Multi-split central air conditioner control method and air conditioner system

Abstract

The invention discloses a multi-split central air conditioner control method and an air conditioner system, and relates to the technical field of air conditioner control. According to the multi-split central air conditioner control method, control decision time sequence data of a plurality of split central air conditioners in a set area are obtained, feature analysis is conducted, and a regulation and control evaluation index set of each split central air conditioner in each time period in the set area is obtained; comprehensively analyzing the control evaluation index set of each time period of each online central air conditioner in the set area to obtain a comprehensive perception regulation index of each time period of each online central air conditioner in the set area; according to the method, the preset control measures are taken for the corresponding online central air conditioners in the set area on the basis of the comprehensive sensing adjustment indexes of all the time periods, so that the operation mode of each online central air conditioner can be flexibly adjusted, and it is ensured that the air conditioners can respond to load fluctuation and environment changes in real time; therefore, high efficiency and comfort of the air conditioner are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning control, and in particular to a multi-connected central air conditioning control method and an air conditioning system. Background Art

[0002] With the continuous development of intelligent building management, central air-conditioning systems are becoming more and more common in large-scale commercial venues, office buildings, hotels and other buildings. Multi-split central air-conditioning systems, as an energy-saving, efficient and flexible air-conditioning solution, are widely used in commercial, office and residential fields. They use an external host to connect multiple indoor devices to work in a network, which can achieve centralized control and independent adjustment, which not only improves the energy efficiency of the air-conditioning system, but also can meet the different needs of different rooms and areas. However, traditional multi-split central air-conditioning control systems usually rely on simple temperature control or load control strategies, which are difficult to adapt to dynamic changes in the environment and system in real time, resulting in low energy efficiency and difficulty in ensuring comfort. Especially when facing complex indoor environmental changes and changing user needs, the system's adjustment and control capabilities are very limited.

[0003] Prior art, such as the patent application with publication number CN114234370B, discloses a multi-split air conditioner control method, device, and multi-split air conditioner. The method includes: determining the lumped room temperature of a building where the multi-split air conditioner is located within a first preset time period in the past, where the lumped room temperature reflects the building's heat storage state; determining a lumped room temperature variation range based on the current set temperature of each indoor unit; obtaining meteorological data for the area where the air conditioner is located within a second preset time period in the future; calculating a lumped room temperature sequence within the second preset time period in the future based on the lumped room temperature and meteorological data within the first preset time period in the past, the lumped room temperature variation range, the meteorological data within the second preset time period in the future, and the current air conditioner energy consumption, thereby minimizing the total energy consumption of the air conditioner within the second preset time period in the future; and controlling the air conditioner operation using the lumped room temperature sequence within the second preset time period in the future as the set temperature sequence within the second preset time period in the future. The present invention utilizes the building's heat storage potential while ensuring the room temperature is within a comfortable range, minimizing the total energy consumption during future air conditioner operation.

[0004] Based on the above solution, it is found that the limitations of the existing technology include at least the following problems: the existing technology lacks flexible dynamic adjustment capabilities, that is, it is difficult to automatically optimize according to the actual usage environment and real-time data changes, which can easily lead to insufficient energy efficiency or uneven temperature, affecting comfort and increasing energy consumption. For example, in different hotel rooms and at different time periods, due to different room environments and uneven occupants, the air conditioner may be difficult to effectively adjust according to real-time conditions, which can easily lead to some rooms being overcooled or overheated, resulting in uneven temperature and affecting guests' comfort. However, the existing method is difficult to flexibly adjust the control window, which makes it difficult for the air conditioner to be adjusted in time when the load fluctuates greatly, resulting in energy waste and insufficient comfort. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a multi-split central air-conditioning control method and air-conditioning system, which solves the problems of the existing technology lacking flexible dynamic adjustment capabilities, making it difficult to adjust the air-conditioning according to environmental changes and real-time data, and easily leading to insufficient energy efficiency and uneven temperature.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a multi-connected central air-conditioning control method, comprising the following steps: obtaining control decision timing data of several online central air-conditioners in a set area, wherein the control decision data include equipment operation timing data, indoor thermal imaging timing data, and environmental perception timing data; performing feature analysis on the control decision timing data of each online central air-conditioner in the set area, and obtaining a set of regulation and control evaluation indexes for each time period of each online central air-conditioner in the set area, including an equipment energy efficiency synchronization index, a thermal coordination index, and an environmental interference index; and performing a comprehensive analysis on the control evaluation index set for each time period of each online central air-conditioner in the set area, and obtaining a comprehensive perception regulation index for each time period of each online central air-conditioner in the set area; and taking preset control measures for the corresponding online central air-conditioners in the set area based on the comprehensive perception regulation index of each time period.

[0007] Furthermore, the specific formula for calculating the comprehensive perception adjustment index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, ZtK is the comprehensive perception adjustment index of a certain online central air conditioner in the set area during a certain period of time, SbN is the equipment energy efficiency synchronization index of a certain online central air conditioner in the set area during a certain period of time, α1 is the energy efficiency synchronization adjustment coefficient stored in the database, RgX is the thermal coordination index of a certain online central air conditioner in the set area during a certain period of time, α2 is the thermal coordination adjustment coefficient stored in the database, HyR is the environmental interference index of a certain online central air conditioner in the set area during a certain period of time, α3 is the environmental interference adjustment coefficient stored in the database, and α4 is the superposition adjustment coefficient stored in the database.

[0008] Furthermore, the equipment operation sequence data includes the air supply intensity index, wind swing angle value, electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, air supply static pressure value, and refrigerant evaporation pressure value for each time period. The specific steps for obtaining the equipment energy efficiency synchronization index for each time period of each online central air conditioner in the set area are as follows: comprehensively analyze the equipment operation sequence data of each online central air conditioner in the set area to obtain the equipment evaluation index set for each time period of each online central air conditioner in the set area, including the wind domain coordination index and the refrigerant flow control index; and comprehensively analyze the equipment evaluation index set for each time period of each online central air conditioner in the set area to obtain the equipment energy efficiency synchronization index for each time period of each online central air conditioner in the set area.

[0009] Furthermore, the specific steps for obtaining the equipment evaluation index set for each time period of each online central air conditioner in the set area are as follows: read the air supply intensity index, wind swing angle value, and air supply static pressure value of each online central air conditioner in the set area for each time period, and perform a comprehensive analysis to obtain the wind domain coordination index of each online central air conditioner in the set area for each time period; read the electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, and refrigerant evaporation pressure value of each online central air conditioner in the set area for each time period, and perform a comprehensive analysis to obtain the refrigerant flow control index of each online central air conditioner in the set area for each time period.

