Large central air conditioner multi-model optimization control method based on final target temperature

Abstract

The invention relates to a final target temperature-based large central air conditioner multi-model optimization control method, which relates to the technical field of air conditioner system control, and comprises the following steps: S1, locally constructing physical models of an actual building and a central air conditioner; s2, establishing a corresponding control strategy designed for each scene in a multi-scene simulation environment; s3, acquiring current environmental parameters of the actual building, and inputting the parameters into the physical model; s4, a real-time control feedback mechanism is established, and the current indoor temperature and the current external environment change are continuously obtained in each time step length; s5, inputting the output data of the central air conditioner and the environmental parameters into the actual building physical model to predict the temperature in the actual building; and S6, all the models are iterated through an optimization algorithm, the working efficiency of the central air conditioner can be remarkably improved through an optimization control strategy, meanwhile, a real-time feedback mechanism can quickly respond to environment changes, and it is ensured that the indoor temperature meets the comfort standard all the time.

Description

Technical Field

[0001] The present application relates to the technical field of air-conditioning system control, and in particular to a multi-model optimization control method for large central air-conditioning based on a final target temperature. Background Art

[0002] The comfort of the indoor environment directly affects people's work efficiency and quality of life. Therefore, it is crucial to regulate and maintain appropriate indoor temperature, humidity and air quality. Among these factors, temperature is one of the key parameters affecting comfort. Large buildings are usually equipped with central air-conditioning systems, which usually consist of multiple components such as chillers, fans, cooling towers, and piping systems. Its operation is affected by many factors, including the design of the building, external climate, indoor heat load, etc.

[0003] In modern buildings, large central air-conditioning systems are one of the key technologies to ensure indoor environmental comfort and energy efficiency. However, traditional central air-conditioning control systems have the following major problems: 1. Lack of real-time data support: Many existing systems cannot obtain key parameters such as external environmental temperature and humidity in real time, resulting in the inability to adjust the control strategy in a timely manner; 2. Single model limitation: Most traditional systems rely on a single mathematical model for prediction and control, which is difficult to adapt to complex and changeable actual environmental conditions; 3. Energy waste: Due to the inaccurate control strategy, it often leads to excessive energy consumption and increased operating costs; 4. Poor user experience: It is impossible to flexibly adjust the temperature according to the needs of different areas, affecting the user's comfort.

[0004] In view of the above-mentioned related technologies, a solution is now proposed. Summary of the invention

[0005] The purpose of this application is to provide a multi-model optimization control method for large central air conditioners based on the final target temperature to solve the technical problems in the prior art such as lack of real-time data support, single model limitation, energy waste and poor user experience.

[0006] The large-scale central air conditioning multi-model optimization control method based on the final target temperature provided in this application adopts the following technical solutions: The multi-model optimization control method for large central air conditioning based on the final target temperature includes the following steps: S1. Build physical models of actual buildings and central air conditioners locally; S2. Establish a multi-scenario simulation environment and design corresponding control strategies for each scenario; S3, obtaining current environmental parameters of the actual building, and inputting the parameters into the physical model; S4. Establish a real-time control feedback mechanism to continuously obtain the current indoor temperature and external environment changes in each time step; S5, entering the central air-conditioning output data and the environmental parameters into the actual building physical model to predict the temperature in the actual building; S6. All models are iterated using an optimization algorithm to determine the best control curve for prediction.

[0007] By adopting the above technical solutions, a physical model of the actual building and central air-conditioning system is established locally. The model should take into account factors such as the building's geometry, material properties, window location and its thermal insulation performance, and reflect the performance parameters of the central air-conditioning, such as cooling / heating capacity, air flow path, etc., to ensure that the model is highly consistent with the actual scene. Next, a multi-scenario simulation environment is established based on the physical model. Design corresponding control strategies for different scenarios such as different seasons and different weather conditions. These strategies should cover temperature regulation, humidity control and air circulation to adapt to different user needs and environmental changes, and obtain the current environmental parameters of the actual building in real time, including indoor temperature, humidity, wind speed, external temperature, etc. These parameters are input into the constructed physical model to provide basic data for subsequent control decisions, establish a real-time control feedback mechanism, and continuously monitor the current indoor temperature and external environmental changes in each time step. Through high-frequency data collection and processing, ensure that the system can respond to environmental changes in a timely manner and maintain indoor comfort. Then, based on the output data of the central air-conditioning and the current environmental parameters, this information is entered into the physical model to predict the temperature changes in the actual building. This process requires combining historical data with current status, using algorithms to accurately predict temperature, and finally, iterating all models using optimization algorithms to determine the best control curve based on the prediction results. The optimization algorithm can combine a variety of modern optimization techniques such as genetic algorithms and particle swarm optimization to achieve the best balance between maximum energy efficiency and user comfort.

