously adapt to environmental changes and possess strong anti-interference capabilities. Finally, it outputs the globally optimal control parameter set with the minimum comprehensive cost function J, i.e., the optimized sequence of parameters such as set temperature, wind speed, and wind direction for each air conditioner in the future time period.

[0124] In this embodiment, for the collective comfort index, the user's subjective feedback is correlated with the objective environmental state to construct a collective comfort model and determine the corresponding collective comfort index. Specifically, a unique user identifier ID is assigned to each person in each space to distinguish the data of different users. That is, each person entering the office is assigned a user identifier. At this time, the identification and labeling of personnel are carried out through smart cameras, etc. In addition, when the user gives feedback, the environmental parameters of the area where the user is located are collected, including air temperature, relative humidity, and air velocity, and a state vector S_t is formed. Here, the user feedback is the active adjustment behavior of the user through a mobile APP, smart panel or voice command, such as "increase the temperature by 1°C". Each feedback behavior is quantified into an action on the environment, such as (Action, A_t). Further, the user's body surface temperature is estimated by an infrared thermal imaging sensor, and the user's facial expression and behavioral posture are analyzed by the camera, such as whether the user appears irritable, sweating, or curling up, taking off a coat, etc., and these are converted into a reward value (Reward, R_t).

[0125] In this embodiment, the above data will be stored in the form of time-series data, with each record being (User ID, Timestamp, State S_t, Action A_t / Reward R_t), forming a comfort experience dataset. Furthermore, a reinforcement learning framework is adopted, treating each user as an independent learning environment. Specifically, the State is S_t, representing the environmental state vector in which the user is located. The Action is the user's suggestion to adjust environmental parameters. For example, in one embodiment, A_t is a suggestion to increase the temperature by 0.5°C generated by a voice command. The Reward is R_t, representing the user's immediate comfort feedback in this state.

[0126] Furthermore, using the aforementioned comfort experience dataset, an initial training process is performed on each user's personal comfort model. Specifically, ,in, γ represents the learning rate, and γ represents the discount factor. This represents the old value estimate of the state-action pair (S_t, A_t) before this update. In other words, it's the estimated long-term reward of performing action A_t in state S_t before updating Q with the existing knowledge. This represents the new value estimate for the state-action pair (S_t, A_t) after this update. Specifically, the updated value estimate is obtained by multiplying the learning rate α by the difference in value estimates and adding it to the old value estimate. Here, the learning rate α is... The value range is usually between 0 and 1. A value close to 0 indicates that the personal comfort model learns new experiences more slowly and relies more on previously learned knowledge. A value close to 1 indicates that the personal comfort model tends to favor the rapid adoption of information from new experiences, while 'a' represents the state of... A specific action among all the possible actions that can be taken. Indicates the next state Next, from all possible actions a, select the action-value function Q( a) The Q value corresponding to the action that achieves the maximum value. Through iteration, the personal comfort model learns the long-term value that the action A_t can bring in state S_t. Specifically, whenever a new user feedback is received, the personal comfort model is immediately updated incrementally with the new empirical data (S_t, A_t, R_t, S_{t+1}) so that the model can adapt to changes in user preferences.

[0127] In this embodiment, for each functional area, the output of the personal comfort model of all personnel is weighted and aggregated. Specifically, for each user i in the area, their current environmental state S is input into their personal comfort model to obtain their predicted comfort score Q_i, and a weight w_i is assigned to each user. This weight can be determined based on their identity, activity level, and feedback frequency. Thus, the collective comfort C_collective=(∑(w_i×C_i)) / (∑ w_i) is calculated. The value of C_collective is the collective comfort index J_comfort in the multi-objective optimization cost function, which is used to guide the optimization process.

[0128] S500: Based on the air conditioning control parameter set, the output parameters of each air conditioning unit are gradually adjusted through a switching mechanism.

[0129] In some embodiments, step S500 above includes:

[0130] Obtain the current operating parameter set of each air conditioner unit, and calculate the absolute difference of each parameter of each air conditioner unit in combination with the air conditioner control parameter set;

[0131] Based on the absolute difference, a transition path is generated using distributed linear interpolation. The transition path is a sequence of parameter outputs from the current time to the intermediate time point between the current time and the preset time.