[0010] Furthermore, the indoor thermal imaging time series data includes the temperature value and two-dimensional coordinates of each pixel point in each frame of indoor thermal imaging in each time period. The specific steps of obtaining the thermal coordination index of each online central air conditioner in the set area for each time period are as follows: inputting the indoor thermal imaging time series data of each online central air conditioner in the set area into a pre-trained thermal recognition model for comprehensive analysis to obtain a set of thermal evaluation indexes for each time period of each online central air conditioner in the set area, including a thermal comfort index, a thermal response index, and a thermal imbalance monitoring index; and performing a comprehensive analysis on the set of thermal evaluation indexes for each time period of each online central air conditioner in the set area to obtain a thermal coordination index for each time period of each online central air conditioner in the set area.

[0011] Furthermore, the thermal perception recognition model is specifically a convolutional neural network, which includes an input layer, a convolutional layer, a timing processing layer, a time-distributed fully connected layer, and an output layer. The specific steps of obtaining a thermal perception evaluation index set for each time period of each online central air conditioner in a set area are as follows: in the input layer of the convolutional neural network, the indoor thermal imaging time series data of each online central air conditioner in the set area is received and preprocessed; in the convolutional layer of the convolutional neural network, the preprocessed indoor thermal imaging time series data of each online central air conditioner in the set area is subjected to feature extraction processing to obtain a feature atlas set of each frame of indoor thermal imaging of each time period of each online central air conditioner in the set area; in the timing processing layer of the convolutional neural network, the .... The feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area is subjected to time series association processing to obtain a time series feature vector set of each online central air conditioner in each time period in the set area; in the time distribution fully connected layer of the convolutional neural network, the time series feature vector of each time period of each online central air conditioner in the set area is activated and converted to obtain a spatiotemporal feature vector of each time period of each online central air conditioner in the set area; in the output layer of the convolutional neural network, the spatiotemporal feature vector of each time period of each online central air conditioner in the set area is regression predicted to obtain a thermal comfort index, a thermal response index and a thermal imbalance monitoring index of each time period of each online central air conditioner in the set area.

[0012] Furthermore, the specific formula for calculating the thermal coordination index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, RgX is the thermal coordination index of a certain online central air conditioner in the set area during a certain period of time, RsD is the thermal comfort index of a certain online central air conditioner in the set area during a certain period of time, φ1 is the comfort adjustment coefficient stored in the database, YgF is the thermal response index of a certain online central air conditioner in the set area during a certain period of time, φ2 is the thermal response adjustment coefficient stored in the database, RsH is the thermal imbalance monitoring index of a certain online central air conditioner in the set area during a certain period of time, φ3 is the thermal imbalance adjustment coefficient stored in the database, and φ4 is the interaction adjustment coefficient stored in the database.

[0013] Furthermore, the environmental perception time series data includes the ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient air pressure fluctuation index for each time period. The specific steps for obtaining the environmental interference index for each time period of each online central air conditioner in the set area are as follows: normalizing the ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient air pressure fluctuation index for each time period of each online central air conditioner in the set area; and comprehensively analyzing the normalized ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient air pressure fluctuation index for each time period of each online central air conditioner in the set area to obtain the environmental interference index for each time period of each online central air conditioner in the set area.

[0014] Furthermore, the specific steps of taking preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception adjustment index of each time period are as follows: reading the comprehensive perception adjustment index of each online central air conditioner in the set area for each time period, and performing a comprehensive analysis to obtain the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period, and performing judgment and analysis with the predicted comprehensive perception adjustment index threshold value respectively; if the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period is lower than the preset predicted comprehensive perception adjustment index threshold value, then taking the first adjustment control measure; if the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period is higher than or equal to the preset predicted comprehensive perception adjustment index threshold value, then taking the second adjustment control measure.

[0015] A multi-connected central air-conditioning system includes: a data acquisition module for acquiring control decision time series data of several online central air conditioners in a set area, wherein the control decision data include equipment operation time series data, indoor thermal imaging time series data, and environmental perception time series data; a data feature analysis module for performing feature analysis on the control decision time series data of each online central air conditioner in the set area, and obtaining a set of control evaluation indexes for each time period of each online central air conditioner in the set area, including an equipment energy efficiency synchronization index, a thermal coordination index, and an environmental interference index; a comprehensive control analysis module for performing comprehensive analysis on the set of control evaluation indexes for each time period of each online central air conditioner in the set area, and obtaining a comprehensive perception regulation index for each time period of each online central air conditioner in the set area; and a control regulation feedback module for taking preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception regulation index of each time period.

[0016] The present invention has the following beneficial effects:

[0017] (1) The multi-connected central air-conditioning control method can flexibly adjust the operating mode of each connected central air-conditioning unit through dynamic predictive control based on real-time data and calculation of the comprehensive perception adjustment index, ensuring that the air-conditioning unit can respond to load fluctuations and environmental changes in real time. Through trend analysis and adaptive adjustment of the control window, it can make precise adjustments according to the actual needs of different rooms and different time periods, thereby ensuring that the air-conditioning system can adjust the output power in time to avoid excessively high or low temperatures, thereby effectively maintaining temperature balance and comfort, and avoiding excessive operation of the air-conditioning equipment at low loads, achieving energy-saving optimization, and then significantly improving the energy efficiency of the air-conditioning unit, avoiding energy consumption and temperature differences, and being able to maintain optimal performance under variable environmental conditions, thereby ensuring the efficiency and comfort of the air-conditioning unit.

[0018] (2) The multi-split central air-conditioning control method realizes the load balancing and energy-saving control of the multi-split air-conditioning by comprehensively analyzing the equipment operation data and using the particle swarm optimization algorithm and genetic algorithm to perform weighted optimization on the equipment evaluation index. That is, based on the wind domain coordination index and refrigerant flow control index of the equipment, the load distribution between the equipment is dynamically adjusted. For example, in a high-load or low-load environment, the working mode of the air-conditioning can be automatically adjusted to ensure the optimization of the overall energy efficiency, thereby effectively improving the overall energy efficiency. At the same time, the intelligent adjustment capability of the air-conditioning equipment reduces energy consumption and ensures the efficient operation of the air-conditioning in different usage scenarios, thereby achieving both energy saving and comfort.

[0019] (3) This multi-connected central air-conditioning control method uses a deep analysis of thermal imaging time series data and a convolutional neural network to extract the heat source characteristics of the indoor environment, and then performs time series analysis to effectively capture the dynamic trend of temperature changes. It combines the thermal comfort index and the thermal response index to accurately reflect the human body's perception of temperature changes. Through the analysis of the thermal imbalance monitoring index, it can timely identify indoor hot and cold unevenness and adjust the air-conditioning operation strategy accordingly, thereby avoiding excessive temperature differences or uncomfortable temperature fluctuations. It can then ensure that the air-conditioning operates intelligently and optimally under temperature changes and heat source distribution, and ensure comfort while improving user experience.

