Glass closed space microclimate regulation and control method based on CFD-IL coupling model

Abstract

The invention relates to the technical field of building environment control and intelligent regulation and control, and discloses a glass closed space microclimate regulation and control method based on a CFD-IL coupling model. The method comprises the following steps: firstly, establishing a CFD-IL coupling model comprising a three-dimensional flow field simulation module and a dynamic parameter optimization module, collecting multi-dimensional environmental parameters such as temperature, humidity, air velocity and illumination intensity in a glass closed space in real time, inputting the model to predict future microclimate parameter distribution, and generating a multi-target regulation and control instruction set according to the future microclimate parameter distribution; and parameters of environment control equipment such as an air-conditioning system and a ventilation device are dynamically adjusted. Meanwhile, based on an environmental parameter feedback optimization model weight coefficient and a boundary condition, a flow field simulation boundary condition can be corrected according to external meteorological data. The method can accurately regulate and control microclimate, saves energy, reduces consumption, improves the self-adaptive capacity of the system, and has wide application prospects in closed spaces such as glass greenhouses, exhibition halls and office buildings.

Description

Technical Field

[0001] The present invention relates to the technical field of building environment control and intelligent regulation, and specifically to a microclimate regulation method for glass-enclosed spaces based on a CFD-IL coupling model. Background Art

[0002] In the field of modern architecture, glass-enclosed spaces are widely used due to their unique lighting and aesthetic advantages. For example, various greenhouses are used for plant cultivation, exhibition halls are used for displaying exhibits, and office buildings create a comfortable working environment. However, the special structure of glass-enclosed spaces makes it difficult to regulate the internal microclimate.

[0003] On the one hand, although the high light transmittance of glass is beneficial to lighting, the intensity of solar radiation changes significantly in different seasons and periods, resulting in drastic fluctuations in indoor temperature. In summer, strong solar radiation causes the indoor temperature to rise rapidly. If effective heat dissipation cannot be achieved, the excessive temperature not only affects human comfort but is also extremely unfavorable for plant growth in greenhouses and the preservation of exhibits in exhibition halls, possibly causing plant wilting and exhibit damage. In winter, indoor heat is easily dissipated through the glass, resulting in too low indoor temperature and increased heating energy consumption.

[0004] On the other hand, the air circulation and humidity distribution in glass-enclosed spaces are also difficult to be uniform. Due to the relatively enclosed space, the air flow is not smooth, and there are easily stuffy or humid conditions in local areas. In greenhouses, poor air circulation affects the photosynthesis and respiration of plants and hinders plant growth. In office places, the poor air environment reduces the work efficiency of personnel and even causes health problems. Traditional ventilation and humidification equipment often cannot accurately control the air flow rate and humidity in different areas and are difficult to meet the diverse needs of complex spaces.

[0005] Most of the existing microclimate regulation methods have limitations. Some simple temperature control systems only control the start and stop of air conditioners based on the feedback of temperature sensors, without comprehensively considering other environmental factors such as humidity and light, and the regulation effect is single and rough. Although some complex regulation systems consider multi-parameter control, the models and algorithms used lack adaptability and cannot optimize the regulation strategy in real time according to the environmental changes in the space. For example, in greenhouse environments, the environmental parameter requirements vary greatly at different growth stages of plants, and traditional regulation methods are difficult to flexibly respond to these changes, resulting in resource waste and poor regulation effects. Moreover, when dealing with complex space structures and dynamic environmental changes, it is difficult to balance the calculation accuracy and efficiency of the models in the existing technology, and it cannot meet the requirements of precise regulation and rapid response in practical applications.

