Cold source group control method, device, equipment and storage medium based on artificial intelligence

Abstract

The present application relates to a cold source group control method, device, equipment and storage medium based on artificial intelligence, the method comprising obtaining comfort prediction information based on a preset comfort prediction model; obtaining cold capacity prediction information based on a preset cold capacity prediction model; analyzing the comfort prediction information and the cold capacity prediction information together to determine whether the prediction information meets the preset energy-saving index, and obtaining parameter model input information according to the determination result; inputting the parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model; and controlling the cold source group control system based on the parameter control information. The present application has the effect of improving the balance between energy saving and user comfort of the cold source group control system, and achieving the effect of reducing energy consumption while improving user comfort.

Description

Technical Field

[0001] The present invention relates to the technical field of cold source group control, and in particular to a cold source group control method, device, equipment and storage medium based on artificial intelligence. Background Art

[0002] The cold source group control system is mainly used to analyze, classify, process and judge the information of various electromechanical equipment in the central air-conditioning system, adopt the most optimized control means and combine modern computer technology to conduct comprehensive and effective monitoring and management of various system equipment, so that all equipment is always in an orderly, coordinated, efficient and orderly state of operation, to ensure a comfortable and safe environment in the building, save energy consumption and various daily management expenses, ensure the full operation of the system, and make the investment get a good return.

[0003] On the premise of meeting the comfort requirements, the cold source group control system can minimize the operating costs of the building by rationally organizing the operation of equipment, that is, optimizing the system control with the lowest energy consumption value as the control target. The cold source group control system software has an energy-saving program that can control the equipment to operate reasonably. Through the computer control program, each device is monitored and controlled, and the power consumption of all equipment is uniformly allocated, which can achieve the optimal control of the power load, effectively save electricity and reduce waste; in addition, the cold source group control system can monitor, diagnose and record the operation status of the equipment, detect and eliminate faults at an early stage, and issue maintenance and maintenance notices in time to ensure that the equipment is always in good working condition and the normal operation of the building. The operating status of all controlled equipment is always under the monitoring of the system. The system can provide complete operation records of the equipment and can regularly print out maintenance and maintenance notices, so as to ensure that the maintenance personnel do not perform equipment maintenance in advance or on time, so as to extend the operating life of the equipment, that is, reduce the operating costs of the building.

[0004] In the related technology, the refrigeration system control node has multiple PID (proportional integral differential) control points, such as the cooling water system control loop composed of cooling towers, cooling pumps, and chillers. Its PID control is to adjust the fan speed, pump speed, and number of pumps through the cooling water outlet temperature setting value, cooling water temperature difference, and outlet water flow. A complete PID must have input, control output, and feedback signals. When the system is determined, the above parameters are determined, and the external adjustable value is the setting value. For a cooling tower, the setting value includes the number of cooling towers, the number of cooling tower fans, the cooling tower outlet water temperature, etc. These values ​​have a threshold range. However, PID control works based on preset algorithms and parameters, which have been determined during the system design and debugging stages. During actual operation, the system's operating status will be affected by many factors, such as ambient temperature, humidity, load changes, etc. Changes in these factors will cause deviations between the actual operating status of the system and the preset parameters, making it difficult for PID control to accurately adjust the system's operating status; and, PID control can usually only be adjusted based on a preset threshold range, and cannot be dynamically adjusted based on real-time system status and user needs. In a refrigeration system, the optimal operating state of the system will be different in different time periods and under different load conditions. If PID control cannot be dynamically adjusted based on real-time conditions, it will be difficult to achieve a balance between energy saving and user comfort. Summary of the invention

[0005] In order to improve the balance between energy saving and user comfort in a cold source group control system, and to reduce energy consumption while improving user comfort, the present application provides a cold source group control method, device, equipment and storage medium based on artificial intelligence.

[0006] In the first aspect, the above-mentioned invention objective of the present application is achieved through the following technical solutions:

[0007] A cold source group control method based on artificial intelligence, the cold source group control method based on artificial intelligence comprising:

[0008] Based on a preset comfort prediction model, obtain comfort prediction information;

[0009] Based on the preset cooling capacity prediction model, obtain cooling capacity prediction information;

[0010] Analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information according to the determination result;

[0011] Inputting the parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model;

[0012] Based on the parameter control information, the cold source group control system is controlled.

[0013] By adopting the above technical solutions, the comfort prediction model and the cooling capacity prediction model are large artificial intelligence models. The comfort prediction model is used to determine the comfort of users who are currently using the cold source group control system under different cold source terminal parameters, and the cooling capacity prediction model is used to determine the cooling capacity demand of users who are currently using the cold source group control system. User comfort and cooling capacity demand are both parameter information that represent user needs, and are both related to user experience. However, user comfort and cooling capacity demand are not equal. Even if different users have the same cooling capacity demand, there may be differences in user comfort. On the contrary, when different users have the same user comfort, different users may have different cooling capacity demands. Therefore, through the comfort prediction model and the cooling capacity prediction model, user comfort and cooling capacity demand are analyzed respectively, and user demand conditions are explored from more angles, thereby discovering more In addition, the hierarchical fusion model is used to determine the optimal control scheme of the cold source group control system under different user needs. The hierarchical fusion model divides the overall refrigeration system into multiple layers. For example, the chiller and related equipment are divided into one layer, the terminal air-conditioning equipment is divided into one layer, and the refrigeration object equipment is divided into one layer. The energy efficiency and load conditions in the refrigeration system of each layer are judged respectively, and they are adapted to the control objects of the cold source group control system, so as to realize adaptive control and precise control of the control objects of the cold source group control system while meeting the user comfort, reduce the interference of intermediate variables, reduce the disconnection and conflict of the control and optimization systems, thereby improving the balance between energy saving and user comfort of the cold source group control system, and achieving reduced energy consumption while improving user comfort.

[0014] In a preferred example, the present application may be further configured as follows: the comfort prediction model is obtained by: acquiring user comfort reference data, the user comfort reference data including user feedback data, environmental parameter data, user behavior data and physiological parameter data;

[0015] Performing model training based on the user comfort reference data to generate a comfort prediction model;

[0016] The cooling capacity prediction model is obtained by:

[0017] Acquiring cooling demand reference data, the cooling demand reference data including historical cooling demand data, building and facility data, time series data, environmental factor data, population density data, and energy price data;

[0018] Model training is performed based on the cooling demand reference data to generate a cooling prediction model.