[0020] (4) The multi-split central air-conditioning system realizes highly automated and intelligent regulation of the multi-split central air-conditioning system through the data acquisition module, the data feature analysis module and the comprehensive control analysis module. The data acquisition module collects time series data in real time, and conducts in-depth analysis of these data through the data feature analysis module to extract key control evaluation indexes. The comprehensive control analysis module further summarizes these data to generate a predicted comprehensive perception adjustment index for the next period, providing an accurate basis for the automatic adjustment of the air-conditioning system. The control adjustment feedback module automatically adjusts the operation strategy of each air-conditioning device according to the predicted comprehensive perception adjustment index for the next period generated in real time, thereby ensuring that the air-conditioning system can optimize the operating status in real time, reduce manual intervention, improve the overall energy efficiency and operating efficiency of the system, and then improve the response speed and flexibility of the air-conditioning to adapt to complex and changing environmental conditions.

[0021] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of a multi-connected central air-conditioning control method of the present invention.

[0023] Figure 2 The present invention is a flowchart of the specific steps for obtaining the equipment energy efficiency synchronization index of each online central air conditioner in each time period in a set area in a multi-connected central air conditioner control method.

[0024] Figure 3 This is a timing diagram of the thermal comfort index of a certain online central air conditioner in a set area in a multi-connected central air conditioner control method of the present invention.

[0025] Figure 4 This is a timing diagram of the thermal response index of a certain online central air conditioner in a set area in a multi-connected central air conditioner control method of the present invention.

[0026] Figure 5 This is a timing diagram of the thermal imbalance monitoring index of a certain online central air conditioner in a set area in a multi-connected central air conditioner control method of the present invention.

[0027] Figure 6 This is a block diagram of a multi-connected central air-conditioning system of the present invention. DETAILED DESCRIPTION

[0028] See also Figure 1, an embodiment of the present invention provides a technical solution: a multi-connected central air-conditioning control method, comprising the following steps: obtaining control decision time series data of several online central air-conditioners (in each hotel room) in a set area (such as a hotel), the control decision data including equipment operation time series data, indoor thermal imaging time series data, and environmental perception time series data (of the hotel room where each online central air-conditioner is located); performing feature analysis on the control decision time series data of each online central air-conditioner in the set area, and obtaining a regulation evaluation index set for each time period (such as 5 minutes as a time period) of each online central air-conditioner in the set area, including an equipment energy efficiency synchronization index, a thermal coordination index, and an environmental interference index; and performing a comprehensive analysis on the control evaluation index set for each time period of each online central air-conditioner in the set area, and obtaining a comprehensive perception regulation index for each time period of each online central air-conditioner in the set area; and taking preset control measures for the corresponding online central air-conditioners in the set area based on the comprehensive perception regulation index of each time period.

[0029] The specific formula for calculating the comprehensive perception adjustment index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, ZtK is the comprehensive perception adjustment index of a certain online central air conditioner in the set area during a certain period of time, SbN is the equipment energy efficiency synchronization index of a certain online central air conditioner in the set area during a certain period of time, α1 is the energy efficiency synchronization adjustment coefficient stored in the database, RgX is the thermal coordination index of a certain online central air conditioner in the set area during a certain period of time, α2 is the thermal coordination adjustment coefficient stored in the database, HyR is the environmental interference index of a certain online central air conditioner in the set area during a certain period of time, α3 is the environmental interference adjustment coefficient stored in the database, and α4 is the superposition adjustment coefficient stored in the database.

[0030] It needs to be explained that the formula 1+α4*ln(1+SbN)*(1+RgX 2 ) This item is used to adjust the superposition effect between the equipment energy efficiency synchronization index and the thermal coordination index to avoid the comprehensive perception adjustment index being too high or too low.

[0031] α1, α2, α3, and α4 can be obtained through the following steps: using historical data, combined with the equipment energy efficiency synchronization index, thermal coordination index, and environmental interference index, to conduct statistical regression analysis, quantify the specific impact of each factor on the comprehensive perception regulation index, and thus fit the initial weight value; secondly, using the sensitivity analysis method, adjust the value range of each coefficient, observe its impact on the comprehensive perception regulation evaluation results, ensure the stability and rationality of the model, and based on the characteristics and actual situation of the online central air-conditioning, correct and optimize the preliminary fitting coefficients, and finally determine the coefficient value applicable to the specific online central air-conditioning.

[0032] Specifically, if Figure 2 As shown, the equipment operation sequence data includes the air supply intensity index, wind swing angle value, electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, air supply static pressure value, and refrigerant evaporation pressure value of each time period. The specific steps for obtaining the equipment energy efficiency synchronization index of each online central air conditioner in the set area for each time period are as follows: the equipment operation sequence data of each online central air conditioner in the set area are comprehensively analyzed to obtain the equipment evaluation index set of each online central air conditioner in the set area for each time period, including the wind domain coordination index and the refrigerant flow control index; and the equipment evaluation index set of each online central air conditioner in the set area for each time period is comprehensively analyzed (that is, weighted processing, and the weights of the wind domain coordination index and the refrigerant flow control index can be obtained by the following steps: set initial weights for the wind domain coordination index and the refrigerant flow control index. For example, the initial weights of the wind domain coordination index and the refrigerant flow control index are both 0.5, and the sum of the two weights is always 1, and define the objective function, that is, calculate the equipment efficiency of each online central air conditioner in each time period. The weighted formula for the energy efficiency synchronization index optimizes the equipment energy efficiency synchronization index by adjusting the weights of the wind domain coordination index and the refrigerant flow control index. A genetic algorithm is then used to search for the optimal weight combination. Specifically, an initial population is generated, in which each individual in the population represents a set of randomly generated weights. Based on the weight combination of each individual, their fitness, i.e., the equipment energy efficiency synchronization index, is calculated. The higher the fitness, the more the weight combination can optimize the equipment energy efficiency. The optimal individual, i.e., the weight combination, is selected based on the fitness. A crossover operation is then used to generate new individuals, i.e., new weight combinations. Some individuals are mutated to ensure population diversity and avoid falling into local optimal solutions. Through multiple iterations, e.g., 100 generations, the genetic algorithm will continuously optimize the individuals, i.e., the weight combinations, until the optimal weight combination is found. This weight combination is then used as the weight corresponding to the wind domain coordination index and the refrigerant flow control index. The equipment energy efficiency synchronization index (which measures the coordination and cooperation of the air conditioners during the air supply and cooling processes) is obtained for each online central air conditioner in the set area during each time period.

[0033] Among them, the air supply intensity index is the actual output intensity of the air supply of the air conditioner, which can be obtained by obtaining the operating frequency of the indoor unit (the average operating frequency at each time point in the period, and the operating frequency at each time point can be obtained by the power meter), the wind speed value (the average wind speed at each time point in the period, and the wind speed at each time point can be obtained by the wind speed sensor), and the supply air temperature value (the average supply air temperature at each time point in the period, and the supply air temperature at each time point can be obtained by the NTC thermistor sensor), and performing normalization processing. Based on the normalization processing result, weighted processing is performed, and the result obtained is the air supply intensity index.

[0034] The wind swing angle value indicates the swing angle of the indoor unit's air supply. It can obtain the angle value at each time point in the period through the angle information fed back by the rotary encoder stored in the database (reading the digital signal provided by the rotary encoder), and perform averaging processing. The result is the wind swing angle value.