[0132] Based on the parameter output sequence, the output parameters of each air conditioning unit are gradually adjusted;

[0133] During the adjustment process, environmental parameters and the operating status of each air conditioner are monitored in real time, and the transition path is dynamically updated based on the monitoring results.

[0134] In this embodiment, when regulating each air conditioning unit, it is necessary to assess the difference between the current operating parameter set of the air conditioning unit and the preferred air conditioning regulation parameter set, thereby generating a transition path and regulating the parameters of the air conditioning unit to avoid environmental abrupt changes caused by uneven coordination among the air conditioning units, which would lead to environmental imbalance in each functional area and make it impossible to guarantee the overall comfort.

[0135] In this embodiment, the current operating parameter set of each air conditioning unit is obtained, namely the temperature setpoint T_curr, the wind speed setpoint V_curr, and the air conditioning control parameter set obtained through the above step S400. The absolute difference of each parameter is calculated as Δt=|T_target-T_curr|, Δv=|V_target-V_curr|, where T_target and V_target represent the temperature control value and wind speed control value in the air conditioning control parameter set, respectively. Δt represents the absolute difference between the temperature control value T_target and the temperature setpoint T_curr, and Δv represents the absolute difference between the wind speed control value V_target and the wind speed setpoint V_curr. In some other embodiments, wind direction control is also included, which is determined according to actual needs.

[0136] Furthermore, based on the aforementioned differences, a transition path is generated. This transition path defines a sequence of setpoints for each parameter at a series of intermediate time points from the current moment to a future moment. Specifically, it is set using a step-by-step linear interpolation method. For example, for the temperature parameter, if it is decided to transition in N steps, the temperature setpoint for the k-th step (k from 1 to N) is: T_set(k) = T_curr + (k / N) × Δt. The total duration and number of steps N for completing this transition path can be dynamically adjusted according to the magnitude of the difference and the system inertia. When the difference is larger, the total duration can be appropriately extended. For systems with high inertia, such as slow temperature changes, the step size (each step) can be set longer. The wind speed and wind direction parameters are determined using the same method as described above, and will not be elaborated further here.

[0137] Specifically, based on the parameter output sequence generated by the aforementioned step-by-step linear interpolation method, the air conditioning system can issue control commands to each individual air conditioner unit in a step-by-step and progressive manner. For example, if the goal is to reduce the temperature from 26℃ to 24℃, and a 5-step transition is adopted with a 2-minute interval between each step, the air conditioning system will issue commands sequentially as follows: 25.6℃ -> 25.2℃ -> 24.8℃ -> 24.4℃ -> 24.0℃. This avoids the discomfort and equipment shock caused by jumping directly from 26℃ to 24℃.

[0138] In this embodiment, during the entire transition period, environmental parameters and the operating status of each air conditioner are continuously monitored. Specifically, the environmental parameters are obtained through indoor environmental monitoring sensors to capture the actual changes in the indoor environment, such as the actual rate of temperature drop. At the same time, the operating status of the air conditioner compressor and fan is monitored to ensure stability. The monitored environmental parameters are compared with the expected changes in the transition path. If a deviation is found, the subsequent transition path is dynamically updated. For example, the adjustment range of subsequent steps is automatically increased or the total transition time is appropriately extended to generate a new transition curve that better reflects the current situation. This ensures that the transition process remains smooth and stable even when faced with interference.

[0139] In some embodiments, a unified adjustment timing table is established for all air conditioning units that need to be switched. This adjustment timing table determines the parameter output sequence and switching time of each air conditioning unit, avoiding confusion when multiple air conditioning units are controlled. In addition, the same adjustment interval can be set for multiple air conditioning units to avoid short-term environmental parameter differences between regions caused by asynchronous adjustments, thus ensuring the consistency of the overall environment.