[0008] Preferably, in step S1, a physical model of the actual building is constructed locally, and the construction steps are as follows: A1. Establish a building heat conduction model: ; Wherein, T represents temperature distribution; α represents thermal diffusivity; Q represents heat source term, and the central air conditioning temperature distribution in the actual building is established by the formula; A2. Space and time discretization: Use finite difference methods to discretize space into a grid of points: Define grid spacing , and , represents the spacing on the x, y, and z axes; The temperature distribution is discretized into the temperature values ​​of the grid nodes , and , where i, j, k represent the spatial index of the node; Using an explicit method leads to discretization of time: ; The time step is , the time discretization is ; Temperature discreteness is , represents the temperature value at the nth moment; A3. Combining the heat conduction equation and the discretization result, the discrete equation group is obtained as follows: ; A4. Set boundary conditions according to the actual situation of the building, where the fixed temperature boundary conditions are: ; The fixed heat flow boundary conditions are: ; Initialize the temperature distribution throughout the actual building: ; A5. Step by step solve the temperature distribution of each time step according to the discrete equations ,when When it is less than a set threshold, the iteration and solution are stopped to obtain the physical model of the actual building at different times and spaces.

[0009] By adopting the above scheme, based on detailed heat conduction theory and discretization method, the high accuracy of the building temperature distribution model is ensured. By setting a variety of boundary conditions, the model can flexibly adapt to different building structures and usage scenarios. At the same time, combined with real-time data, the gradual solution mechanism can quickly respond to changes in the external environment and improve the system's responsiveness. The established model framework has good scalability and can easily integrate other influencing factors such as humidity and wind speed.

[0010] Preferably, in step S1, the central air conditioning physical model is constructed using the following expression: ; in, represents the total cooling load; k represents the thermal conductivity of the material; A represents the heat transfer area; Indicates the external temperature; represents the internal temperature; c represents the specific heat capacity of water; m represents the mass flow rate; Indicates the return water temperature; represents the water supply temperature; n represents the number of people; q represents the calorific value per person; represents the heat dissipation of electrical equipment; eff represents the system efficiency. The central air-conditioning physical model constructed by the expression can express the operating efficiency and energy consumption of the central air-conditioning.

[0011] By adopting the above scheme, multiple factors such as heat conduction, fluid dynamics and human physiological characteristics are comprehensively considered, which can fully reflect the performance and operation status of the air-conditioning system. The model can accurately calculate the cooling load and help designers and engineers to formulate reasonable plans in the initial design and operation adjustment process, so as to achieve the optimization of overall energy efficiency. By real-time input of data such as temperature inside and outside the building, heat dissipation of personnel and equipment, the model can provide data support for the regulation of central air-conditioning, making the decision more scientific and reasonable. The model can be adjusted and optimized according to different building types and functional requirements, and has good adaptability and flexibility. By optimizing the energy-saving effect obtained by the model, the central air-conditioning system can reduce energy consumption and reduce environmental burden while ensuring comfort, thereby helping to achieve sustainable development goals.

[0012] Preferably, in step S2, a machine learning method is used for different control scenarios to predict and evaluate temperature changes in different scenarios and integrate them into the optimization process. The steps of the machine learning method are as follows: D1. Data collection and processing: Collecting and cleaning the indoor and outdoor environmental parameters of the actual building; D2. Extract key features that affect indoor temperature from the collected data; D3. Use the computational model to make scene judgments on key features and output the predicted temperature; D4. Input the collected new environmental data into the trained model in real time to predict the indoor temperature at the next moment.

[0013] By adopting the above solution, machine learning can quickly process and analyze large amounts of data, improve temperature prediction efficiency, and support real-time decision-making. Through feature extraction and scenario judgment, it can improve the accuracy of temperature prediction and help optimize central air-conditioning control strategies. The constructed model can be adjusted for different control scenarios, adapt to changing environmental conditions, and enhance the flexibility of the system. The automated data processing and temperature prediction process reduces the need for manual intervention and improves the operational stability of the system. As new data is continuously collected, the machine learning model has the ability to continue learning, and can continuously optimize and improve prediction accuracy in practical applications.

[0014] Preferably, in step S3, the environmental parameters are obtained as follows: B1. Sensor deployment: installing sensors in key areas of the actual building; B2. Data collection: Transmit sensor data to the data processing center through wired and wireless methods; B3. Data preprocessing: cleaning the collected data; B4. Data analysis and feedback: Real-time monitoring of environmental parameters in the building, analysis of historical data through statistical analysis, and establishment of a model between environmental parameters and energy consumption. The model is expressed as follows: ; Among them, y represents energy consumption; Indicates environmental parameters; represents the model parameters; Denotes the error term, and the best estimate of the model parameters is determined by minimizing the sum of squared errors: ; Where m is the number of samples; represents the energy consumption of the i-th sample; represents the value of the ith environmental parameter, and through the above calculation, the environmental parameters are collected and optimized to establish an energy consumption prediction model.