[0006] Therefore, it is of great practical significance to develop a small climate control method for glass-enclosed spaces that can comprehensively consider various environmental factors, adaptively optimize control strategies, and improve control accuracy and efficiency. This can not only enhance the user experience and application value of glass-enclosed spaces, but also reduce energy consumption and achieve sustainable development. Summary of the Invention

[0007] The purpose of the present invention is to provide a small climate control method for glass-enclosed spaces based on the CFD-IL coupling model to solve the problems raised in the above background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: A small climate control method for glass-enclosed spaces based on the CFD-IL coupling model, where the full name of CFD is Computational Fluid Dynamics, representing Computational Fluid Dynamics, and the full name of IL is Iterative Learning, representing Iterative Learning. The coupling model emphasizes the deep integration of the flow field simulation ability of CFD and the dynamic optimization ability of IL. The method specifically includes: Establish a CFD-IL coupling model for the glass-enclosed space, which includes a three-dimensional flow field simulation module based on computational fluid dynamics and a dynamic parameter optimization module based on the iterative learning algorithm; Real-time collect multi-dimensional environmental parameters in the glass-enclosed space, including temperature, humidity, air velocity, and light intensity; Input the multi-dimensional environmental parameters into the CFD-IL coupling model, and predict the distribution of small climate parameters within a preset future time period through flow field simulation and iterative learning; Generate a multi-objective control instruction set according to the prediction results, and the instruction set includes the operating mode, power, and action range of the environmental control equipment; Dynamically adjust the parameters of the environmental control equipment, including the coordinated control strategies of the air conditioning system, ventilation device, humidifier, and sunshade mechanism; Based on the updated environmental parameter feedback, optimize the weight coefficients and boundary conditions of the CFD-IL coupling model through the iterative learning algorithm.

[0009] Preferably, the steps of establishing the CFD-IL coupling model include: Divide the grid topology structure of the glass-enclosed space and set the initial fluid mechanics boundary conditions; Construct the objective function of the iterative learning algorithm, and the objective function takes the uniformity of environmental parameters, energy consumption minimization, and equipment response speed as optimization variables; Train the CFD-IL coupling model with historical environmental data, and dynamically correlate the flow field simulation results with the parameter correction amount of iterative learning.

[0010] Preferably, the step of collecting the multi-dimensional environmental parameters includes: Arrange a distributed sensor network in a glass enclosed space to obtain the gradient distribution data of temperature and humidity in real time; Collect the spatial vectors of air velocity through a laser velocimeter; Use a light sensor to record the light intensity and incident angle in different areas.

[0011] Preferably, the step of generating the multi-objective regulation instruction set includes: Compare the predicted microclimate parameter distribution with the preset comfort interval to generate a deviation matrix; Based on the deviation matrix, solve the optimal parameter combination of the environmental control equipment through a multi-objective particle swarm optimization algorithm; Convert the optimal parameter combination into an instruction sequence executable by the equipment and assign priority weights.

[0012] Preferably, the step of dynamically adjusting the environmental control equipment includes: According to the instruction sequence, adjust the supply air temperature, air velocity and direction of the air conditioning system; Control the start-stop frequency and opening angle of the ventilation device; Synchronously adjust the atomization particle concentration of the humidifier and the unfolding area of the sunshade mechanism.

[0013] Preferably, the step of optimizing the CFD-IL coupling model includes: Calculate the residuals between the actual environmental parameters and the predicted values and construct a residual matrix; Update the gain coefficient of the iterative learning algorithm through an adaptive Kalman filter; Correct the turbulence equation coefficients of the CFD-IL coupling model and the boundary conditions of the locally refined mesh area.

[0014] Preferably, the method further includes: Within a preset time interval, re-divide the mesh density of the CFD-IL coupling model and dynamically adjust the simulation step size according to the environmental parameter change rate.

[0015] Preferably, the iterative learning algorithm includes: Construct a recursive neural network based on time series, input the historical regulation instructions and the corresponding environmental parameter feedback; Optimize the weights of the hidden layer nodes of the neural network through the backpropagation algorithm; Output the parameter correction amount for the next time step and fuse it into the initial conditions of the CFD-IL coupling model.

[0016] Preferably, the method further includes: When it is detected that the environmental parameters of a local area exceed the preset threshold, the emergency control mode is activated and the corresponding environmental control equipment of the area is adjusted first; In emergency control mode, the global optimization algorithm is suspended and a rule-based control strategy is used to respond quickly.

[0017] Preferably, the method further comprises: Connecting external meteorological data to the CFD-IL coupling model, including outdoor temperature, wind speed and solar radiation intensity; The external boundary conditions of the flow field simulation are dynamically corrected through the coupling analysis of external data and internal parameters.