[0019] By adopting the above technical solution, the comfort prediction model and the cooling prediction model judge user needs from different angles respectively. Therefore, the training methods and training samples of the comfort prediction model and the cooling prediction model are different. The training parameter data of the comfort prediction model include user feedback data, environmental parameter data, user behavior data and physiological parameter data, etc. The user needs represented by the user comfort reference data are more inclined to the needs in terms of the user's actual feelings. For example, the user feedback data includes the comfort rating or feedback directly provided by the user, such as obtained through questionnaires, online evaluations or satisfaction scores within the application. Therefore, model training through user feedback data can directly and accurately predict the user's actual needs. The training data for the cooling prediction model include historical cooling demand data, building and facility data, time series data, environmental factor data, population density data, and energy price data. The user demand represented by the cooling demand reference data is more inclined to represent the user demand in combination with objective environmental factors. For example, by collecting data such as the type, area, and usage of buildings and facilities, and combining it with the cooling output data, a more targeted training sample is constructed. In this way, the model training through building and facility data can take into account the demand characteristics of users in buildings currently using the cold source group control system, thereby accurately predicting the user's cooling demand.

[0020] In a preferred example, the present application can be further configured as follows: the hierarchical fusion model is obtained by:

[0021] Acquire physical facility data, and establish a facility physical model based on the physical facility data;

[0022] Acquire historical load data, determine load drop according to the historical load data, and obtain load stratification information;

[0023] Based on the load stratification information, establishing a load model;

[0024] The adjustable parameter information is obtained, the facility physical model and the load model are integrated, the adjustable parameter information is used as a model input, and a hierarchical fusion model is generated.

[0025] By adopting the above technical solution, the physical facility data represents the relevant data of the physical facilities of the building currently using the cold source group control system, including the relevant data of each device of the cold source group control system, the relevant data of each device of the refrigeration system, the relevant data of the user equipment on the refrigeration terminal, the relevant data of the refrigeration area of ​​each refrigeration terminal, the building sensor data and the user data of the refrigeration area of ​​each refrigeration terminal. By associating each load-side object on the transmission path of the cold load with the load usage, a facility physical model is established to facilitate a clear and intuitive display of the load consumption, thereby determining the energy utilization efficiency of each energy usage node; the historical load data refers to the historical load data of each node of the facility physical model. On the transmission path of the cold load, not every node will have load consumption transfer. When analyzing the control parameters of the cold source group control system selected under different user needs, if there is no actual load consumption transfer, there is no need to consider the load situation at this node. Therefore, the load drop is determined by the historical load data, the nodes where the actual load change can cause energy consumption change are extracted, the load model is established, and then the facility physical model and the load model are integrated, so that the hierarchical fusion model is more practical.

[0026] In a preferred example, the present application may be further configured as follows: the acquisition of adjustable parameter information, fusion of the facility physical model and the load model, and use of the adjustable parameter information as model input to generate a hierarchical fusion model specifically include:

[0027] Acquire adjustable parameter information and historical parameter adjustment information, and use the historical parameter adjustment information and the historical load data as a training data set;

[0028] Based on a preset model fusion algorithm, the facility physical model and the load model are fused to generate a fusion model, wherein the model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross fusion algorithm;

[0029] The adjustable parameter information is used as a model input, and the fusion model is trained based on the training data set to generate a hierarchical fusion model.

[0030] By adopting the above technical solution, the adjustable parameter information refers to the equipment parameters in the refrigeration system that can be adjusted, and because the parameter adjustment is automatically performed through the cold source group control system, the adjustable parameter information refers to the parameters that can be automatically adjusted through the cold source group control system. By using the adjustable parameter information as the model input, it is convenient to learn the adjustment amount of each adjustable parameter under different user load requirements; when the facility physical model and the load model are integrated, the model fusion is performed through a variety of preset model fusion algorithms, including a linear fitting algorithm, a physical control fitting algorithm and a cross-fusion algorithm. For example, when the fusion is performed through the linear fitting algorithm, the linear change trend of the load of each node on the transmission path of the cold load represented by the load model is considered, and the control linear relationship between each physical device represented by the facility physical model and the controlled device is considered, and then the linear change trend of the load of each node on the transmission path of the cold load represented by the load model and the control linear relationship between each physical device represented by the facility physical model and the controlled device are fitted, so that each node in the facility physical model and the load model can be better integrated. In this way, by using a variety of different fusion algorithms, the adaptability of the fusion of the facility physical model and the load model can be guaranteed, and the output accuracy of the hierarchical fusion model can be guaranteed.

[0031] In a preferred example, the present application can be further configured as follows: the combined comfort prediction information and the cooling capacity prediction information are analyzed to determine whether the prediction information meets the preset energy-saving index, and according to the determination result, parameter model input information is obtained, specifically including:

[0032] Converting the comfort level prediction information into comfort level cooling capacity prediction information;

[0033] Comparing the comfort cooling capacity prediction information with the cooling capacity prediction information, and obtaining high comfort predicted cooling capacity information according to the comparison result;

[0034] It is determined whether the high comfort predicted cooling capacity information meets the preset energy-saving index, and according to the determination result, parameter model input information is obtained.