[0015] By adopting the above solution and deploying sensors in key areas, comprehensive monitoring of environmental parameters can be achieved, the accuracy and timeliness of data collection can be improved, the preprocessing step ensures the high quality of the collected data, reduces the prediction error caused by data noise, and improves the credibility of the model. Through statistical analysis and regression modeling, the relationship between environmental factors and energy consumption can be accurately captured, providing a solid basis for subsequent algorithm optimization. The real-time monitoring and feedback mechanism of environmental parameters enables the energy consumption prediction model to respond quickly to changes, implement dynamic adjustment and optimize control strategies, and the model can be adjusted and optimized at any time according to new data input. It has good scalability and can gradually integrate more environmental variables and influencing factors to improve the adaptability of the model.

[0016] Preferably, in step S4, the output power and control strategy of the central air conditioner are dynamically adjusted by comparing the real-time data with the predicted data, and database services are used to process and store the real-time data, unify the format of data from different sources and perform standardized processing, and use the stream processing framework to analyze the real-time data.

[0017] By adopting the above scheme, and the technical effects and advantages of the above technical scheme, the design content that fits the above paragraph is described, with clear logic and standardized wording. By adjusting the output power of the air conditioner according to real-time needs, energy waste is reduced, the energy efficiency ratio of the system is improved, and contributions are made to sustainable development goals. The centralized management and unified format of data are realized through database services, which improves the convenience of data integration and reduces the complexity of subsequent data analysis. The real-time analysis capabilities of the application stream processing framework enable the system to quickly detect environmental changes, identify abnormal data, and take corresponding measures to improve the intelligence and adaptability of the system. The system can flexibly adjust the operation strategy according to different external environmental conditions, weather changes, personnel activities, etc., to ensure indoor comfort while maintaining efficient energy consumption management.

[0018] Preferably, in step S5, the data are input into the actual building physical model, so that the actual building model simulates the thermal environment changes inside the building according to the input data, and the temperature in the actual building is calculated and predicted using the finite element method. The finite element method is expressed by the heat conduction equation, and the steady-state heat conduction equation formula is as follows: ; The formula is combined with the transient heat conduction equation to describe the heat conduction equation, and the formula of the transient heat conduction equation is as follows: ; in, represents temperature; k represents the thermal conductivity of the material; Q represents the heat source term per unit volume; represents the gradient operator; is a divergence operator, which is expressed by the operation of the above equations and the discretized finite element method in matrix form as follows: ; in, represents the hot melt matrix; represents the thermal conductivity matrix; represents the temperature node vector; represents the heat source term vector, which is solved numerically by the formula to obtain the temperature distribution inside the building.

[0019] By adopting the above scheme and combining the steady-state and transient heat conduction equations, the complex heat conduction process in the building can be fully described, and the accuracy of the simulation can be improved. Through the discretization processing of the finite element method, the temperature distribution problem of large-scale buildings can be effectively solved, and real-time calculation under complex working conditions can be supported. The model can adapt to different building materials and structures, and allows the flexible introduction of various heat sources such as internal heating and external temperature changes, so that it can maintain effectiveness in a changing environment. The calculation results can display the temperature distribution inside the building through visualization technology, which is convenient for designers and engineers to intuitively analyze and optimize the building thermal environment. The real-time monitoring data can be combined with the model prediction to achieve dynamic adjustment and optimization, and improve the operating efficiency and energy utilization of the air-conditioning system.

[0020] Preferably, in step S6, the logic steps of the optimization algorithm are as follows: C1. According to the final target temperature Design objective function J: ; Controlling the input data as chromosome genes; C2. Randomly generate a set of control strategies represented by an encoding vector: ; in, represents the control input at time point n; C3. Evaluate the fitness of each individual in the population and use the objective function G to calculate the fitness value of each individual: ; C4. Select the better individuals for reproduction according to the fitness, and randomly select a pair of parents from the selected individuals to perform a crossover operation to produce offspring. The crossover operation is as follows: ; C5. Randomly mutate the offspring, the mutation method is: ; in, Represents the random perturbation value; the offspring generated by selection, crossover and mutation are placed into the new population C6. Repeat steps C3 to C5 until the termination condition is met. The termination condition is that the maximum number of generations is reached. Finally, the individual with the highest fitness is selected from the population as the optimal control strategy to predict the corresponding control input.