[0018] Compared with the prior art, the present invention has the following beneficial effects: In terms of precise control of microclimate, by real-time collection of multi-dimensional environmental parameters such as temperature, humidity, air flow rate and light intensity, and using CFD-IL coupling model for flow field simulation and iterative learning prediction, the distribution of microclimate parameters in the glass-enclosed space within the preset time period in the future can be accurately grasped. Taking the glass greenhouse as an example, under different seasons and weather conditions, this method can accurately predict the temperature change trend of each area in the greenhouse, and adjust environmental control equipment such as air conditioning system, ventilation device, humidifier and sunshade mechanism in advance. When it is predicted that the temperature in a certain area is about to rise, the sunshade mechanism is activated in advance to block the sunlight, and the air supply temperature and wind speed of the air conditioning system are adjusted at the same time to accurately maintain the suitable temperature range required for plant growth, avoiding the adverse effects of excessively high or low temperatures on plant growth. For exhibition halls, the environmental parameters of the exhibition area can be accurately controlled to ensure that the exhibits are in the best preservation environment and reduce the risk of damage to the exhibits by environmental factors.

[0019] In terms of energy saving and consumption reduction, the objective function of the iterative learning algorithm constructed by the present invention takes the uniformity of environmental parameters, minimization of energy consumption and equipment response speed as optimization variables. Through the historical environmental data training model, the flow field simulation results are dynamically associated with the parameter correction amount of iterative learning to realize the intelligent collaborative control of environmental control equipment. For example, in an office space, when there are fewer people, the power of the air-conditioning system is automatically reduced according to the indoor environmental parameters and the prediction results. At the same time, the start and stop frequency and opening and closing angle of the ventilation device are reasonably adjusted to minimize energy consumption while ensuring indoor comfort. Compared with traditional control methods, it can effectively reduce energy consumption, save operating costs, and conform to the concept of green buildings and sustainable development.

[0020] In terms of the system's adaptability, based on the feedback of updated environmental parameters, the weight coefficients and boundary conditions of the CFD-IL coupling model are optimized through an iterative learning algorithm. As time goes by and the environment changes, the model can continuously self-adjust and optimize. When different seasons alternate, significant changes occur in outdoor meteorological conditions. The model can dynamically correct the external boundary conditions of the flow field simulation based on the coupled analysis of external meteorological data (such as outdoor temperature, wind speed, and solar radiation intensity) and internal parameters. At the same time, an adaptive Kalman filter is used to update the gain coefficient of the iterative learning algorithm, and correct the turbulence equation coefficients of the CFD model and the boundary conditions of the locally refined mesh area, so that the model always maintains high prediction accuracy and regulation performance, adapting to the complex and changeable glass-enclosed space environment.

[0021] In response to emergencies, when it is detected that the environmental parameters in a local area exceed the preset threshold, the emergency regulation mode is activated. In a greenhouse, if the temperature in a certain area suddenly becomes too high, the system will first adjust the corresponding environmental control equipment in that area, such as increasing the cooling power of the air conditioner in that area to quickly reduce the temperature and ensure the normal growth of plants. In the emergency regulation mode, the global optimization algorithm is suspended, and a rule-based control strategy is adopted for rapid response. This mechanism ensures that effective measures can be taken quickly in case of emergencies to avoid serious damage to people, items, or plants in the space caused by environmental deterioration.

[0022] In addition, re-dividing the grid density of the CFD model at preset time intervals and dynamically adjusting the simulation step size according to the change rate of environmental parameters can improve the calculation efficiency while ensuring the calculation accuracy. The grid is encrypted in areas where environmental parameters change violently to improve the simulation accuracy of the model for local complex flow fields; the grid size is appropriately increased in areas with gentle changes to reduce the amount of calculation. This dynamic adjustment strategy enables the regulation method of the present invention to operate efficiently in glass-enclosed spaces of different scales and complexities, and has wide applicability. Description of the Drawings

[0023] Figure 1 It is the working principle diagram of the small climate regulation method for the glass-enclosed space based on the CFD-IL coupling model described in the present invention; Figure 2 It is the step diagram of multi-dimensional environmental parameter acquisition; Figure 3 It is the step diagram of generating a multi-objective regulation instruction set; Figure 4 It is the step diagram of optimizing the CFD-IL coupling model. Detailed Embodiments