[0035] The opening value of the electronic expansion valve is the size of the expansion valve opening, that is, the degree of control of the refrigerant flow. The function of the expansion valve is to regulate the flow of refrigerant from the high-pressure side to the low-pressure side. The larger the opening value, the more refrigerant flows into the evaporator and the stronger the cooling effect. The expansion valve opening at each time point in the period can be obtained through a position sensor (such as a Hall effect sensor on the electronic expansion valve, which is based on the Hall effect principle and detects changes in the magnetic field to sense the position and determines the valve opening by sensing the position change of the magnet on the valve core) and averaged. The result is the opening value of the electronic expansion valve.

[0036] The refrigerant return air temperature is the temperature of the refrigerant in the evaporator after absorbing heat and returning to the compressor. The temperature of the refrigerant in the evaporator after absorbing heat and returning to the compressor at each time point during the period can be obtained through the temperature sensor in the refrigerant pipeline, and the average value is processed. The result is the refrigerant return air temperature.

[0037] The evaporator coil temperature value is the temperature of the evaporator coil surface, which reflects the heat absorption of the refrigerant in the evaporator. The temperature of the evaporator coil surface at each time point in the period can be obtained by the temperature sensor and averaged. The result is the evaporator coil temperature value.

[0038] The supply air static pressure value is the static pressure in the air duct of the air conditioning supply system. Too high or too low static pressure is usually a sign of air duct obstruction or system mismatch, which affects the flow of air and indirectly affects the cooling or heating efficiency. The static pressure at each time point in the period can be obtained through a pressure differential sensor and averaged. The result is the supply air static pressure value.

[0039] The refrigerant evaporation pressure value is the pressure generated when the refrigerant absorbs heat and vaporizes in the evaporator, reflecting the working efficiency of the refrigeration system and the regulation of the refrigerant flow. The refrigerant evaporation pressure at each time point in the period can be obtained through a pressure sensor and averaged. The result is the refrigerant evaporation pressure value.

[0040] The specific steps for obtaining the equipment evaluation index set for each time period of each online central air conditioner in the set area are as follows: read the air supply intensity index, wind swing angle value, and air supply static pressure value of each online central air conditioner in the set area for each time period, and conduct a comprehensive analysis (i.e., first perform standardization processing, and perform weighted processing based on the standardization processing results, and update the weights corresponding to the air supply intensity index, wind swing angle value, and air supply static pressure value based on the particle swarm optimization algorithm), and obtain the wind domain coordination index of each time period of each online central air conditioner in the set area (used to measure the control effect of the air conditioner in terms of air supply, reflecting the uniformity and comfort of the air flow of the air conditioner); read the electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, and refrigerant evaporation pressure value of each online central air conditioner in the set area for each time period, and conduct a comprehensive analysis (i.e., first perform standardization processing, and then perform weighted processing based on the standardization processing results, and then update the weights corresponding to the air supply intensity index, wind swing angle value, and air supply static pressure value). The normalization processing is performed, and weighted processing is performed based on the normalization processing result. The weights corresponding to the electronic expansion valve opening value, the refrigerant return air temperature value, the evaporator coil temperature value, and the refrigerant evaporation pressure value are updated based on the particle swarm optimization algorithm, that is, multiple particles are initialized, each particle corresponds to a set of weight values corresponding to the electronic expansion valve opening value, the refrigerant return air temperature value, the evaporator coil temperature value, and the refrigerant evaporation pressure value, and the fitness function, that is, the refrigerant flow control index, is calculated according to the current parameters, such as minimizing the error, and based on the individual best position, such as the historical best and the global best position, such as the group best, the speed and position of the particle are updated to find the optimal weight combination), and the refrigerant flow control index of each online central air conditioner in the set area in each time period is obtained (which measures the ability of the air conditioner to regulate the refrigerant flow, and reflects the cooling effect of the air conditioner in the refrigeration process and the adjustment accuracy of the refrigerant flow).

[0041] In this implementation scheme, by standardizing and weighting analysis of multiple important operating parameters, such as the air supply intensity index, wind swing angle value, electronic expansion valve opening value, etc., the energy efficiency synchronization index of each air conditioner can be dynamically adjusted according to the actual performance of different equipment, thereby optimizing the overall operating efficiency of the air-conditioning equipment, improving energy efficiency, and reducing energy waste. Secondly, by calculating the wind domain coordination index, the uniformity of the air supply of the air conditioner can be effectively measured and optimized, and the temperature imbalance caused by poor air flow can be avoided. Finally, the weight of the refrigerant flow control index is updated through the particle swarm optimization algorithm, so that the refrigerant flow can be optimized and adjusted according to actual needs. Especially in the case of large load fluctuations and environmental changes, the air conditioner can adapt to the changes in real time and improve the adjustment accuracy of the refrigerant flow, thereby achieving more efficient cooling or heating, reducing system overload and insufficient energy efficiency, and thus improving overall stability and reliability.

[0042] Specifically, the indoor thermal imaging time series data includes the temperature value and two-dimensional coordinates of each pixel point in each frame of indoor thermal imaging in each time period. The specific steps of obtaining the thermal coordination index of each online central air conditioner in the set area for each time period are as follows: the indoor thermal imaging time series data of each online central air conditioner in the set area is input into the pre-trained thermal recognition model for comprehensive analysis to obtain a set of thermal evaluation indexes for each time period of each online central air conditioner in the set area, including a thermal comfort index (human body comfort in the current environment), a thermal response index (human body response intensity to indoor temperature changes), and a thermal imbalance monitoring index (indicating indoor temperature unevenness, i.e., cold and hot imbalance); a comprehensive analysis is performed on the set of thermal evaluation indexes for each time period of each online central air conditioner in the set area to obtain a thermal coordination index for each time period of each online central air conditioner in the set area.