[0021] By adopting the above scheme and introducing evolutionary algorithms, the central air-conditioning control strategy can be intelligently optimized, energy consumption can be reduced and the indoor temperature comfort can be maintained. The algorithm can automatically adjust the control strategy according to real-time feedback, improve the system's adaptability to environmental changes, and actively respond to temperature fluctuations. The evolutionary algorithm has a strong global search capability, which can effectively avoid local optimal solutions and ensure that a control strategy close to the optimal is found. The algorithm design allows flexible definition and adjustment of the control input, is simple and easy to use, and is suitable for application in a variety of control scenarios. Through the selection, crossover and mutation operations in the genetic algorithm, better control effects can be achieved with less computational effort, thereby improving the system's operating efficiency.

[0022] In summary, the present application includes at least one of the following beneficial technical effects: 1. Optimizing control strategies can significantly reduce energy consumption and improve the working efficiency of central air conditioners; 2. The real-time feedback mechanism can quickly respond to environmental changes to ensure that the indoor temperature always meets the comfort standard; 3. Through multi-scenario simulation environment, the system can adapt to different usage scenarios, facilitating subsequent expansion and adjustment; 4. Based on the combination of physical models and real-time data, the prediction results are more accurate, reducing the impact of temperature fluctuations on comfort; 5. The use of optimization algorithms enables the control strategy to be continuously iterated and improved to adapt to the ever-changing environment and user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of the multi-model optimization control method of a large central air conditioner based on the final target temperature of the present application; Figure 2 It is a flowchart of the construction of the actual building physical model of this application; Figure 3 is a flow chart of the machine learning method of this application; Figure 4 It is the environmental parameter acquisition flow chart of this application; Figure 5 It is the logic flow chart of the optimization algorithm of this application. DETAILED DESCRIPTION

[0024] The following is combined with Figure 1 - Attachment Figure 5 , further details of this application are given.

[0025] The embodiments of the present application disclose a multi-model optimization control method for a large central air conditioner based on a final target temperature.

[0026] Reference Figure 1 , a large central air conditioning multi-model optimization control method based on the final target temperature includes the following steps: S1. Build physical models of actual buildings and central air conditioners locally; S2. Establish a multi-scenario simulation environment and design corresponding control strategies for each scenario; S3, obtaining the current environmental parameters of the actual building and inputting the parameters into the physical model; S4. Establish a real-time control feedback mechanism to continuously obtain the current indoor temperature and external environment changes in each time step; S5, inputting the central air-conditioning output data and environmental parameters into the actual building physical model to predict the temperature in the actual building; S6. All models are iterated using an optimization algorithm to determine the best control curve for prediction.

[0027] Specifically, first, a physical model of the actual building and central air-conditioning system is established locally. The model should take into account factors such as the building's geometry, material properties, window location and its thermal insulation performance, and reflect the performance parameters of the central air-conditioning, such as cooling / heating capacity, air flow path, etc., to ensure that the model is highly consistent with the actual scene; Next, a multi-scenario simulation environment is established based on the physical model. Design corresponding control strategies for different scenarios such as different seasons and different weather conditions. These strategies should cover temperature regulation, humidity control and air circulation to adapt to different user needs and environmental changes; obtain the current environmental parameters of the actual building in real time, including indoor temperature, humidity, wind speed, external temperature, etc. These parameters are input into the constructed physical model to provide basic data for subsequent control decisions; establish a real-time control feedback mechanism to continuously monitor the current indoor temperature and external environmental changes in each time step. Through high-frequency data collection and processing, ensure that the system can respond to environmental changes in a timely manner and maintain indoor comfort; based on the output data of the central air-conditioning and the current environmental parameters, enter this information into the physical model to predict the temperature changes in the actual building. This process requires combining historical data with current status and using algorithms to accurately predict temperature; finally, all models are iterated using optimization algorithms to determine the best control curve based on the prediction results. The optimization algorithm can be combined with a variety of modern optimization techniques such as genetic algorithms and particle swarm optimization to achieve the best balance between maximum energy efficiency and user comfort.

[0028] Reference Figure 2 In step S1, a physical model of the actual building is constructed locally. The construction steps are as follows: A1. Establish a building heat conduction model: ; Wherein, T represents temperature distribution; α represents thermal diffusivity; Q represents heat source term, and the central air conditioning temperature distribution in the actual building is established by the formula; A2. Space and time discretization: Use finite difference methods to discretize space into a grid of points: Define grid spacing , and , represents the spacing on the x, y, and z axes; The temperature distribution is discretized into the temperature values ​​of the grid nodes , and , where i, j, k represent the spatial index of the node; Using an explicit method leads to discretization of time: ; The time step is , the time discretization is ; Temperature discreteness is , represents the temperature value at the nth moment; A3. Combining the heat conduction equation and the discretization result, the discrete equation group is obtained as follows: ; A4. Set boundary conditions according to the actual situation of the building, where the fixed temperature boundary conditions are: ; The fixed heat flow boundary conditions are: ; Initialize the temperature distribution throughout the actual building: ; A5. Step by step solve the temperature distribution of each time step according to the discrete equations ,when When it is less than a set threshold, the iteration and solution are stopped to obtain the physical model of the actual building at different times and spaces.