[0140] Please see Figure 2 As shown, the present invention also provides an intelligent air conditioning control system, the system comprising:

[0141] First processing module 201: used to collect outdoor environmental data, indoor environmental data and personnel status data in real time, and to fuse the data to generate a comprehensive dataset;

[0142] Second processing module 202: used to construct a dynamic temperature gradient field based on the comprehensive dataset;

[0143] The third processing module 203 is used to divide functional areas and calculate the influence relationship matrix in conjunction with the dynamic temperature gradient field.

[0144] The fourth processing module 204 is used to obtain the air conditioning control parameter set by employing a multi-objective optimization algorithm with the influence relationship matrix as the core constraint.

[0145] The fifth processing module 205 is used to gradually adjust the output parameters of each air conditioner unit based on the air conditioner control parameter set through a switching mechanism.

[0146] It is understandable that, such as Figure 1 The content of the air conditioning intelligent control method embodiments shown is applicable to the air conditioning intelligent control system embodiments. The specific functions implemented by the air conditioning intelligent control system embodiments are the same as those shown in the examples. Figure 1 The air conditioning intelligent control method shown in the embodiment is the same, and the beneficial effects achieved are the same as those shown. Figure 1 The beneficial effects achieved by the air conditioning intelligent control method embodiment shown are also the same.

[0147] It should be noted that the information interaction and execution process between the above systems are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0148] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0149] Please see Figure 3 As shown, this embodiment of the invention also provides a computer device 3, including: a memory 302 and a processor 301, and a computer program 303 stored in the memory 302. When the computer program 303 is executed on the processor 301, it implements the air conditioning intelligent control method as described in any of the above methods.

[0150] The computer device 3 may be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will understand that... Figure 3 The computer device 3 is merely an example and does not constitute a limitation on the computer device 3. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0151] The processor 301 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0152] In some embodiments, the memory 302 may be an internal storage unit of the computer device 3, such as a hard disk or memory of the computer device 3. In other embodiments, the memory 302 may be an external storage device of the computer device 3, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 3. Furthermore, the memory 302 may include both internal and external storage units of the computer device 3. The memory 302 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 302 can also be used to temporarily store data that has been output or will be output.

[0153] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the intelligent air conditioning control method as described in any of the above methods.

[0154] In this embodiment, if the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographic device / computer device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0155] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An intelligent conditioning method for an air conditioner, characterized by, include: Real-time collection of outdoor environmental data, indoor environmental data, and personnel status data; and fusion of the data to generate a comprehensive dataset. Based on the comprehensive dataset, a dynamic temperature gradient field is constructed; The functional areas are divided, and the influence relationship matrix is ​​calculated based on the dynamic temperature gradient field. Using the aforementioned influence relationship matrix as the core constraint, a multi-objective optimization algorithm is employed to obtain the air conditioning control parameter set; based on the air conditioning control parameter set, the output parameters of each individual air conditioner are progressively adjusted through a switching mechanism. The process involves real-time acquisition of outdoor environmental data, indoor environmental data, and personnel status data, followed by fusion of these data to generate a comprehensive dataset. This includes: integrating outdoor meteorological sensors, indoor environmental monitoring sensors, and personnel status detection sensors to collect data from multiple data sources; preprocessing and spatiotemporally aligning the multidimensional data sources; and fusing the spatiotemporally aligned multidimensional data sources to generate a comprehensive dataset. The step of constructing a dynamic temperature gradient field based on the comprehensive dataset includes: acquiring the real-time operating parameters and physical locations of all individual air conditioning units; constructing a temperature diffusion model for each individual air conditioning unit; calculating the temperature distribution of a single individual air conditioning unit; generating a sub-gradient field; using a field superposition algorithm, vector superimposing the sub-gradient fields of all individual air conditioning units with known heat sources to generate a composite temperature difference gradient field; and fusing the predicted environmental values ​​of the composite temperature difference gradient field with the collected data values ​​of the comprehensive dataset to dynamically adjust the composite temperature difference gradient field. The process of dividing functional areas and calculating the influence relationship matrix based on the dynamic temperature gradient field includes: dynamically dividing multiple functional areas based on the dynamic temperature gradient field, personnel status data, and indoor space layout, wherein each functional area is assigned a unique identifier and preset temperature and humidity target values; analyzing the thermal influence relationship between each functional area based on the dynamic temperature gradient field to determine the temperature gradient intensity and influence direction factor; determining the airflow path between each functional area based on the indoor space layout to generate a path unobstructedness coefficient; and calculating the matrix elements of the influence relationship matrix by combining the temperature gradient intensity, the influence direction factor, and the path unobstructedness coefficient. The process involves using the influence relationship matrix as the core constraint and employing a multi-objective optimization algorithm to obtain a set of air conditioning control parameters. This includes: constructing a multi-objective cost function with the objectives of minimizing total air conditioning energy consumption, maximizing equipment operational stability, and maximizing overall comfort; using a distributed model predictive control algorithm to simulate the evolution of indoor states under different combinations of air conditioning parameters within a preset time period; and iteratively solving the multi-objective cost function to output a set of air conditioning control parameters constrained by the influence relationship matrix.