[0035] By adopting the above technical scheme, the comfort prediction information output by the comfort prediction model represents the user's future comfort information, and the comfort prediction information is relatively abstract information and cannot directly represent the user's demand. Therefore, the comfort prediction information is converted into comfort cooling capacity prediction information. For example, the correlation between the user's comfort and the user's cooling capacity demand is determined through the training data of the comfort prediction model and the cooling capacity prediction model, and the comfort prediction information is converted into comfort cooling capacity prediction information through the correlation between the user's comfort and the user's cooling capacity demand. In order to achieve the purpose of meeting the user's comfort requirements, it is necessary to use the cooling capacity prediction information that represents a higher user comfort while meeting the energy-saving indicators. Therefore, by comparing the comfort cooling capacity prediction information and the cooling capacity prediction information, when the ambient temperature is high, the comfort cooling capacity prediction information and the cooling capacity prediction information are different. The one with higher prediction information, that is, higher cooling capacity, will be able to improve user comfort more. When the ambient temperature is moderate, it is necessary to consider the gap between the suitable human body temperature and the cooling capacity under the current ambient temperature. The smaller the gap, the better the user comfort can be improved. Therefore, according to the current ambient temperature and seasonal factors, the cooling capacity prediction information that can better improve the user comfort is selected from the comfort cooling capacity prediction information and the cooling capacity prediction information to obtain high-comfort predicted cooling capacity information. If the high-comfort predicted cooling capacity information can meet the energy-saving index, the high-comfort predicted cooling capacity information can be used as the input of the hierarchical fusion model, so that the parameter control information obtained based on the hierarchical fusion model can meet the user comfort while meeting the energy-saving index, thereby improving the balance between energy saving and user comfort of the cold source group control system, and achieving reduced energy consumption while improving user comfort.

[0036] In the second aspect, the above invention objective of the present application is achieved through the following technical solutions:

[0037] A cold source group control device based on artificial intelligence, the cold source group control device based on artificial intelligence comprising:

[0038] A comfort prediction module, used for obtaining comfort prediction information based on a preset comfort prediction model;

[0039] A cooling capacity prediction module is used to obtain cooling capacity prediction information based on a preset cooling capacity prediction model;

[0040] An input analysis module is used to analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information according to the determination result;

[0041] A model analysis module, used for inputting the parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model;

[0042] The control module is used to control the cold source group control system based on the parameter control information.

[0043] Optionally, the cold source group control device based on artificial intelligence also includes:

[0044] A comfort data acquisition module, used to acquire user comfort reference data, wherein the user comfort reference data includes user feedback data, environmental parameter data, user behavior data and physiological parameter data;

[0045] A comfort model generation module, used for performing model training based on the user comfort reference data to generate a comfort prediction model;

[0046] A cooling demand data acquisition module is used to acquire cooling demand reference data, wherein the cooling demand reference data includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data and energy price data; a cooling demand model generation module is used to perform model training based on the cooling demand reference data to generate a cooling prediction model.

[0047] Optionally, the cold source group control device based on artificial intelligence also includes:

[0048] A physical model generation module is used to obtain physical facility data and establish a facility physical model based on the physical facility data; a load stratification analysis module is used to obtain historical load data, determine the load drop based on the historical load data, and obtain load stratification information;

[0049] A load model generation module, used to establish a load model based on the load stratification information;

[0050] The fusion model module is used to obtain adjustable parameter information, fuse the facility physical model and the load model, use the adjustable parameter information as model input, and generate a hierarchical fusion model.

[0051] In the third aspect, the above invention objective of the present application is achieved through the following technical solutions:

[0052] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned cold source group control method based on artificial intelligence when executing the computer program.

[0053] In a fourth aspect, the above-mentioned invention objective of the present application is achieved through the following technical solutions:

[0054] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned cold source group control method based on artificial intelligence.

[0055] In summary, the present application includes at least one of the following beneficial technical effects:

[0056] 1. Through the comfort prediction model and the cooling capacity prediction model, the user comfort and cooling capacity demand are analyzed respectively, and the user demand situation is explored from more angles, so as to find more practical and accurate user needs, so as to carry out targeted control of the cold source group control system; in addition, the hierarchical fusion model is used to determine the optimal control scheme of the cold source group control system under different user needs. The hierarchical fusion model divides the overall refrigeration system into multiple layers. For example, the chiller and related equipment are divided into one layer, the terminal air-conditioning equipment is divided into one layer, and the refrigeration object equipment is divided into one layer. The energy efficiency and load situation in the refrigeration system of each layer are determined respectively, and adapted to the control object of the cold source group control system, so as to realize adaptive control and precise control of the control object of the cold source group control system while satisfying the user comfort, reduce the interference of intermediate variables, reduce the disconnection and conflict of the control and optimization system, thereby improving the balance between energy saving and user comfort of the cold source group control system, and realizing the improvement of user comfort while reducing energy consumption;

[0057] 2. Model training through user feedback data can directly and accurately predict the user's comfort level; the training data used for the cooling prediction model includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data and energy price data, etc. The user demand represented by the cooling demand reference data is more inclined to represent the user demand in combination with objective environmental factors. For example, by collecting data such as the type, area and usage of buildings and facilities, and combining it with cooling output data, a more targeted training sample is constructed. In this way, model training through building and facility data can take into account the demand characteristics of users in buildings currently using the cooling source group control system, thereby accurately predicting the user's cooling demand;

[0058] 3. The physical facility data represents the relevant data of the physical facilities of the building currently using the cold source group control system, including the relevant data of each device of the cold source group control system, the relevant data of each device of the refrigeration system, the relevant data of the user equipment on the refrigeration terminal, the relevant data of the refrigeration area of ​​each refrigeration terminal, the building sensor data and the user data of the refrigeration area of ​​each refrigeration terminal. By associating each load-side object on the transmission path of the cold load with the load usage, a facility physical model is established to facilitate a clear and intuitive display of the load consumption, thereby determining the energy utilization efficiency of each energy usage node; the historical load data refers to the historical load data of each node of the facility physical model. On the transmission path of the cold load, not every node will have load consumption transfer. When analyzing the control parameters of the optimal cold source group control system under different user needs, if there is no actual load consumption transfer, there is no need to consider the load situation at this node. Therefore, the load gap is determined by the historical load data, and the nodes where the actual load change can lead to energy consumption changes are extracted to establish a load model, and then the facility physical model and the load model are integrated to make the hierarchical fusion model more practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of an implementation of a cold source group control method based on artificial intelligence in an embodiment of the present application;

[0060] Figure 2 is a flowchart of the method for obtaining the comfort prediction model in the embodiment of the present application;

[0061] Figure 3 is a flowchart of the method for obtaining the cooling capacity prediction model in the embodiment of the present application;

[0062] Figure 4 is a flowchart of the method for obtaining a hierarchical fusion model in an embodiment of the present application;

[0063] Figure 5 It is a flow chart for implementing S004 of the cold source group control method based on artificial intelligence in an embodiment of the present application;

[0064] Figure 6 is a flowchart for implementing S30 of the cold source group control method based on artificial intelligence in an embodiment of the present application;

[0065] Figure 7 It is a principle block diagram of a cold source group control device based on artificial intelligence in an embodiment of the present application;

[0066] Figure 8 It is a diagram of the internal structure of a computer device in an embodiment of the present application. DETAILED DESCRIPTION

[0067] The following is combined with Figure 1-8 This application is described in further detail.