[0029] Specifically, first, based on the heat conduction theory, a heat conduction model inside the building is established. Set T temperature distribution; α thermal diffusivity; Q heat source term, and use the corresponding heat conduction equation to describe the temperature distribution of the central air-conditioning system in the actual building. This equation can be expressed as a function of space and time, reflecting the changes in the thermal state of the building at different locations and times; The finite difference method is used to discretize the space into grid points for numerical calculation. The specific steps are as follows: Define grid spacing , and , represents the spatial discrete step length of the spacing on the x, y, and z axes; Discretize the temperature distribution into temperature values ​​at the mesh nodes , and , where i, j, k represent the spatial index of the node; For the time direction, an explicit method is used for discretization, and the time step is , the time discretization is ; Temperature discreteness is , represents the temperature value at the nth moment; Combining the heat conduction equation with the discretization results of space and time, a discrete set of equations is derived. This set of equations will be used for subsequent numerical solutions, usually in the following form: ; According to the actual situation of the building, set appropriate boundary conditions to reflect the impact of the real environment on the temperature distribution. The following types of boundary conditions can be set to gradually iterate and solve the temperature distribution of each time step according to the discrete equations. By setting the convergence threshold, the iteration is stopped when the current temperature change is less than the set threshold. This will ensure the accuracy of the solution so that the final building model can accurately reflect its temperature state at different times and spaces.

[0030] Reference Figure 1 ,In step S1, the central air conditioning physical model is constructed using the following expression: ; in, represents the total cooling load; k represents the thermal conductivity of the material; A represents the heat transfer area; Indicates the external temperature; represents the internal temperature; c represents the specific heat capacity of water; m represents the mass flow rate; Indicates the return water temperature; represents the water supply temperature; n represents the number of people; q represents the calorific value per person; represents the heat dissipation of electrical equipment; eff represents the system efficiency. The central air-conditioning physical model constructed by the expression can express the operating efficiency and energy consumption of the central air-conditioning.

[0031] Specifically, determine the material properties of the building to obtain the thermal conductivity k; measure and calculate each heat transfer area through the building design blueprint; install sensors to monitor the external and internal temperatures in real time; determine the mass flow rate and water supply temperature of cooling water through a water pump flow meter; count the number of people indoors and their heat generation to evaluate the heat dissipation of the equipment; substitute the collected data into the above expression for calculation to obtain the total cooling load; use computing software such as MATLAB, ANSYS, etc. to perform numerical simulation on the model to analyze the air conditioning performance and operating efficiency under different conditions; evaluate the system efficiency eff through simulation results and determine the energy efficiency ratio; adjust system parameters such as flow rate, water supply temperature, etc. according to the calculation results to optimize energy consumption and improve system performance.

[0032] Reference Figure 3 In step S2, the temperature changes in different control scenarios are predicted and evaluated by machine learning methods and integrated into the optimization process. The steps of the machine learning method are as follows: D1. Data collection and processing: Collect and clean the indoor and outdoor environmental parameters of the actual building; D2. Extract key features that affect indoor temperature from the collected data; D3. Use the computational model to make scene judgments on key features and output the predicted temperature; D4. Input the collected new environmental data into the trained model in real time to predict the indoor temperature at the next moment.

[0033] Specifically, first, indoor and outdoor environmental parameters such as temperature, humidity, wind speed, light intensity, etc. are collected from actual buildings. The collected data needs to be preprocessed, including cleaning, removing outliers and filling missing values ​​to ensure the quality and accuracy of the data. This process lays a solid foundation for subsequent analysis; based on the processed data, the key features that have a significant impact on indoor temperature changes are extracted. These features may include but are not limited to: external temperature fluctuations, the number of people indoors, the heat generated by electrical equipment, solar radiation, etc. The most representative and influential features are selected from the entire data set using feature selection algorithms or correlation analysis techniques to improve the prediction accuracy of the model; computational models such as regression models, decision trees, support vector machines, etc. are used to perform scene judgment on the extracted key features to output the predicted temperature. At this stage, by training the model, the relationship between the features and the indoor temperature is learned, so that the model can automatically infer the corresponding temperature changes when facing different scenarios. At the same time, cross-validation and other techniques are used to ensure the generalization ability and robustness of the model; in the implementation of the optimization process, the newly collected environmental data needs to be input into the trained machine learning model in real time to predict the indoor temperature at the next moment. Through this process, it is possible to respond quickly to environmental changes and ensure that the central air-conditioning system can be adjusted in time to meet the set comfort standards.