[0043] The thermal perception recognition model is specifically a convolutional neural network, which includes an input layer, a convolution layer, a time series processing layer, a time distribution fully connected layer, and an output layer. The specific steps of obtaining the thermal perception evaluation index set of each online central air conditioner in the set area for each time period are as follows: In the input layer of the convolutional neural network, the indoor thermal imaging time series data of each online central air conditioner in the set area (that is, the temperature value and two-dimensional coordinates of each pixel point in each frame of indoor thermal imaging in each time period) are received and preprocessed; in the convolution layer of the convolutional neural network, the indoor thermal imaging time series data of each online central air conditioner in the set area after preprocessing are subjected to feature extraction processing (using multiple convolution kernels, such as a 3x3 convolution kernel sliding on each frame image, through The convolution operation extracts local features, including but not limited to the temperature difference, heat source distribution and temperature change pattern of the human body area. The convolution operation detects the temperature difference area and heat source area in the image through weighted sum calculation, thereby realizing the recognition of the human body area. Each convolution kernel will generate a feature map. These feature maps represent the local features in the image, such as the heat source location of the human body, i.e. shoulders, head, back, etc., as well as local changes, heat source distribution, temperature difference area, etc. Through the action of multiple convolution kernels, the convolution layer can extract multiple spatial features in the image, and the local features extracted by the convolution kernel are aggregated into a feature map. Each feature map represents a specific feature in the thermal imaging image, and the maximum pooling operation is performed to further reduce the spatial size of the feature map. At the same time The key information of the human body heat source area is retained. The pooling operation can reduce the computational complexity and highlight the important features in the image while keeping the features of the human body area unchanged). The feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area is obtained (that is, multiple four-dimensional feature maps of each time period, which contain key information such as human body heat source and temperature difference area in the thermal imaging image); in the time series processing layer of the convolutional neural network, the feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area is subjected to time series association processing (first, the feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area is flattened, that is, it is flattened into three-dimensional data, then each online The feature atlas of each frame of indoor thermal imaging for each time period of the central air conditioner is converted into a feature vector for each frame of indoor thermal imaging for each time period of each online central air conditioner. Then, the long short-term memory network stored in the database is called to perform time series processing. This uses its internal memory units to capture long-term dependencies in the time series, thereby learning the temporal trends of human thermal sensation. For example, after the air conditioner status changes, the human body's thermal sensation tends to lag, and the LSTM can capture this reaction delay. The output of the LSTM is the time series feature vector corresponding to each time step, that is, each frame of indoor thermal imaging within that time period (such as changes in human thermal sensation, changes in temperature difference areas, etc.). This yields a time series feature vector set for each time period of each online central air conditioner in the set area.In the time-distributed fully connected layer of the convolutional neural network, the time series feature vectors of each period of each online central air conditioner in the set area are activated and converted (the feature vectors of each period are weighted and summed, and the input feature vectors of each period are weighted and summed through the weight matrix of the fully connected layer to learn the relative importance of each feature in the period. At this time, the feature vectors will be transformed nonlinearly after weighted processing through activation functions such as ReLU and Sigmoid, introducing the nonlinear ability of the model, so that the model can learn more complex spatiotemporal patterns and features. The role of the activation function is to enhance the representation ability of the network and avoid the limitations of linear mapping. , the activated output will generate a new spatiotemporal feature vector, including but not limited to the spatial distribution changes of the human body's heat source area, the temporal changes of the temperature difference area and heat flow distribution, the temporal trend of temperature uniformity, the dynamic changes of the human body's thermal sensation distribution, and the characteristics of the temporal changes of the human body's response intensity to temperature changes). The spatiotemporal feature vector of each time period of each online central air conditioner in the set area is obtained (including but not limited to the heat source location of different parts of the human body, the temperature values of different parts of the human body, the heat source distribution area, the temperature gradient, the direction and intensity of heat flow, the temperature difference area, the temperature uniformity, the local temperature difference, the size of the human body's thermal sensation area, the change of the thermal sensation area, the thermal sensation intensity, etc.);In the output layer of the convolutional neural network, the spatiotemporal feature vectors of each time period of each online central air conditioner in the set area are subjected to regression prediction processing (i.e., for the thermal comfort index, the human body heat source position and temperature value in the spatiotemporal feature vector are extracted, that is, the heat source distribution of the human body, such as the head, shoulders, back, etc., directly affects the thermal comfort. Parts with higher temperature values are more likely to cause discomfort, temperature difference areas, large temperature difference areas, such as the transition between hot and cold areas can easily lead to discomfort, especially when the temperature difference is large, temperature uniformity, high uniformity of temperature distribution means a comfortable environment, low uniformity can easily cause discomfort, the size and change of the human body's thermal sensitivity area, the size of the human body's thermal sensitivity area is usually proportional to the comfort, the larger the area, the stronger the heat perception, and related features, etc. For the thermal response index, the change of the thermal sensitivity area in the spatiotemporal feature vector is extracted, the expansion or contraction of the human body's thermal sensitivity area reflects the intensity of the human body's response to temperature changes, thermal intensity, local temperature difference, when the temperature difference in the local area is large, the human body reacts Stronger, possibly causing discomfort, and related features, etc. For the thermal imbalance monitoring index, temperature difference areas are extracted. By calculating the temperature difference areas in the image, areas with large temperature differences are analyzed. These areas generally indicate the risk of thermal imbalance. Local temperature differences are calculated. By calculating the temperature difference fluctuations within the local area, the degree of excessive difference between hot and cold in the local area, the temperature difference fluctuations, and the phenomenon that excessive temperature difference fluctuations can easily lead to thermal imbalance, etc., the various features related to the thermal comfort index, thermal response index, and thermal imbalance monitoring index are normalized to the same range. Then, based on a weighted sum method, the contribution of each feature to thermal comfort is weighted and summed. These weighted features are mapped to the final thermal imbalance monitoring index through regression analysis, such as linear regression or other machine learning models. The above index results are all mapped to the range of 0-1 using the Sigmoid activation function. The thermal comfort index, thermal response index, and thermal imbalance monitoring index for each online central air conditioner in the set area and each time period are obtained.

[0044] The input layer is used to receive thermal imaging data, including each frame of thermal imaging image and its timestamp in each time period.

[0045] The convolutional layer is used to extract the spatial features of thermal imaging images, such as heat source distribution, temperature difference area, etc.

[0046] The temporal processing layer is used to capture the dynamic changes of thermal imaging images over time and learn temporal features such as changes in thermal areas and temperature fluctuations.

[0047] The time-distributed fully connected layer is used to process the time series features of each time period and convert them into the final spatiotemporal feature vector.

[0048] The output layer is used to generate thermal comfort index, thermal response index and thermal imbalance monitoring index.

[0049] And the pre-training steps of the convolutional neural network are as follows:

[0050] A thermal imaging dataset is obtained, including several sets of labeled thermal imaging data, and divided into an imaging training set and an imaging verification set.

[0051] Initialize the convolutional neural network, that is, randomly initialize the model weights and set hyperparameters, such as setting the learning rate (for example, 0.001), controlling the pace of model parameter updates, selecting an appropriate optimizer (such as Adam, SGD, etc.) to update the network weights, and selecting a suitable loss function based on the task. For regression tasks (such as predicting indices such as thermal comfort), the mean square error (MSE) can be used.

[0052] Training is performed based on the imaging training set, the number of training cycles is set (such as 50 times), and forward propagation is performed in sequence in each training cycle (the training data of this batch is input into the convolutional neural network for forward propagation. In the forward propagation, the network calculates the output of each layer until the final prediction result is generated, that is, the convolution layer extracts spatial features, the time series processing layer processes time dependencies, and the output layer is used to generate the final prediction output), loss is calculated (a loss function such as mean square error is used to calculate the difference between the network prediction output and the actual label. For example, MSE is used to calculate the difference between the predicted value and the actual value of the thermal comfort index), back propagation (through the back propagation algorithm, the gradient of the loss function for each weight in the network is calculated, and the network weights are updated according to the gradient. An optimizer such as Adam is usually used to minimize the loss function), and weight update (an optimizer such as Adam is used to update the weights in the convolutional neural network according to the calculated gradient to optimize the performance of the model).