[0068] The cold source group control method based on artificial intelligence is applied to the cold source group control system. The cold source group control system is a distributed, open control system to meet the technical requirements of decentralized control and centralized management. The cold source group control system adopts a three-layer network structure: management layer, control layer and field layer. The management layer network adopts TCP / IP network transmission, and the data transmission rate is not less than 10Mbps. The control layer network adopts bus network transmission and complies with Modbus and CAN standard protocols; all DDC controllers in the control layer communicate in a point-to-point manner through the control layer bus, independent of the upper layer equipment; the field layer is various sensors and actuators.

[0069] The cold source group control system forms a complete system through its own cloud platform, device manager, DDC controller, mobile monitoring software, computer monitoring software, etc. Computers, tablets, mobile phones and other terminal devices can be used to monitor various devices quickly and conveniently. The device manager has its own MODBUS, BACNET, OPC and other commonly used communication protocol drivers, which can easily manage third-party device systems; in addition, the monitoring interface of the cold source group control system uses pure HTML5, CSS and JAVASCRIPT general technologies, which can be directly embedded in other systems.

[0070] The cloud platform is a new generation of building energy-saving monitoring and management system platform based on TCP / IP. It has the characteristics of high reliability, strong compatibility, affordable price, and easy installation. It can easily meet the monitoring system requirements of various environments. When the cloud platform is used as an integrated server, multiple brands of control systems in the same or different projects can be integrated into the same management platform to achieve data interconnection. The configuration parameters of the cloud platform are as follows:

[0071] 1) Panel interface: 6 9-pin COM serial ports, 2 TCP / IP Gigabit Ethernet interfaces, 2 USB2.0 interfaces, 1 VGA interface, 1 HDMI interface, 1 POWER, RESET switch, 1 POWER_LED, HDD_LED indicator light, 1 WIFI antenna, DV working power interface, etc.;

[0072] 2) Directly access RS485 and RS323 communication interfaces;

[0073] 3) Working power supply 12DV, comes with adapter;

[0074] 4) Fanless closed cooling system, operating temperature 0~50℃, storage temperature -10~60℃, relative humidity 0-95%, no condensation;

[0075] 5) Overall dimensions: 172x107x54 (length x width x height), installation method: desktop or wall-mounted;

[0076] 6) Other parameters: 1.86GHz dual-core processor; Intel NM10 high-speed chipset; DDR3 memory, support 4GB; 1 2.5-inch SATA hard disk slot and 1 MSATA slot; etc.

[0077] 7) With operating system and AI cold source group control client software;

[0078] 8) Wireless routing function can be realized through partial configuration;

[0079] 9) Supports WEB browser access, just enter the management server IP address in the address bar (the IP address can be freely set by the user).

[0080] 11) The cold source group control system client can be freely developed according to project requirements. The main functions include: various communication protocol drivers (Modbus, BACnet, OPC, etc.), flexible equipment monitoring point browsing and collection functions, powerful configuration page function production, configuration diagram editing is all completed in the form of pictures; etc.

[0081] 12) Real-time reflection of device network connection and data value changes; more flexible history record setting function.

[0082] 13) Able to complete the logic control and data collection of mainstream building automation systems in the market (such as Honeywell, Trend, SIEMENS, etc.), chillers (Carrier, McQuay, etc.), VAV air-conditioning systems, networked fan coil systems, smart meters (smart electricity meters, smart water meters, etc.), networked actuators / sensors, etc.

[0083] The cold source group control system monitors and controls the central air conditioning refrigeration system including the refrigeration host, freezing pump, cooling pump, cooling tower, electric valve, etc. The monitoring and control parameters are shown in Table 1:

[0084]

[0085]

[0086] Table 1 Monitoring and control parameters of cold source group control system

[0087] In one embodiment, if Figure 1 As shown, the present application discloses a cold source group control method based on artificial intelligence, which specifically includes the following steps:

[0088] S10: Acquire comfort prediction information based on a preset comfort prediction model.

[0089] In this embodiment, the comfort prediction model refers to a model used to predict the comfort of the user under different cold source terminal parameters.

[0090] Specifically, the operation and management personnel of the cold source group control system can set a control cycle within a predetermined time period by themselves. The control cycle refers to the cycle of controlling the cold source group control system based on artificial intelligence analysis of the current load situation. For example, it can be set to 30 minutes, 45 minutes or 60 minutes. Therefore, when it comes to the control cycle, based on the preset comfort prediction model for predicting the user's comfort under different cold source terminal parameters, combined with the current time point and the collected data, the user's next comfort situation is predicted to obtain comfort prediction information. The collected data types include data of the same type as the training sample data of the comfort prediction model, such as user feedback data, environmental parameter data, user behavior data, and physiological parameter data.

[0091] S20: Based on a preset cooling capacity prediction model, obtain cooling capacity prediction information.

[0092] In this embodiment, the cooling capacity prediction model refers to a model used to predict the cooling capacity demand of users using the cooling source group control system.

[0093] Specifically, based on a preset cooling capacity prediction model for predicting the cooling capacity demand of users using the cooling source group control system, combined with the current time point and the collected data, the user's next cooling capacity demand of the refrigeration system is predicted to obtain cooling capacity prediction information. The collected data types include the same data types as the training sample data types of the cooling capacity prediction model, such as building and facility data, environmental factor data, population density data, etc.

[0094] S30: Analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information based on the determination result.