[0034] Reference Figure 4 In step S3, the environmental parameters are obtained as follows: B1. Sensor deployment: Install sensors in key areas of actual buildings; B2. Data collection: Transmit sensor data to the data processing center through wired and wireless methods; B3. Data preprocessing: cleaning the collected data; B4. Data analysis and feedback: Real-time monitoring of environmental parameters in the building, analysis of historical data through statistical analysis, and establishment of a model between environmental parameters and energy consumption. The model is expressed as follows: ; Among them, y represents energy consumption; Indicates environmental parameters; represents the model parameters; Denotes the error term, and the best estimate of the model parameters is determined by minimizing the sum of squared errors: ; Where m is the number of samples; represents the energy consumption of the i-th sample; represents the value of the ith environmental parameter, and through the above calculation, the environmental parameters are collected and optimized to establish an energy consumption prediction model.

[0035] Specifically, a variety of sensors are installed in key areas of actual buildings (such as air-conditioning vents, near windows, in the center of the room, etc.). These sensors may include temperature sensors, humidity sensors, carbon dioxide concentration sensors, and light sensors to comprehensively monitor environmental parameters and provide a basis for subsequent data analysis; the data collected by the sensors are transmitted to the data processing center in real time through wired and wireless methods. A method combining wireless transmission solutions (such as Wi-Fi, Zigbee) and wired networks (such as Ethernet) is used to ensure the stability and security of data transmission and provide support for real-time monitoring; the collected sensor data is pre-processed to improve data quality. This step includes data cleaning, removing erroneous data, filling missing values, and processing outliers to ensure that the data used for analysis and modeling is accurate and reliable; real-time monitoring of environmental parameters in the building, and statistical analysis of historical data. Using regression analysis and other methods, a model between environmental parameters and energy consumption is established, expressed as: ; Among them, y represents energy consumption, X represents environmental parameters, are model parameters, The best estimate of the model parameters is determined by minimizing the sum of squared errors.

[0036] Reference Figure 1 ,In step S4, by comparing the real-time data with the predicted data, the output power and control strategy of the ,central air conditioner are dynamically adjusted. At the same time, the database service is used to ,process and store the real-time data, the data from different sources are unified in ,format and standardized, and the stream processing framework is used to ,analyze the real-time data.

[0037] Specifically, by comparing real-time data with predicted data, the output power and control strategy of the central air conditioner are dynamically adjusted. At the same time, database services are used to process and store real-time data, and data from different sources are standardized and processed in a unified format. The stream processing framework is used to analyze real-time data, and efficient database services are used to store real-time operation data. By establishing a unified data storage format, data collected from different sources such as sensors, environmental monitoring equipment, and user feedback are standardized to ensure consistency and comparability between different data sources. This step not only improves data management efficiency, but also simplifies the subsequent data analysis process. In the collected data, data standardization is performed to solve the problem that the data comes from different sources and has different formats. This process includes unified measurement units, format conversion, and missing value processing, making the data easier for subsequent analysis and model application, and using the stream processing framework to analyze real-time data. These stream processing tools can efficiently process high-speed data streams, calculate, analyze, and respond in real time. Through the analysis of real-time data streams, the system can: Realize dynamic data monitoring and quickly identify abnormal situations; Real-time aggregation and statistics of data provide a basis for adjusting the output power of central air conditioners.

[0038] Reference Figure 1 In step S5, the data is input into the actual building physical model, so that the actual building model simulates the changes in the thermal environment inside the building according to the input data, and the finite element method is used to calculate and predict the temperature in the actual building. The finite element method is expressed by the heat conduction equation, and its steady-state heat conduction equation formula is as follows: ; The formula is combined with the transient heat conduction equation to describe the heat conduction equation, and the formula of the transient heat conduction equation is as follows: ; in, represents temperature; k represents the thermal conductivity of the material; Q represents the heat source term per unit volume; represents the gradient operator; is a divergence operator, which is expressed by the operation of the above equations and the discretized finite element method in matrix form as follows: ; in, represents the hot melt matrix; represents the thermal conductivity matrix; represents the temperature node vector; represents the heat source term vector, which is solved numerically by the formula to obtain the temperature distribution inside the building.

[0039] Specifically, in order to solve the above heat conduction equation, it is necessary to discretize it and express it using the finite element method in matrix form; the discretized equations are solved by numerical methods to obtain the temperature distribution inside the building. Time stepping can be performed using explicit or implicit methods to gradually deduce the temperature changes inside the building; The actual monitoring data and model parameters such as material thermal conductivity, specific heat capacity, density and heat source distribution are entered into the actual building physical model. Based on the collected data, the geometric model of the building is constructed and the finite element mesh is divided to ensure the accuracy of the model. The boundary conditions of the model are set, including outdoor temperature, indoor set temperature, heat source distribution, etc., in order to accurately simulate the thermal environment in the building. The simulation model is run through numerical solution tools to calculate the temperature distribution in different areas inside the building and generate the corresponding thermal environment change curve. The simulation results are analyzed to evaluate the thermal environment inside the building, and improvement suggestions are provided based on the temperature distribution to optimize the operation strategy of the air-conditioning system.