[0053] After each training cycle, an evaluation analysis is performed based on the imaging validation set. Data is extracted from the imaging validation set, that is, a batch of data is extracted from the imaging validation set. The same preprocessing process as the training set is performed, but no gradient update is performed. The validation data is input into the trained model, forward propagation is performed, and the network prediction results are calculated. The performance of the model on the validation set is calculated using predetermined evaluation indicators such as MSE, MAE, etc., that is, the difference between the predicted values of the thermal comfort index, thermal response index and thermal imbalance monitoring index and the corresponding actual values is evaluated, and the loss and evaluation indicators of each training cycle are recorded. If the loss decreases and the evaluation indicators improve, the model is learning and optimizing. During the model training process, the early stopping method can be used to avoid overfitting. If the evaluation indicators of the validation set do not improve in several consecutive cycles, the training is stopped. If the performance on the validation set is poor, the learning rate, batch size or other hyperparameters can be adjusted, or the model can be fine-tuned.

[0054] After training is complete and evaluated on the validation set, save the trained model.

[0055] The specific formula for calculating the thermal coordination index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, RgX is the thermal coordination index of a certain online central air conditioner in the set area during a certain period of time, RsD is the thermal comfort index of a certain online central air conditioner in the set area during a certain period of time, φ1 is the comfort adjustment coefficient stored in the database, YgF is the thermal response index of a certain online central air conditioner in the set area during a certain period of time, φ2 is the thermal response adjustment coefficient stored in the database, RsH is the thermal imbalance monitoring index of a certain online central air conditioner in the set area during a certain period of time, φ3 is the thermal imbalance adjustment coefficient stored in the database, and φ4 is the interaction adjustment coefficient stored in the database.

[0056] It should be explained that φ1, φ2, φ3, and φ4 can be obtained through the following steps: based on historical data, the initial influence weights of each variable (thermal comfort index, thermal response index, and thermal imbalance monitoring index) on the thermal coordination index are determined through statistical regression analysis. Then, the value range of the coefficient is adjusted using the sensitivity analysis method to evaluate the stability and applicability of these parameters to the formula output. Next, the weights are further fitted through model optimization (such as machine learning algorithms or multi-objective optimization) to ensure that the formula can accurately reflect the impact of the actual air-conditioning state on the human body.

[0057] The specific implementation example of calculating the thermal coordination index of a certain online central air conditioner in a set area during a certain period is as follows. The existing data includes the thermal comfort index, thermal response index, and thermal imbalance monitoring index of a certain online central air conditioner in the set area during 5 periods, as shown in Table 1 and Figure 3-5 As shown:

[0058] Table 1 Example of a time series data set of thermal perception evaluation index of an online central air conditioner in a given area

[0059]

[0060]

[0061] The comfort adjustment coefficient φ1 stored in the database is approximately: 0.524;

[0062] The thermal response adjustment coefficient φ2 stored in the database is approximately: 1.257;

[0063] The thermal imbalance adjustment coefficient φ3 stored in the database is approximately: 0.427;

[0064] The interaction adjustment coefficient φ4 stored in the database is approximately: 0.317;

[0065] Substituting the data in Table 1 and the above adjustment coefficient into the specific formula for calculating the thermal coordination index of a certain online central air conditioner in a set area during a certain period of time, we obtain:

[0066] The thermal coordination index of the first period of a certain online central air conditioner in the set area = 0.752 0.524 ×ln(1+1.257×0.624)×(1-0.108) 0.427 ×exp(-0.317×0.752×0.624)≈0.409;

[0067] Set the thermal coordination index of the second period of a central air conditioner in the area to 0.783 0.524 ×ln(1+1.257×0.613)×(1-0.097) 0.427 ×exp(-0.317×0.783×0.613)≈0.415;

[0068] The thermal coordination index of the third period of a central air conditioner in the set area = 0.761 0.524 ×ln(1+1.257×0.618)×(1-0.104) 0.427 ×exp(-0.317×0.761×0.618)≈0.410;

[0069] The thermal coordination index of the fourth period of a central air conditioner in the set area = 0.763 0.524 ×ln(1+1.257×0.0.621)×(1-0.107) 0.427 ×exp(-0.317×0.763×0.621)≈0.411;

[0070] Set the thermal coordination index of a central air conditioner in the area for the fifth period = 0.758 0.524 ×ln(1+1.257×0.634)×(1-0.105) 0.427 ×exp(-0.317×0.758×0.634)≈0.413.

[0071] In this implementation, by inputting thermal imaging time series data and combining it with deep learning of convolutional neural networks and long short-term memory networks, multi-dimensional data such as thermal comfort index, thermal response index and thermal imbalance monitoring index can be obtained in real time, so that the changing pattern of indoor temperature can be accurately identified, and the air-conditioning operation can be dynamically adjusted to ensure that the temperature perceived by the human body is in the comfort range, thereby effectively avoiding the discomfort caused by excessive local temperature differences and achieving the accuracy of temperature regulation. Secondly, by comprehensively analyzing the thermal evaluation index in each time period, the operating status of the air-conditioning can be adjusted according to the human body's response to temperature changes, and the load can be adjusted in time to avoid excessive operation. Finally, through time series processing and eigenvector analysis, the long-term trend of indoor heat source changes, such as changes in heat source distribution, temperature difference fluctuations, etc., can be captured, so that the heat source changes can be responded to intelligently and the operating mode can be adjusted in time to adapt to the needs of different time periods, ensuring a long-term stable comfortable environment and energy-saving effects.

[0072] Specifically, the environmental perception time series data includes the (indoor) ambient temperature fluctuation index, (indoor) ambient humidity fluctuation index, air flow turbulence index, and (indoor) ambient pressure fluctuation index for each time period. The specific steps for obtaining the environmental interference index for each time period of each online central air conditioner in the set area are as follows: normalize the ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient pressure fluctuation index for each time period of each online central air conditioner in the set area; and conduct a comprehensive analysis of the ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient pressure fluctuation index for each time period of each online central air conditioner in the set area after normalization (i.e., weighted processing, and the weights of the ambient temperature fluctuation index, ambient humidity fluctuation index, air flow turbulence index, and ambient pressure fluctuation index are consistent with the weight acquisition logic of the wind domain coordination index and the refrigerant flow control index, and are all obtained through genetic algorithms), to obtain the environmental interference index for each time period of each online central air conditioner in the set area.

[0073] The ambient temperature fluctuation index is the degree of fluctuation of the indoor ambient temperature over time during the period. The indoor ambient temperature value at each time point in the period can be obtained by a temperature sensor and subjected to standard deviation processing. The result is the ambient temperature fluctuation index.

[0074] The ambient humidity fluctuation index is the degree of fluctuation of indoor ambient humidity over time during the period. The indoor ambient humidity value at each time point in the period can be obtained by a humidity sensor and subjected to standard deviation processing. The result is the ambient humidity fluctuation index.