[0095] Specifically, user comfort and cooling demand are both parameter information representing user needs, and are both related to user experience. However, user comfort and cooling demand are not equal. Even if different users have the same cooling demand, their user comfort may be different. Conversely, different users may have different cooling demands when their user comfort is the same. Therefore, the comfort prediction information and the cooling prediction information are analyzed together to determine whether the cooling capacity represented by the cooling prediction information is equivalent to the user comfort represented by the comfort prediction information during actual operation, and whether the energy consumption required for the cooling capacity represented by the cooling prediction information meets the preset energy-saving index. Based on the judgment result, the parameter model input information is obtained. For example, by judging When it is necessary to achieve the user comfort level indicated by the comfort prediction information, the required cooling capacity of the refrigeration system is compared with the cooling capacity indicated by the cooling capacity prediction information to determine whether they are equal, and the cooling capacity indicated by the cooling capacity prediction information is adjusted according to the judgment result so that the user's comfort needs can be met. In addition, the energy consumption of the refrigeration system is calculated by simulating and predicting the cooling capacity indicated by the adjusted cooling capacity prediction information, and the simulated predicted energy consumption is compared with the preset energy-saving index to determine whether the preset energy-saving index can be met. If it is determined that it cannot be met, the cooling capacity is adjusted accordingly according to the comparison result to obtain the cooling capacity that can meet the user's comfort needs and meet the preset energy-saving index, and the parameter model input information is obtained.

[0096] It should be noted that, when simulating and predicting the cooling capacity represented by the adjusted cooling capacity prediction information after manufacturing adjustment, the energy consumption of the refrigeration system is simulated according to the various parameters of the refrigeration system under the normal state before the parameter adjustment, that is, the normal state without optimizing the refrigeration system. The simulation prediction is for preliminarily judging the energy consumption value. Therefore, the simulated predicted energy consumption is the energy consumption data under the normal state, and the preset energy-saving index is also the energy consumption range of the refrigeration system under the normal state, that is, there is a margin for comparing the energy consumption data of the refrigeration system under the normal state before the parameter adjustment. The energy consumption of the refrigeration system after the parameter adjustment will be lower than the energy consumption of the refrigeration system before the parameter adjustment. Therefore, the energy consumption of the refrigeration system after the parameter adjustment is lower than the lower limit of the energy consumption range of the preset energy-saving index, and the energy consumption of the refrigeration system before the parameter adjustment only needs to be within the energy consumption range of the preset energy-saving index.

[0097] S40: Inputting the parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model.

[0098] In this embodiment, the hierarchical fusion model refers to a model used to determine the optimal control scheme of the cold source group control system under different user requirements.

[0099] Specifically, the parameter model input information is input into a preset hierarchical fusion model, and the hierarchical fusion model is used to determine the model of the preferred control scheme of the cold source group control system under different user needs. Therefore, the hierarchical fusion model outputs at least one cold source group control system control scheme, and the parameter model input information is used as the input information of the hierarchical fusion model. Through the hierarchical fusion model, the preferred control scheme for the refrigeration system under the cooling capacity represented by the parameter model input information is determined, thereby obtaining the preferred cold source group control system control scheme, that is, the parameter control information.

[0100] In this embodiment, the hierarchical fusion model divides the overall refrigeration system into multiple layers based on the monitoring and control objects of the cold source group control system. For example, the chiller and related equipment are divided into one layer, the terminal air-conditioning equipment is divided into one layer, and the refrigeration object equipment is divided into one layer. The energy efficiency and load conditions in the refrigeration system of each layer are judged respectively under the cooling capacity represented by the parameter model input information, so as to find the most optimal control scheme for the cold source group control system.

[0101] S50: Controlling the cold source group control system based on the parameter control information.

[0102] Specifically, after the parameter control information is obtained, the parameter control information is sent to the cold source group control system to control the cold source group control system, thereby achieving optimization of the refrigeration system.

[0103] In this embodiment, the parameter control information may include the number of chillers in operation, the order of chiller operation, the switch control of the butterfly valve, the frequency of the chilled water pump, the start and stop control of the chilled water pump, the start and stop control of the cooling tower, the start and stop control of the cooling water pump, the frequency of the cooling water pump, the opening of the electric valve on the primary side of the plate heat exchanger, etc.

[0104] In addition, the cold source group control system uploads 0-1000 characters every 10 seconds. When the cold source group control method based on artificial intelligence is used, 0-1000 characters are sent to the cold source group control system every 10 seconds. When the cold source group control system receives characters without change for 10 consecutive times, it automatically exits the energy-saving mode of the cold source group control method based on artificial intelligence, and operates in the normal control mode, and reports the communication fault. When the characters received are different for 10 times, the communication fault is automatically restored, and the system operates in the energy-saving control mode again.

[0105] Normal control mode: The cold source group control system automatically executes all control logic, including equipment addition and subtraction, speed adjustment, cooling mode switching, bypass, charging and discharging, etc., and the DCIM monitors the status information;

[0106] Energy-saving control mode: The cold source group control system accepts the cold source group control method based on artificial intelligence, and issues instructions to complete specified actions. The instruction package includes the number of equipment, speed / power / temperature / pressure difference and other control loop target values, cooling mode switching, charging and discharging control, etc. The group control system acts according to the instructions issued by the energy-saving algorithm.

[0107] In one embodiment, if Figure 2 As shown in Figure 2, the comfort prediction model is obtained by:

[0108] S01: Acquire user comfort reference data, where the user comfort reference data includes user feedback data, environmental parameter data, user behavior data, and physiological parameter data.

[0109] In this embodiment, the user comfort reference data refers to reference data used to determine the user comfort.

[0110] Specifically, user comfort reference data is obtained, which includes user feedback data, environmental parameter data, user behavior data and physiological parameter data, wherein user feedback data includes comfort ratings or feedback directly provided by users, such as obtained through questionnaires, online evaluations or satisfaction ratings within applications; environmental parameter data includes environmental parameters closely related to user comfort, such as temperature, humidity, wind speed and air pressure; user behavior data includes user behavior under cooling environment conditions, such as turning on and off air conditioners, adjusting clothing and using fans, which can indirectly reflect the user's comfort perception; physiological parameter data includes heart rate, body temperature and skin humidity, which can directly reflect the user's physiological comfort. It should be noted that each of the above user comfort reference data is associated with the cooling capacity data of the cooling system at the same time.

[0111] S02: Perform model training based on user comfort reference data to generate a comfort prediction model.