2. The method of claim 1, wherein, The spatiotemporal alignment includes dynamic timestamp calibration and sampling frequency adjustment; The dynamic timestamp calibration specifically includes: for sensors that can be synchronized by command, the moment when the acquisition command is received is used as the data calibration timestamp; for sensors that cannot be synchronized by command, the data calibration timestamp is calculated using a time drift estimation model. The sampling frequency adjustment specifically includes: using a signal resampling method to unify all data sequences to a predetermined target sampling rate.

3. The method as described in claim 1, characterized in that, The step of fusing the predicted environmental values ​​of the composite temperature gradient field with the collected data values ​​of the comprehensive dataset to dynamically adjust the composite temperature gradient field includes: The indoor environmental data is compared with the predicted environmental values ​​by using a data fusion algorithm, and the parameters of the temperature diffusion model are dynamically adjusted to calibrate the composite temperature gradient field. When calibrating the composite temperature gradient field, dynamic environmental data is used as a boundary condition to drive the refresh of the composite temperature gradient field. The dynamic environmental data includes the opening and closing status of doors and windows, personnel status data, and outdoor environmental data.

4. The method of claim 1, wherein, The process of determining the airflow paths between the functional areas based on the indoor spatial layout and generating a path smoothness coefficient includes: Based on the dynamic temperature gradient field, an airflow path from one functional area to another is determined, and the airflow path is a virtual pipe connecting the key points of the two functional areas. Calculate the projected area of ​​the virtual pipe in the indoor space, generate a local projected area, and based on the indoor space layout, determine and calculate the sum of the projected areas of each obstruction in the virtual pipe, and generate an obstruction projected area. The path accessibility coefficient is calculated based on the local projected area and the obstructed projected area.

5. The method of claim 1, wherein, The step of gradually adjusting the output parameters of each individual air conditioner unit based on the air conditioning control parameter set through a switching mechanism includes: Obtain the current operating parameter set of each air conditioner unit, and calculate the absolute difference of each parameter of each air conditioner unit in combination with the air conditioner control parameter set; Based on the absolute difference, a transition path is generated using distributed linear interpolation. The transition path is a sequence of parameter outputs from the current time to the intermediate time point between the current time and the preset time. Based on the parameter output sequence, the output parameters of each air conditioning unit are gradually adjusted; During the adjustment process, environmental parameters and the operating status of each air conditioner are monitored in real time, and the transition path is dynamically updated based on the monitoring results.

6. An intelligent air conditioning regulation system for implementing the method of any one of claims 1-5, characterized in that, include: The first processing module is used to collect outdoor environmental data, indoor environmental data, and personnel status data in real time, and to merge the data to generate a comprehensive dataset. The second processing module is used to construct a dynamic temperature gradient field based on the comprehensive dataset. The third processing module is used to divide functional areas and calculate the influence relationship matrix in conjunction with the dynamic temperature gradient field. The fourth processing module is used to obtain the air conditioning control parameter set by employing a multi-objective optimization algorithm with the aforementioned influence relationship matrix as the core constraint. The fifth processing module is used to gradually adjust the output parameters of each individual air conditioner unit based on the set of air conditioner control parameters through a switching mechanism.

Images