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0025] Please refer to Figures 1-4 , the present invention provides a technical solution: a method for regulating the microclimate in a glass enclosed space based on a CFD-IL coupling model, and the method includes: Establish a CFD-IL coupling model: The constructed model includes a three-dimensional flow field simulation module based on computational fluid dynamics and a dynamic parameter optimization module based on an iterative learning algorithm. By analyzing the glass enclosed space, its grid topology structure is divided, and the initial hydrodynamic boundary conditions are set, providing a basic framework for subsequent flow field simulation. At the same time, an objective function of the iterative learning algorithm is constructed, with the uniformity of environmental parameters, energy consumption minimization, and equipment response speed as optimization variables, enabling the model to continuously optimize according to actual needs.

[0026] Collect multi-dimensional environmental parameters in real time: Corresponding sensors are arranged in the glass enclosed space to collect multi-dimensional environmental parameters such as temperature, humidity, air velocity, and light intensity in real time. These parameters are the basic data for the model to make predictions and regulations, and their accuracy and real-time nature are crucial.

[0027] Predict the distribution of microclimate parameters: The collected multi-dimensional environmental parameters are input into the CFD-IL coupling model, and the model uses methods of flow field simulation and iterative learning to predict the distribution of microclimate parameters in the glass enclosed space within a preset future time period. This step provides a basis for generating subsequent regulation instructions.

[0028] Generate a multi-objective regulation instruction set: According to the predicted results of the microclimate parameter distribution, a multi-objective regulation instruction set including the operating mode, power, and action range of environmental control devices is generated. This instruction set is the key to realizing microclimate regulation. By reasonably setting device parameters, the microclimate in the glass enclosed space can reach an ideal state.

[0029] Dynamically adjust the parameters of environmental control devices: According to the generated regulation instruction set, the parameters of environmental control devices such as air conditioning systems, ventilation devices, humidifiers, and shading mechanisms are dynamically adjusted. By coordinately controlling these devices, precise regulation of the microclimate in the glass enclosed space is achieved.

[0030] Optimizing the CFD-IL Coupling Model: Based on the feedback of updated environmental parameters, the weight coefficients and boundary conditions of the CFD-IL coupling model are optimized using an iterative learning algorithm. As the environment changes, the model parameters are continuously adjusted to improve the prediction accuracy and control effect of the model.

[0031] The present invention will be further described below in conjunction with Examples 1 to 6: Example 1: When establishing the CFD-IL coupling model, the glass enclosed space needs to be meshed first. Taking a large glass greenhouse as an example, the greenhouse is 50 meters long, 30 meters wide, and 8 meters high. Using professional meshing software, according to the internal structure of the greenhouse and the expected flow field changes, a non-uniform meshing method is adopted. In areas with large flow field changes, such as near heat sources (such as heating equipment), ventilation openings, and plant cultivation areas, the mesh is refined so that the mesh size reaches 0.2 m × 0.2 m × 0.2 m; while in the middle area of the space with relatively small flow field changes, the mesh size is set to 0.5 m × 0.5 m × 0.5 m, which not only ensures the calculation accuracy but also controls the calculation amount. After the meshing is completed, the initial fluid mechanics boundary conditions are set. For example, at the ventilation opening, according to the design parameters of the ventilation equipment, the wind speed is set to 2 m / s, the temperature is 25 °C, and the relative humidity is 60%; at the wall boundary, a no-slip boundary condition is set, that is, the velocity at the contact between the fluid and the wall is 0.

[0032] Construct the objective function of the iterative learning algorithm, with the uniformity of environmental parameters, energy consumption minimization, and equipment response speed as the optimization variables. Assume that the uniformity of environmental parameters is measured by the standard deviation of temperature and the standard deviation of humidity , the energy consumption is represented by the total energy consumption , and the equipment response speed is represented by the response time . The objective function can be expressed as: , where , , , are weight coefficients, which are set according to actual needs and importance. For example, during the plant growth stage, more attention is paid to the uniformity of temperature and humidity, and and can be set to larger values, such as , ; during the period of energy shortage, the value of can be increased, such as ; and when it is necessary to quickly respond to sudden environmental changes, the value of can be correspondingly increased, such as .