[0112] Specifically, the user comfort reference data is cleaned, screened and standardized to ensure accuracy, completeness and consistency. In addition, the number of samples of different categories (comfortable / uncomfortable) in the training samples is balanced to avoid excessive bias of the model towards a certain category. The processed user comfort reference data is used for model training to generate a comfort prediction model.

[0113] like Figure 3 As shown in the figure, the cooling capacity prediction model is obtained by:

[0114] S03: Acquire cooling demand reference data, which includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data, and energy price data.

[0115] In this embodiment, the cooling capacity demand reference data refers to reference data used to determine the user's demand for cooling capacity.

[0116] Specifically, obtain cooling demand reference data, which includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data and energy price data, among which building and facility data include data such as the type, area and usage of buildings and facilities; environmental factor data include data such as temperature, humidity, season and weather conditions (such as sunny, rainy, cloudy) that have a significant impact on cooling demand; time series data is used to represent the time series characteristics of user cooling demand, that is, different demand patterns are shown at different time points. It should be noted that each of the above cooling demand reference data is associated with the cooling capacity data of the refrigeration system at the same time.

[0117] S04: Perform model training based on cooling demand reference data to generate a cooling prediction model.

[0118] Specifically, the cooling demand reference data is cleaned, screened and standardized to ensure the accuracy, completeness and consistency of the user comfort reference data. In addition, the number of training samples covering different scenarios and conditions is balanced to avoid excessive bias of the model towards a certain category. The processed cooling demand reference data is used for model training to generate a cooling prediction model.

[0119] In one embodiment, if Figure 4 As shown, the hierarchical fusion model is obtained by:

[0120] S001: Acquire physical facility data, and establish a facility physical model based on the physical facility data.

[0121] In this embodiment, the physical facility data refers to the relevant data of the physical facilities of the building currently using the cold source group control system.

[0122] Specifically, physical facility data is obtained. In this embodiment, the physical facility data includes relevant data of each device in the cold source group control system, relevant data of each device in the refrigeration system, relevant data of user devices on the refrigeration terminal, relevant data of the refrigeration area of ​​each refrigeration terminal, building sensor data and user data of the refrigeration area of ​​each refrigeration terminal, etc., that is, the physical facility data refers to the relevant data of each load-side object on the transmission path of the cold load, for example, the model specification of each device, the node name and the load data of the device during operation, and the number of users in the refrigeration area of ​​each refrigeration terminal, etc. The physical facility data also includes parent-child management node relationships, for example, the association relationship between the supervision of each device in the cold source group control system and the equipment of the refrigeration system supervised, the association relationship between the cooling water circulation system in the refrigeration system and the equipment in the cooling water circulation system, etc.; therefore, on the basis of the parent-child management node relationship, a facility physical model is established through the relevant data of each physical facility, and the facility physical model is used to represent the association relationship and the supervision relationship of each physical facility of the building using the cold source group control system.

[0123] S002: Obtain historical load data, determine the load difference based on the historical load data, and obtain load stratification information.

[0124] In this embodiment, the historical load data refers to the historical load data for each node of the facility physical model.

[0125] Specifically, based on each load-side node on the transmission path of the cold load represented by the facility physical model, the load data of each load-side node on the transmission path of each cold load, that is, the historical load data, is obtained. On the transmission path of the cold load, not every node will have load consumption transfer. When analyzing the optimal control parameters of the cold source group control system under different user needs, if there is no actual load consumption transfer, there is no need to consider the load situation at this node. Therefore, the historical load data is analyzed to determine the nodes where actual load changes have occurred between each associated node and can cause energy consumption changes, and the nodes are extracted to obtain load stratification information.

[0126] S003: Establish a load model based on load stratification information.

[0127] Specifically, based on the load stratification information indicating the nodes where actual load changes may lead to energy consumption changes, a load model is established using the historical load data corresponding to the node. The load model represents the load changes when the actual load changes on the transmission path of the cold load lead to energy consumption changes.

[0128] S004: Obtain adjustable parameter information, integrate the facility physical model and the load model, use the adjustable parameter information as model input, and generate a hierarchical fusion model.

[0129] Specifically, after obtaining a load model that represents the load change situation when the actual load change on the transmission path of the cold load leads to a change in energy consumption, and a facility physical model that represents the association relationship and supervision relationship of each physical facility on the transmission path of the cold load, the facility physical model and the load model are fused to obtain a model that represents the load change situation of each physical facility on the transmission path of the cold load, and adjustable parameter information is obtained, that is, the equipment parameters in the refrigeration system that can be adjusted by the cold source group control system. The adjustable parameter information is used as the input of the model that represents the load change situation of each physical facility on the transmission path of the cold load, and the historical adjustment parameters, historical load data and physical facility data corresponding to the adjustable parameter information are used to train the model to generate a hierarchical fusion model.

[0130] In one embodiment, if Figure 5 As shown, in step S004, the adjustable parameter information is obtained, the facility physical model and the load model are integrated, and the adjustable parameter information is used as the model input to generate a hierarchical fusion model, which specifically includes:

[0131] S0041: Obtain adjustable parameter information and historical parameter adjustment information, and use the historical parameter adjustment information and historical load data as training data sets.

[0132] Specifically, adjustable parameter information and historical parameter adjustment information corresponding to the adjustable parameter information are obtained, and the historical parameter adjustment information and historical load data are used as training data sets.

[0133] S0042: Based on a preset model fusion algorithm, the facility physical model and the load model are integrated to generate a fusion model. The model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross-fusion algorithm.

[0134] Specifically, based on a preset model fusion algorithm, the facility physical model and the load model are integrated. In this embodiment, the model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross-fusion algorithm. When integrating the facility physical model and the load model, the order of using the model fusion algorithm is to first use the physical control fitting algorithm, then the linear fitting algorithm, and finally the cross-fusion algorithm. In this embodiment, when the physical control fitting algorithm is used to fuse the facility physical model and the load model, each node is associated based on the control relationship represented by the facility physical model, and then the load data represented by the load model is integrated, and the load data of the corresponding association relationship is assigned a data association relationship based on the association relationship, so that each physical facility and load data on the transmission path of the cold load have an association relationship; when the linear fitting algorithm is used for fusion, the linear change trend of the load of each node on the transmission path of the cold load represented by the load model is considered, as well as the control linear relationship between each physical device represented by the facility physical model and the controlled device, and then the linear change trend of the load of each node on the transmission path of the cold load represented by the load model and the control linear relationship between each physical device represented by the facility physical model and the controlled device are fitted; when the cross-fusion algorithm is used for fusion, in addition to considering the control linear relationship between each physical device and the controlled device, it will further consider whether there is a correlation between each physical device represented by the fitted facility physical model and each other device, and cross-associate the corresponding devices based on the corresponding correlation relationship.