[0075] The airflow turbulence index is the degree of turbulence of indoor air flow during the period. It can obtain the indoor environmental wind speed value at each time point in the period through the wind speed sensor and perform standard deviation processing. The result is the airflow turbulence index.

[0076] The ambient air pressure fluctuation index is the degree of fluctuation of the indoor ambient air pressure over time during the period. The indoor air pressure and temperature values at each time point in the period can be obtained by the air pressure sensor and processed with the standard deviation. The result is the ambient air pressure fluctuation index.

[0077] In this implementation scheme, by normalizing four key environmental parameters, namely ambient temperature fluctuation, humidity fluctuation, airflow turbulence and air pressure fluctuation, the interference factors in the indoor environment can be fully quantified, and the environmental interference index can be accurately obtained, thereby ensuring that the air-conditioning system can perceive and respond to environmental changes in real time, and identify the source of fluctuations that may affect comfort, so as to adjust the operating mode and maintain the stability and comfort of the indoor environment. Secondly, a genetic algorithm is used to perform weighted analysis on these environmental parameters, and by dynamically adjusting the weights of each environmental index, the air-conditioning system can flexibly optimize the adjustment strategy according to the degree of influence of different environmental fluctuations. Finally, by comprehensively analyzing multiple interference factors in the environment, the air-conditioning can intelligently predict and adapt to various environmental conditions, especially in complex and changing environments, so as to foresee environmental changes in advance and make corresponding adjustments, thereby improving the stability and long-term operation efficiency of the air-conditioning.

[0078] Specifically, the specific steps of taking preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception adjustment index of each time period are as follows: read the comprehensive perception adjustment index of each online central air conditioner in the set area for each time period, and perform comprehensive analysis to obtain the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period, and perform judgment analysis with the predicted comprehensive perception adjustment index threshold value respectively; if the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period is lower than the preset predicted comprehensive perception adjustment index threshold value, then (for the online central air conditioner) take the first adjustment control measure, which is specifically: appropriately increase the output to make up for the insufficient energy efficiency and temperature deviation, ensure that the indoor temperature is within the set range, increase the wind speed of the air conditioner, improve air circulation, help the temperature to quickly balance, reduce temperature difference, and improve comfort, and adjust the load distribution to reallocate the load to equipment with better performance (that is, the predicted comprehensive perception adjustment index is higher than the preset predicted comprehensive perception adjustment index threshold value). The system can also be used to measure the comprehensive perception adjustment index threshold of each online central air conditioner in the set area), avoid overloading of individual equipment, improve overall energy efficiency, and enhance the sensitivity of environmental sensors, monitor and feedback environmental data such as temperature and humidity in real time, so that the air conditioner can respond quickly according to real-time environmental changes; if the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next period is higher than or equal to the preset predicted comprehensive perception adjustment index threshold, then (for the online central air conditioner) a second adjustment control measure is taken, which is specifically: reduce the output of the air conditioner to avoid over-adjustment, reduce energy consumption, and maintain a suitable indoor temperature, reduce wind speed, reduce the intensity of air flow, reduce unnecessary energy consumption, and avoid noise and discomfort caused by excessive wind speed of the air conditioner. According to the load and demand of the system, the working mode of the equipment is automatically adjusted, such as switching some equipment to standby or low-power operation state to avoid unnecessary energy waste, and adjust the temperature setting value in time to make it more in line with actual needs, avoid the temperature being too low or too high, and causing unnecessary energy efficiency consumption.

[0079] The specific steps of obtaining the predicted comprehensive perceived regulation index for the next time period of each online central air conditioner in the set area are as follows: performing trend analysis on the comprehensive perceived regulation index for each time period of each online central air conditioner in the set area to obtain the change rate of the comprehensive perceived regulation index for several groups of adjacent time periods of each online central air conditioner in the set area, and dynamically selecting a prediction window for each online central air conditioner in the set area based on specific rules (such as error minimization and similarity of the change rate of the comprehensive perceived regulation index) (for example, if the similarity of the change rate of the comprehensive perceived regulation index for three consecutive time periods is within a preset interval, the three consecutive time periods are set as one window). Then, averaging the comprehensive perceived regulation index for each time period within the prediction window of each online central air conditioner in the set area to obtain a predicted reference value for each online central air conditioner in the set area, and performing weighted averaging on the change rate of the comprehensive perceived regulation index for each group of adjacent time periods within the prediction window of each online central air conditioner to obtain a predicted change value for each online central air conditioner in the set area. Coupled processing is performed with the predicted reference value to obtain the predicted comprehensive perceived regulation index for the next time period of each online central air conditioner in the set area.

[0080] In this implementation scheme, by comprehensively analyzing the comprehensive perception adjustment index of each time period and predicting the change trend of the next time period, each air conditioner can adjust its working status in a timely and accurate manner and ensure that the indoor temperature is always maintained within a comfortable range. Secondly, through trend analysis and weighted average processing, each air conditioner can predict future change trends based on the environmental data of each time period and make advance adjustments, so that each air conditioner can cope with changing environmental conditions, such as external climate changes, room load fluctuations, etc., thereby improving the adaptability and stability of each air conditioner. Finally, by using prediction windows and weighted averages, each air conditioner can optimize the adjustment strategy to ensure real-time response to changes in accordance with actual needs.

[0081] See also Figure 6, an embodiment of the present invention provides a technical solution: a multi-connected central air-conditioning system, comprising: a data acquisition module, used to acquire control decision time series data of several online central air conditioners in a set area, the control decision data including equipment operation time series data, indoor thermal imaging time series data, and environmental perception time series data; a data feature analysis module, used to perform feature analysis on the control decision time series data of each online central air conditioner in the set area, and obtain a control evaluation index set for each time period of each online central air conditioner in the set area, including an equipment energy efficiency synchronization index, a thermal coordination index, and an environmental interference index; a comprehensive control analysis module, used to perform comprehensive analysis on the control evaluation index set for each time period of each online central air conditioner in the set area, and obtain a comprehensive perception adjustment index for each time period of each online central air conditioner in the set area; a control adjustment feedback module, used to take preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception adjustment index of each time period.

[0082] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0083] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A multi-split central air conditioner control method, characterized in that: The following steps are involved: Obtain control decision time series data for several online central air conditioners in a set area, wherein the control decision data includes equipment operation time series data, indoor thermal imaging time series data, and environmental perception time series data; Perform feature analysis on the control decision time series data of each online central air conditioner in the set area, and obtain a set of control evaluation indexes for each online central air conditioner in each time period in the set area, including the equipment energy efficiency synchronization index, thermal coordination index, and environmental interference index. And conduct a comprehensive analysis on the control evaluation index set of each online central air conditioner in each time period in the set area to obtain the comprehensive perception adjustment index of each online central air conditioner in each time period in the set area; Based on the comprehensive perception adjustment index of each time period, preset control measures are taken for the corresponding online central air conditioners in the set area.