[0033] In terms of collecting multi-dimensional environmental parameters, a distributed sensor network is arranged inside the greenhouse. Along the length direction of the greenhouse, a temperature and humidity sensor node is set every 5 meters. Each node is equipped with a temperature sensor and a humidity sensor, with a total of 10 nodes. These sensors can obtain real-time gradient distribution data of temperature and humidity. For example, at noon in summer, the sensor near the top detects a temperature of 32 °C and a relative humidity of 50%, while the sensor near the ground planting area detects a temperature of 28 °C and a relative humidity of 65%.

[0034] A laser velocimeter is used to collect the spatial vector of air velocity. The laser velocimeter is installed at the center of the greenhouse. By emitting laser beams, the air velocity in different directions is measured. For example, when the ventilation equipment is turned on, it is measured that the air velocity near the ventilation opening in the horizontal direction is 1.5 m / s, and the air velocity at a height of 2 meters from the ground in the vertical direction is 0.8 m / s.

[0035] Light sensors are used to record the light intensity and incident angle in different areas. Light sensors are installed in different planting areas of the greenhouse. For example, 3 light sensors are installed in the planting area of light-loving plants and the planting area of shade-tolerant plants respectively. At 10 am on a sunny day, the light sensors in the planting area of light-loving plants detect a light intensity of 8000 lux and an incident angle of 45°; the light intensity in the planting area of shade-tolerant plants is 3000 lux and the incident angle is 30°.

[0036] Example 2: When generating a multi-objective regulation instruction set, it is based on the comparison between the predicted microclimate parameter distribution and the preset comfort interval. For example, for temperature, the preset comfort interval is 22 °C - 28 °C; for humidity, the preset comfort interval is 50% - 70%. Suppose the predicted temperature in a certain area of the greenhouse in the next 1 hour is 30 °C and the humidity is 45%. After comparing with the preset comfort interval, a deviation matrix is generated. The temperature deviation is , and the humidity deviation is .

[0037] Based on this deviation matrix, the optimal parameter combination of the environmental control equipment is solved through a multi-objective particle swarm optimization algorithm. In the multi-objective particle swarm optimization algorithm, each particle represents a set of parameter combinations of the environmental control equipment, such as the supply air temperature 、wind speed of the air conditioning system, the start-stop frequency 、opening and closing angle of the ventilation device, the atomization particle concentration of the humidifier, and the unfolded area of the sunshade mechanism. The particles continuously update their positions and velocities in the search space to find the optimal solution.

[0038] Suppose that after multiple iterative calculations, a set of optimal parameter combinations are obtained: the supply air temperature of the air conditioning system , the wind speed m / s; the start-stop frequency of the ventilation device times / hour, the opening and closing angle ; the atomization particle concentration of the humidifier g / m³; the deployed area of the sunshade mechanism m².

[0039] Convert this set of optimal parameter combinations into an instruction sequence executable by the device and assign priority weights. For example, since the temperature deviation has a greater impact on plant growth, the priority of the instruction to adjust the air conditioning system is set to high; the humidity deviation is relatively small, and the priority of the instruction to adjust the humidifier is set to medium; the priority of the adjustment instructions for the ventilation device and the sunshade mechanism is set to low. According to the priority order, first send instructions to the air conditioning system to adjust the supply air temperature and wind speed so that it operates according to the set parameters; then send instructions to the ventilation device to control the start-stop frequency and opening and closing angle; then send instructions to the humidifier to adjust the atomization particle concentration; finally, send instructions to the sunshade mechanism to adjust the deployed area.

[0040] When dynamically adjusting the environmental control equipment, operate according to the above instruction sequence. For the air conditioning system, adjust the supply air temperature to 24°C and the wind speed to 1.2 m / s by controlling the internal temperature adjustment device and the fan. In terms of the ventilation device, control the opening and closing of the ventilation equipment at a start-stop frequency of 10 times / hour through the controller and adjust the opening and closing angle of the ventilation opening to 60° to achieve ventilation and air exchange. The humidifier increases the air humidity by adjusting the spraying device to increase the atomization particle concentration to 50 g / m³. The sunshade mechanism expands the deployed area to 80 m² through motor drive to block sunlight and reduce the indoor temperature.