[0135] S0043: Use the adjustable parameter information as model input, train the fusion model based on the training data set, and generate a hierarchical fusion model.

[0136] Specifically, the adjustable parameter information is used as the model input, and the fusion model is trained based on the training data set. When the fusion model is trained, it is trained based on the association and correlation relationships in the fusion model to generate a hierarchical fusion model.

[0137] In one embodiment, if Figure 6 As shown, in step S30, the comfort prediction information and the cooling capacity prediction information are analyzed to determine whether the prediction information meets the preset energy-saving index. According to the judgment result, the parameter model input information is obtained, which specifically includes:

[0138] S31: Convert the comfort level prediction information into comfort level cooling capacity prediction information.

[0139] Specifically, the comfort prediction information output by the comfort prediction model represents the user's future comfort information, and the comfort prediction information is relatively abstract information and cannot directly represent the user's needs. Therefore, the comfort prediction information is converted into comfort cooling prediction information. For example, the correlation between the user's comfort and the user's cooling demand is determined through the training data of the comfort prediction model and the cooling prediction model. The comfort prediction information is converted into comfort cooling prediction information through the correlation between the user's comfort and the user's cooling demand to obtain the comfort cooling prediction information.

[0140] S32: Compare the comfort cooling capacity prediction information and the cooling capacity prediction information, and obtain high comfort predicted cooling capacity information according to the comparison result.

[0141] Specifically, in order to achieve the goal of meeting the user's comfort requirements, it is necessary to use the cooling capacity prediction information that represents a higher user comfort level as much as possible while meeting the energy-saving indicators. Therefore, first obtain the preset cooling capacity information with the highest user comfort level at the current time node. When the ambient temperature is high, the higher of the comfort cooling capacity prediction information and the cooling capacity prediction information, that is, the higher cooling capacity, will be able to improve the user's comfort. When the ambient temperature is moderate, it is necessary to consider the gap between the suitable temperature of the human body under the current ambient temperature and the cooling capacity. The smaller the gap, the more it can improve the user's comfort. Therefore, compare the comfort cooling capacity prediction information with the cooling capacity prediction information, and select the cooling capacity prediction information that can better improve the user's comfort from the comfort cooling capacity prediction information and the cooling capacity prediction information, so as to obtain high-comfort predicted cooling capacity information.

[0142] S33: Determine whether the high comfort predicted cooling capacity information meets the preset energy-saving index, and obtain parameter model input information based on the determination result.

[0143] Specifically, after the purpose of meeting the user's comfort requirements is achieved, it is still necessary to determine whether the energy-saving indicators are met. Therefore, it is determined whether the high comfort predicted cooling capacity information meets the preset energy-saving indicators. If it is determined to be satisfied, the high comfort predicted cooling capacity information is used as parameter model input information. If it is determined not to be satisfied, the comfort cooling capacity prediction information and the other predicted cooling capacity information in the cooling capacity prediction information that is not the high comfort predicted cooling capacity information are used as parameter model input information.

[0144] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0145] In one embodiment, a cold source group control device based on artificial intelligence is provided, and the cold source group control device based on artificial intelligence corresponds to the cold source group control method based on artificial intelligence in the above embodiment. Figure 7 As shown, the cold source group control device based on artificial intelligence includes a comfort prediction module, a cooling capacity prediction module, an input analysis module, a model analysis module and a control module. The detailed description of each functional module is as follows:

[0146] A comfort prediction module, used for obtaining comfort prediction information based on a preset comfort prediction model;

[0147] A cooling capacity prediction module is used to obtain cooling capacity prediction information based on a preset cooling capacity prediction model;

[0148] An input analysis module is used to analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information based on the judgment result;

[0149] A model analysis module, used to input parameter model input information into a preset hierarchical fusion model, and obtain parameter control information based on the hierarchical fusion model;

[0150] The control module is used to control the cold source group control system based on parameter control information.

[0151] Optionally, the cold source group control device based on artificial intelligence also includes:

[0152] A comfort data acquisition module is used to acquire user comfort reference data, which includes user feedback data, environmental parameter data, user behavior data and physiological parameter data;

[0153] The comfort model generation module is used to train the model based on the user comfort reference data and generate a comfort prediction model; the cooling demand data acquisition module is used to obtain the cooling demand reference data, which includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data and energy price data;

[0154] The cooling demand model generation module is used to perform model training based on cooling demand reference data and generate a cooling prediction model.

[0155] Optionally, the cold source group control device based on artificial intelligence also includes:

[0156] A physical model generation module is used to obtain physical facility data and establish a facility physical model based on the physical facility data;

[0157] The load stratification analysis module is used to obtain historical load data, determine the load difference based on the historical load data, and obtain load stratification information;

[0158] A load model generation module is used to establish a load model based on load stratification information;

[0159] The fusion model module is used to obtain adjustable parameter information, integrate the facility physical model and the load model, use the adjustable parameter information as model input, and generate a hierarchical fusion model.

[0160] Optionally, the fusion model module includes:

[0161] A data set generation submodule is used to obtain adjustable parameter information and historical parameter adjustment information, and use the historical parameter adjustment information and historical load data as a training data set;

[0162] A fusion submodule is used to generate a fusion model by fusing the facility physical model and the load model based on a preset model fusion algorithm. The model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross fusion algorithm.

[0163] The training submodule is used to take the adjustable parameter information as the model input, train the fusion model based on the training data set, and generate a hierarchical fusion model.

[0164] Optionally, the input analysis module includes:

[0165] A cooling capacity conversion submodule is used to convert the comfort level prediction information into the comfort level cooling capacity prediction information;

[0166] The cooling capacity comparison submodule is used to compare the comfort cooling capacity prediction information and the cooling capacity prediction information, and obtain the high comfort predicted cooling capacity information according to the comparison result;

[0167] The energy-saving judgment module is used to judge whether the high-comfort predicted cooling capacity information meets the preset energy-saving indicators, and obtain the parameter model input information based on the judgment result.