2. The multi-split central air conditioner control method according to claim 1, characterized in that: The specific formula for calculating the comprehensive perception adjustment index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, ZtK, SbN, RgX, and HyR are the comprehensive perception adjustment index, equipment energy efficiency synchronization index, thermal coordination index, and environmental interference index of a certain online central air conditioner in a set area during a certain period of time, respectively. α1, α2, α3, and α4 are the energy efficiency synchronization adjustment coefficient, thermal coordination adjustment coefficient, environmental interference adjustment coefficient, and superposition adjustment coefficient stored in the database, respectively.

3. The multi-split central air conditioner control method according to claim 1, characterized in that: The equipment operation sequence data includes the air supply intensity index, wind swing angle value, electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, air supply static pressure value, and refrigerant evaporation pressure value for each time period. The specific steps for obtaining the equipment energy efficiency synchronization index for each time period of each online central air conditioner in the set area are as follows: Comprehensively analyze the equipment operation sequence data of each online central air conditioner in the set area to obtain a set of equipment evaluation indexes for each time period of each online central air conditioner in the set area, including the wind domain coordination index and the refrigerant flow control index; A comprehensive analysis is performed on the equipment evaluation index set of each online central air conditioner in each time period in the set area to obtain the equipment energy efficiency synchronization index of each online central air conditioner in each time period in the set area.

4. The multi-split central air conditioner control method according to claim 3, characterized in that: The specific steps for obtaining the equipment evaluation index set for each online central air conditioner in each time period in the set area are as follows: Read the air supply intensity index, wind swing angle value, and air supply static pressure value of each online central air conditioner in the set area at each time period, and conduct a comprehensive analysis to obtain the wind domain coordination index of each online central air conditioner in the set area at each time period; The electronic expansion valve opening value, refrigerant return air temperature value, evaporator coil temperature value, and refrigerant evaporation pressure value of each online central air conditioner in the set area in each time period are read, and a comprehensive analysis is performed to obtain the refrigerant flow control index of each online central air conditioner in the set area in each time period.

5. The multi-split central air conditioner control method according to claim 1, characterized in that: The indoor thermal imaging time series data includes the temperature value and two-dimensional coordinates of each pixel point in each frame of indoor thermal imaging in each time period. The specific steps for obtaining the thermal coordination index of each online central air conditioner in the set area in each time period are as follows: The indoor thermal imaging time series data of each online central air conditioner in the set area is input into the pre-trained thermal recognition model for comprehensive analysis. The thermal evaluation index set for each online central air conditioner in the set area at each time period is obtained, including the thermal comfort index, thermal response index, and thermal imbalance monitoring index. A comprehensive analysis is performed on the thermal perception evaluation index set of each online central air conditioner in each time period in the set area to obtain the thermal perception coordination index of each online central air conditioner in each time period in the set area.

6. The multi-split central air conditioner control method according to claim 5, characterized in that: The thermal perception recognition model is specifically a convolutional neural network, which includes an input layer, a convolution layer, a time series processing layer, a time distribution fully connected layer, and an output layer. The specific steps of obtaining the thermal perception evaluation index set for each online central air conditioner in each time period in the set area are as follows: In the input layer of the convolutional neural network, the indoor thermal imaging time series data of each online central air conditioner in the set area is received and preprocessed; In the convolutional layer of the convolutional neural network, feature extraction is performed on the pre-processed indoor thermal imaging time series data of each online central air conditioner in the set area to obtain a feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area; In the time series processing layer of the convolutional neural network, the feature atlas of each frame of indoor thermal imaging of each online central air conditioner in each time period in the set area is subjected to time series association processing to obtain a time series feature vector set of each online central air conditioner in each time period in the set area; In the time-distributed fully connected layer of the convolutional neural network, the time series feature vectors of each online central air conditioner in each time period in the set area are activated and converted to obtain the spatiotemporal feature vectors of each online central air conditioner in each time period in the set area; In the output layer of the convolutional neural network, the spatiotemporal feature vectors of each online central air conditioner in each time period in the set area are subjected to regression prediction processing to obtain the thermal comfort index, thermal response index, and thermal imbalance monitoring index of each online central air conditioner in each time period in the set area.

7. The multi-split central air conditioner control method according to claim 5, characterized in that: The specific formula for calculating the thermal coordination index of a certain online central air conditioner in a set area during a certain period of time is as follows: Among them, RgX, RsD, YgF, and RsH are the thermal coordination index, thermal comfort index, thermal response index, and thermal imbalance monitoring index of a certain online central air conditioner in the set area during a certain period of time, respectively. φ1, φ2, φ3, and φ4 are the comfort adjustment coefficient, thermal response adjustment coefficient, thermal imbalance adjustment coefficient, and interaction adjustment coefficient stored in the database, respectively.

8. The multi-split central air conditioner control method according to claim 1, characterized in that: The environmental perception time series data includes the ambient temperature fluctuation index, ambient humidity fluctuation index, airflow turbulence index, and ambient pressure fluctuation index for each time period. The specific steps for obtaining the environmental interference index for each time period of each online central air conditioner in the set area are as follows: Normalize the ambient temperature fluctuation index, ambient humidity fluctuation index, airflow turbulence index, and ambient pressure fluctuation index of each online central air conditioner in each time period within the set area; A comprehensive analysis is performed on the normalized ambient temperature fluctuation index, ambient humidity fluctuation index, airflow turbulence index, and ambient air pressure fluctuation index of each online central air conditioner in each time period in the set area to obtain the environmental interference index of each online central air conditioner in each time period in the set area.

9. The multi-split central air conditioner control method according to claim 1, characterized in that: The specific steps for taking preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception adjustment index of each time period are as follows: Read the comprehensive perception adjustment index of each online central air conditioner in the set area for each time period, perform comprehensive analysis, obtain the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next time period, and perform judgment analysis with the predicted comprehensive perception adjustment index threshold respectively; If the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next period is lower than the preset predicted comprehensive perception adjustment index threshold, the first adjustment control measure is taken; If the predicted comprehensive perception adjustment index of each online central air conditioner in the set area for the next period is higher than or equal to a preset predicted comprehensive perception adjustment index threshold, the second adjustment control measure is taken.

10. A multi-split central air conditioning system, applying the multi-split central air conditioning control method according to any one of claims 1 to 9, characterized in that: include: A data acquisition module is used to obtain control decision time series data of several online central air conditioners in a set area. The control decision data includes equipment operation time series data, indoor thermal imaging time series data, and environmental perception time series data; The data feature analysis module is used to perform feature analysis on the control decision time series data of each online central air conditioner in the set area, and obtain a set of control evaluation indexes for each online central air conditioner in each time period in the set area, including the equipment energy efficiency synchronization index, thermal coordination index, and environmental interference index; A comprehensive control analysis module is used to comprehensively analyze the control evaluation index set of each online central air conditioner in each time period within the set area to obtain a comprehensive perception adjustment index of each online central air conditioner in each time period within the set area; The control and adjustment feedback module is used to take preset control measures for the corresponding online central air conditioners in the set area based on the comprehensive perception adjustment index of each time period.

Images