[0041] Example 3: When optimizing the CFD-IL coupling model, calculating the residuals between the actual environmental parameters and the predicted values is a key step. For example, at a certain moment, the actual measured temperature at a point in the glass-enclosed space is 26°C, while the model predicts the temperature at this point to be 25.5°C, then the temperature residual is ; the actual measured humidity is 62%, the predicted humidity is 60%, and the humidity residual is . And so on, construct a residual matrix for the residuals of each environmental parameter.

[0042] Update the gain coefficient of the iterative learning algorithm through an adaptive Kalman filter. The adaptive Kalman filter dynamically adjusts the gain coefficient according to information such as the residual matrix and the system noise covariance matrix. Suppose the system noise covariance matrix is , and the measurement noise covariance matrix is , the prediction error covariance matrix is , then the gain coefficient is calculated as follows: , where is the observation matrix, which is used to describe the relationship between the actual measurement value and the predicted value. In this embodiment, is determined according to the corresponding relationship between the actually measured environmental parameters and the model prediction parameters. By continuously updating the gain coefficient , the iterative learning algorithm can more accurately adjust the model parameters.

[0043] Modify the turbulence equation coefficients and the boundary conditions of the locally refined mesh region of the CFD-IL coupling model. For example, in the commonly used turbulence model of the CFD-IL coupling model, according to the residual analysis and the actual flow field situation, the coefficients , , in the turbulence equation are adjusted. Suppose the original , after optimization, it is adjusted to . For the boundary conditions of the locally refined mesh region, such as in the mesh refinement region near the heat source, the originally set heat flux boundary condition is watts per square meter, and according to the actual measurement and the model optimization requirements, it is adjusted to watts per square meter to more accurately simulate the changes in the flow field and temperature field.

[0044] Embodiment 4: Within a preset time interval, re-divide the mesh density of the CFD-IL coupling model and dynamically adjust the simulation step size according to the environmental parameter change rate, which is of great significance for improving the accuracy and calculation efficiency of the model. For example, set the preset time interval to every 6 hours. During the day, as the solar radiation intensity changes, the environmental parameters such as the temperature and air velocity in the glass-enclosed space change relatively fast. Taking an exhibition greenhouse as an example, within the period from 10 am to 4 pm, the mesh is re-divided every 6 hours. According to the monitoring data in the morning, it is found that the temperature and air velocity change violently in the areas near the glass curtain wall and the plant-dense areas. Therefore, when re-dividing the mesh, the meshes in these areas are further refined, and the mesh size is reduced from the original 0.3 m × 0.3 m × 0.3 m to 0.1 m × 0.1 m × 0.1 m, while in other areas with relatively small changes, the mesh size remains unchanged.

[0045] Dynamically adjust the simulation step size according to the environmental parameter change rate. The environmental parameter change rate is obtained by calculating the ratio of the difference between the environmental parameters at adjacent times to the time interval. For example, within a certain 1-hour period, the temperature rises from 25 °C to 27 °C, then the temperature change rate is When the temperature change rate is greater than 1 °C / hour, shorten the simulation step size from the original 0.05 seconds to 0.02 seconds; when the temperature change rate is less than 0.5 °C / hour, extend the simulation step size to 0.1 seconds. In this way, when the environmental parameters change drastically, a smaller simulation step size is adopted to improve the accuracy of the model; when the environmental parameters change slightly, a larger simulation step size is adopted to reduce the computational amount and improve the computational efficiency.

[0046] Example 5: In this embodiment, the iterative learning algorithm is implemented by constructing a recurrent neural network based on time series. Taking a flower cultivation greenhouse as an example, collect the control instructions every hour in the past week, including the temperature setting value of the air conditioning system, the start / stop state of the ventilation device, the working intensity of the humidifier, and the opening / closing degree of the sunshade mechanism, etc. At the same time, collect the corresponding environmental parameter feedback, such as temperature, humidity, light intensity, etc. Arrange these historical data in chronological order as the input of the recurrent neural network.