[0168] For the specific definition of the cold source group control device based on artificial intelligence, please refer to the definition of the cold source group control method based on artificial intelligence above, which will not be repeated here. Each module in the above-mentioned cold source group control device based on artificial intelligence can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0169] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store comfort prediction information, cooling prediction information, parameter model input information and parameter control information, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a cold source group control method based on artificial intelligence is implemented.

[0170] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0171] Based on a preset comfort prediction model, obtain comfort prediction information;

[0172] Based on the preset cooling capacity prediction model, obtain cooling capacity prediction information;

[0173] Analyze the comfort prediction information and cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information based on the judgment result;

[0174] Inputting parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model;

[0175] Based on the parameter control information, the cold source group control system is controlled.

[0176] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0177] Based on a preset comfort prediction model, obtain comfort prediction information;

[0178] Based on the preset cooling capacity prediction model, obtain cooling capacity prediction information;

[0179] Analyze the comfort prediction information and cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information based on the judgment result;

[0180] Inputting parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model;

[0181] Based on the parameter control information, the cold source group control system is controlled.

[0182] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0183] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0184] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A cold source group control method based on artificial intelligence, characterized in that: The cold source group control method based on artificial intelligence includes: Based on a preset comfort prediction model, obtain comfort prediction information; Based on the preset cooling capacity prediction model, obtain cooling capacity prediction information; Analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information according to the determination result; The parameter model input information is input into a preset hierarchical fusion model, and parameter control information is obtained based on the hierarchical fusion model, wherein the hierarchical fusion model is obtained in the following manner: Acquire physical facility data, and establish a facility physical model based on the physical facility data; Acquire historical load data, determine load drop according to the historical load data, and obtain load stratification information; Based on the load stratification information, establishing a load model; Acquiring adjustable parameter information, fusing the facility physical model and the load model, using the adjustable parameter information as a model input, and generating a hierarchical fusion model, specifically including: Acquire adjustable parameter information and historical parameter adjustment information, and use the historical parameter adjustment information and the historical load data as a training data set; Based on a preset model fusion algorithm, the facility physical model and the load model are fused to generate a fusion model, wherein the model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross fusion algorithm; Using the adjustable parameter information as model input, training the fusion model based on the training data set, and generating a hierarchical fusion model; Based on the parameter control information, the cold source group control system is controlled.

2. The cold source group control method based on artificial intelligence according to claim 1 is characterized in that: The comfort prediction model is obtained in the following way: Acquire user comfort reference data, wherein the user comfort reference data includes user feedback data, environmental parameter data, user behavior data and physiological parameter data; Performing model training based on the user comfort reference data to generate a comfort prediction model; The cooling capacity prediction model is obtained by: Acquiring cooling demand reference data, the cooling demand reference data including historical cooling demand data, building and facility data, time series data, environmental factor data, population density data, and energy price data; Model training is performed based on the cooling demand reference data to generate a cooling prediction model.

3. The cold source group control method based on artificial intelligence according to claim 1 is characterized in that: The combining of the comfort prediction information and the cooling capacity prediction information for analysis to determine whether the prediction information meets a preset energy-saving index and obtaining parameter model input information according to the determination result specifically includes: Converting the comfort level prediction information into comfort level cooling capacity prediction information; Comparing the comfort cooling capacity prediction information with the cooling capacity prediction information, and obtaining high comfort predicted cooling capacity information according to the comparison result; It is determined whether the high comfort predicted cooling capacity information meets the preset energy-saving index, and according to the determination result, parameter model input information is obtained.

4. A cold source group control device based on artificial intelligence, characterized in that: The cold source group control device based on artificial intelligence includes: A comfort prediction module, used for obtaining comfort prediction information based on a preset comfort prediction model; A cooling capacity prediction module is used to obtain cooling capacity prediction information based on a preset cooling capacity prediction model; An input analysis module is used to analyze the comfort prediction information and the cooling capacity prediction information to determine whether the prediction information meets the preset energy-saving index, and obtain parameter model input information according to the determination result; A model analysis module, used for inputting the parameter model input information into a preset hierarchical fusion model, and obtaining parameter control information based on the hierarchical fusion model; A physical model generation module, used to obtain physical facility data and establish a facility physical model based on the physical facility data; A load stratification analysis module is used to obtain historical load data, determine the load drop according to the historical load data, and obtain load stratification information; A load model generation module, used to establish a load model based on the load stratification information; A fusion model module, used to obtain adjustable parameter information, fuse the facility physical model and the load model, use the adjustable parameter information as a model input, and generate a hierarchical fusion model; A data set generation submodule is used to obtain adjustable parameter information and historical parameter adjustment information, and use the historical parameter adjustment information and historical load data as a training data set; A fusion submodule is used to generate a fusion model by fusing the facility physical model and the load model based on a preset model fusion algorithm. The model fusion algorithm includes a linear fitting algorithm, a physical control fitting algorithm and a cross fusion algorithm. A training submodule, used for taking the adjustable parameter information as model input, training the fusion model based on the training data set, and generating a hierarchical fusion model; The control module is used to control the cold source group control system based on the parameter control information.

5. The cold source group control device based on artificial intelligence according to claim 4 is characterized in that: The cold source group control device based on artificial intelligence also includes: A comfort data acquisition module, used to acquire user comfort reference data, wherein the user comfort reference data includes user feedback data, environmental parameter data, user behavior data and physiological parameter data; A comfort model generation module, used for performing model training based on the user comfort reference data to generate a comfort prediction model; A cooling demand data acquisition module is used to acquire cooling demand reference data, wherein the cooling demand reference data includes historical cooling demand data, building and facility data, time series data, environmental factor data, population density data and energy price data; The cooling demand model generation module is used to perform model training based on the cooling demand reference data to generate a cooling prediction model.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the cold source group control method based on artificial intelligence as described in any one of claims 1 to 3 are implemented.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the cold source group control method based on artificial intelligence as claimed in any one of claims 1 to 3 are implemented.

Images