[0040] Reference Figure 5 , in step S6, the logical steps of the optimization algorithm are as follows: C1. According to the final target temperature Design objective function J: ; Controlling the input data as chromosome genes; C2. Randomly generate a set of control strategies represented by an encoding vector: ; in, represents the control input at time point n; C3. Evaluate the fitness of each individual in the population and use the objective function G to calculate the fitness value of each individual: ; C4. Select the better individuals for reproduction according to the fitness, and randomly select a pair of parents from the selected individuals to perform a crossover operation to produce offspring. The crossover operation is as follows: ; C5. Randomly mutate the offspring, the mutation method is: ; in, Represents the random perturbation value; the offspring generated by selection, crossover and mutation are placed into the new population C6. Repeat steps C3 to C5 until the termination condition is met. The termination condition is that the maximum number of generations is reached. Finally, the individual with the highest fitness is selected from the population as the optimal control strategy to predict the corresponding control input.

[0041] Specifically, first, the objective function J is designed according to the final target temperature of the building. On this basis, the control input data is used as the chromosome gene to represent the different control strategies of the system, thereby forming a population that can be used for genetic algorithm operation. A set of control strategies is randomly generated and represented by a coding vector, and then the fitness of each individual in the population is evaluated. The fitness value of each individual is calculated using the objective function G; based on the fitness evaluation, the better individuals are selected for reproduction. A pair of parent individuals is randomly selected for crossover operation to generate new offspring; the offspring are randomly mutated to increase the diversity of the population and avoid local optimal solutions; the above steps are repeated until the termination condition is met, which is usually set to reach the maximum number of generations, or the objective function converges. After the optimization is completed, the individual with the highest fitness is selected from the final population as the optimal control strategy, and the corresponding control input is predicted.

[0042] The implementation principle of the embodiment of the present application is as follows: First, a physical model of the actual building and central air-conditioning system is established locally. The model should take into account factors such as the building's geometric shape, material properties, window locations and their thermal insulation performance, and reflect the performance parameters of the central air-conditioning system, such as cooling / heating capacity, air flow path, etc., to ensure that the model is highly consistent with the actual scene; Next, based on the constructed physical model, a multi-scenario simulation environment is established, and corresponding control strategies are designed for different scenarios such as different seasons and different weather conditions. These strategies should cover temperature regulation, humidity control and air circulation to adapt to different user needs and environmental changes; obtain the current environmental parameters of the actual building in real time, including indoor temperature, humidity, wind speed, external temperature, etc. These parameters are input into the constructed physical model to provide basic data for subsequent control decisions; a real-time control feedback mechanism is established to continuously monitor the current indoor temperature and external environmental changes in each time step. Through high-frequency data collection and processing, ensure that the system can respond to environmental changes in a timely manner and maintain indoor comfort; based on the output data of the central air conditioner and the current environmental parameters, this information is entered into the physical model to predict the temperature changes in the actual building. This process requires combining historical data with the current state, and using algorithms to make accurate temperature predictions; finally, all models are iterated using optimization algorithms to determine the optimal control curve based on the prediction results. The optimization algorithm can combine a variety of modern optimization techniques such as genetic algorithms and particle swarm optimization to achieve the best balance between maximum energy efficiency and user comfort.

[0043] The embodiments of this specific implementation are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. The same components are represented by the same figure marks. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A multi-model optimization control method for large central air conditioners based on final target temperature, characterized in that: The following steps are involved: S1. Build physical models of actual buildings and central air conditioners locally; S2. Establish a multi-scenario simulation environment and design corresponding control strategies for each scenario; S3, obtaining current environmental parameters of the actual building, and inputting the parameters into the physical model; S4. Establish a real-time control feedback mechanism to continuously obtain the current indoor temperature and external environment changes in each time step; S5, entering the central air-conditioning output data and the environmental parameters into the actual building physical model to predict the temperature in the actual building; S6. All models are iterated using an optimization algorithm to determine the best control curve for prediction.

2. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 1 is characterized in that: In step S1, a physical model of the actual building is constructed locally, and the construction steps are as follows: A1. Establish a building heat conduction model: ； Wherein, T represents temperature distribution; α represents thermal diffusivity; Q represents heat source term, and the central air conditioning temperature distribution in the actual building is established by the formula; A2. Space and time discretization: Use finite difference methods to discretize space into a grid of points: Define grid spacing , and , represents the spacing on the x, y, and z axes; The temperature distribution is discretized into the temperature values ​​of the grid nodes , and , where i, j, k represent the spatial index of the node; Using an explicit method leads to discretization of time: ； The time step is , the time discretization is ; Temperature discreteness is , represents the temperature value at the nth moment; A3. Combining the heat conduction equation and the discretization result, the discrete equation group is obtained as follows: ； A4. Set boundary conditions according to the actual situation of the building, where the fixed temperature boundary conditions are: ； The fixed heat flow boundary conditions are: ； Initialize the temperature distribution throughout the actual building: ； A5. Step by step solve the temperature distribution of each time step according to the discrete equations ,when When it is less than a set threshold, the iteration and solution are stopped to obtain the physical model of the actual building at different times and spaces.

3. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 1 is characterized in that: In step S1, the central air conditioning physical model is constructed using the following expression: ； in, represents the total cooling load; k represents the thermal conductivity of the material; A represents the heat transfer area; Indicates the external temperature; represents the internal temperature; c represents the specific heat capacity of water; m represents the mass flow rate; Indicates the return water temperature; represents the water supply temperature; n represents the number of people; q represents the calorific value per person; represents the heat dissipation of electrical equipment; eff represents the system efficiency. The central air-conditioning physical model constructed by the expression can express the operating efficiency and energy consumption of the central air-conditioning.

4. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 1 is characterized in that: In step S2, a machine learning method is used to predict and evaluate the temperature changes in different control scenarios and integrate them into the optimization process. The steps of the machine learning method are as follows: D1. Data collection and processing: Collecting and cleaning the indoor and outdoor environmental parameters of the actual building; D2. Extract key features that affect indoor temperature from the collected data; D3. Use the computational model to make scene judgments on key features and output the predicted temperature; D4. Input the collected new environmental data into the trained model in real time to predict the indoor temperature at the next moment.

5. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 1 is characterized in that: In step S3, the environmental parameters are obtained as follows: B1. Sensor deployment: installing sensors in key areas of the actual building; B2. Data collection: Transmit sensor data to the data processing center through wired and wireless methods; B3. Data preprocessing: cleaning the collected data; B4. Data analysis and feedback: Real-time monitoring of environmental parameters in the building, analysis of historical data through statistical analysis, and establishment of a model between environmental parameters and energy consumption. The model is expressed as follows: ； Among them, y represents energy consumption; Indicates environmental parameters; represents the model parameters; Denotes the error term, and the best estimate of the model parameters is determined by minimizing the sum of squared errors: ； Where m is the number of samples; represents the energy consumption of the i-th sample; represents the value of the ith environmental parameter, and through the above calculation, the environmental parameters are collected and optimized to establish an energy consumption prediction model.

6. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 1 is characterized in that: In step S4, the output power and control strategy of the central air conditioner are dynamically adjusted by comparing the real-time data with the predicted data. At the same time, the database service is used to process and store the real-time data, the data from different sources are unified and standardized, and the real-time data is analyzed using the stream processing framework.

7. The large central air conditioning multi-model optimization control method based on final target temperature according to claim 1 is characterized in that: In step S5, the data is input into the actual building physical model, so that the actual building model simulates the thermal environment changes inside the building according to the input data, and uses the finite element method to calculate and predict the temperature in the actual building. The finite element method is expressed by the heat conduction equation, and its steady-state heat conduction equation formula is as follows: ； The formula is combined with the transient heat conduction equation to describe the heat conduction equation, and the formula of the transient heat conduction equation is as follows: ； in, represents temperature; k represents the thermal conductivity of the material; Q represents the heat source term per unit volume; represents the gradient operator; is a divergence operator, which is expressed by the operation of the above equations and the discretized finite element method in matrix form as follows: ； in, represents the hot melt matrix; represents the thermal conductivity matrix; represents the temperature node vector; represents the heat source term vector, which is solved numerically by the formula to obtain the temperature distribution inside the building.

8. The large central air conditioning multi-model optimization control method based on the final target temperature according to claim 7 is characterized in that: In step S6, the logic steps of the optimization algorithm are as follows: C1. According to the final target temperature Design objective function J: ； Controlling the input data as chromosome genes; C2. Randomly generate a set of control strategies represented by an encoding vector: ； in, represents the control input at time point n; C3. Evaluate the fitness of each individual in the population and use the objective function G to calculate the fitness value of each individual: ； C4. Select the better individuals for reproduction according to the fitness, and randomly select a pair of parents from the selected individuals to perform a crossover operation to produce offspring. The crossover operation is as follows: ； C5. Randomly mutate the offspring, the mutation method is: ； in, Represents the random perturbation value; the offspring generated by selection, crossover and mutation are placed into the new population C6. Repeat steps C3 to C5 until the termination condition is met. The termination condition is that the maximum number of generations is reached. Finally, the individual with the highest fitness is selected from the population as the optimal control strategy to predict the corresponding control input.

Images