[0047] The recurrent neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the historical control instructions and environmental parameter feedback data. The hidden layer processes the input data through a recurrent structure to learn the time series features in the data. During the training process, the weights of the hidden layer nodes of the neural network are optimized through the backpropagation algorithm. Assume that the hidden layer has nodes, and the weight between node and node is . Calculate the gradient of the error with respect to the weight through the backpropagation algorithm, and continuously adjust the weight value to minimize the error between the predicted output and the actual output of the network.

[0048] After multiple trainings, the recurrent neural network can output the parameter correction amount for the next time step. For example, the network predicts that the temperature setting value of the air conditioning system needs to be reduced by 1 °C in the next hour, the start / stop frequency of the ventilation device needs to be increased by 2 times / hour, and the working intensity of the humidifier needs to be increased by 10%. Incorporate these parameter correction amounts into the initial conditions of the CFD-IL coupling model. For example, use the adjusted temperature setting value as the initial value of the temperature boundary condition in the CFD-IL coupling model, and the adjustment information of the start / stop frequency of the ventilation device is used to update the flow boundary condition of the ventilation opening in the CFD-IL coupling model. The adjustment of the working intensity of the humidifier is used to correct the parameters of the humidity source term in the CFD-IL coupling model, so as to realize the dynamic optimization of the CFD-IL coupling model and improve the accuracy of model prediction.

[0049] Example 6: During the operation of the glass enclosed space, when the environmental parameters in a local area are detected to exceed the preset threshold, the emergency regulation mode is activated. For example, in a glass exhibition hall, the temperature threshold is set at 30 °C and the humidity threshold is set at 80% in the exhibition area. At a certain moment, the temperature sensor at a corner of the exhibition area detects that the temperature reaches 32 °C, exceeding the preset threshold. At this time, the system immediately activates the emergency regulation mode.

[0050] In the emergency regulation mode, the environmental control equipment corresponding to this area is preferentially adjusted. Due to the high temperature, the air conditioning system near this area is first activated, and its cooling power is adjusted to the maximum, while increasing the air supply volume to quickly reduce the temperature of this area. At the same time, the global optimization algorithm is suspended, and a rule-based control strategy is adopted for rapid response. The rule-based control strategy is set as follows: when the temperature is higher than 30 °C, for every 1 °C increase, the cooling power of the air conditioning system increases by 10%, and the air supply volume increases by 15%; when the humidity is higher than 80%, the humidifier stops working, and the opening frequency of the ventilation device increases by 5 times per hour.

[0051] Connecting external meteorological data to the CFD-IL coupling model is crucial for improving the regulation effect of the model. Taking a glass solar greenhouse as an example, external meteorological data such as outdoor temperature, wind speed, and solar radiation intensity are connected. The outdoor temperature is monitored in real time through sensors , wind speed and solar radiation intensity . In winter, when the outdoor temperature is relatively low, such as reaching -5 °C, through the coupled analysis of external data and internal parameters, the external boundary conditions of the flow field simulation are dynamically corrected. Due to the influence of outdoor cold air, in the CFD-IL coupling model, the heat transfer coefficient of the glass greenhouse wall is appropriately increased. Assuming the original heat transfer coefficient is W / (m·K), it is adjusted to W / (m·K) to more accurately simulate the indoor-outdoor heat exchange process. When the solar radiation intensity is strong, such as reaching 1000 W / m² at noon in summer, according to the solar radiation intensity and the indoor light distribution, the unfolded area of the sunshade mechanism is adjusted, and at the same time, the parameters of the light source term in the CFD-IL coupling model are corrected to effectively regulate the indoor light intensity and temperature.

[0052] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0053] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling microclimate in glass-enclosed space based on CFD-IL coupling model, characterized in that: The following steps are involved: Establishing a CFD-IL coupling model of a glass-enclosed space, the model includes a three-dimensional flow field simulation module based on computational fluid dynamics and a dynamic parameter optimization module based on an iterative learning algorithm; Real-time collection of multi-dimensional environmental parameters in glass-enclosed spaces, including temperature, humidity, air velocity, and light intensity; The multi-dimensional environmental parameters are input into the CFD-IL coupling model, and the distribution of microclimate parameters within a future preset time period is predicted through flow field simulation and iterative learning; Generate a multi-objective control instruction set according to the prediction results, wherein the instruction set includes the operation mode, power and range of action of the environmental control device; Dynamically adjust the parameters of the environmental control equipment, including coordinated control strategies of air conditioning systems, ventilation devices, humidifiers and shading mechanisms; Based on the updated environmental parameter feedback, the weight coefficients and boundary conditions of the CFD-IL coupling model are optimized through the iterative learning algorithm.

2. The method for controlling microclimate in a glass-enclosed space according to claim 1, characterized in that: The steps to establish a CFD-IL coupling model include: Divide the grid topology of the glass-enclosed space and set initial fluid dynamics boundary conditions; Constructing an objective function of an iterative learning algorithm, wherein the objective function takes uniformity of environmental parameters, minimization of energy consumption, and device response speed as optimization variables; The CFD-IL coupling model is trained by historical environmental data to dynamically associate flow field simulation results with iteratively learned parameter corrections.

3. The method for controlling microclimate in a glass-enclosed space as claimed in claim 2, characterized in that: The step of collecting the multi-dimensional environmental parameters includes: Arrange a distributed sensor network in the glass-enclosed space to obtain temperature and humidity gradient distribution data in real time; The spatial vector of air velocity is collected by a laser velocimeter; Use light sensors to record light intensity and incident angle in different areas.

4. The method for controlling microclimate in a glass-enclosed space according to claim 1, characterized in that: The step of generating a multi-objective control instruction set comprises: Compare the predicted microclimate parameter distribution with the preset comfort interval to generate a deviation matrix; Based on the deviation matrix, solving the optimal parameter combination of the environmental control device by a multi-objective particle swarm optimization algorithm; The optimal parameter combination is converted into an instruction sequence executable by the device and a priority weight is assigned.

5. The method for controlling microclimate in a glass-enclosed space as claimed in claim 4, characterized in that: The step of dynamically adjusting the environmental control device comprises: According to the instruction sequence, adjust the air supply temperature, wind speed and direction of the air conditioning system; Control the start / stop frequency and opening / closing angle of the ventilation device; The concentration of atomized particles of the humidifier and the deployment area of ​​the sunshade mechanism are adjusted synchronously.

6. The method for controlling microclimate in a glass-enclosed space according to claim 1, characterized in that: The steps to optimize the CFD-IL coupled model include: Calculate the residuals between the actual environmental parameters and the predicted values, and construct the residual matrix; Updating the gain coefficient of the iterative learning algorithm through an adaptive Kalman filter; The turbulence equation coefficients of the CFD-IL coupling model and the boundary conditions of the local mesh encryption area are modified.

7. The method for controlling microclimate in a glass-enclosed space as claimed in claim 1, characterized in that: Also includes: Within a preset time interval, the grid density of the CFD-IL coupling model is re-divided, and the simulation step size is dynamically adjusted according to the rate of change of the environmental parameters.

8. The method for controlling microclimate in a glass-enclosed space as claimed in claim 1, characterized in that: The iterative learning algorithm includes: Construct a recursive neural network based on time series, input historical control instructions and corresponding environmental parameter feedback; Optimize the hidden layer node weights of the neural network through the back-propagation algorithm; The parameter corrections for the next time step are output and integrated into the initial conditions of the CFD-IL coupling model.

9. The method for controlling microclimate in a glass-enclosed space as claimed in claim 1, characterized in that: Also includes: When it is detected that the environmental parameters of a local area exceed the preset threshold, the emergency control mode is activated and the corresponding environmental control equipment of the area is adjusted first; In emergency control mode, the global optimization algorithm is suspended and a rule-based control strategy is used to respond quickly.

10. The method for controlling microclimate in a glass-enclosed space according to claim 1, characterized in that: The method further comprises: Connecting external meteorological data to the CFD-IL coupling model, including outdoor temperature, wind speed and solar radiation intensity; The external boundary conditions of the flow field simulation are dynamically corrected through the coupling analysis of external data and internal parameters.

Images