Double-air-opening constant-air-volume natural-wind-imitating system based on sliding rail air duct

Abstract

The invention provides a double-air-opening constant-air-volume natural wind simulating system based on a sliding rail air duct, and belongs to the field of natural wind simulating. The double-air-opening constant-air-volume natural wind simulating system comprises a double-air-opening sliding rail structure which comprises a side wall air opening, a top face air opening, a sliding rail assembly and a driving mechanism; the sensor network module comprises a wind speed and direction sensor, a temperature and humidity sensor and a CO2 concentration sensor; the intelligent control module comprises an edge calculation controller and a PLC (Programmable Logic Controller); the feature learning module is used for performing time sequence feature extraction and analysis on the wind speed and direction data and the environmental parameters by adopting an LSTM neural network, and establishing a natural wind feature database; and the cooperative control module is used for establishing a cooperative control algorithm based on double-air-port airflow interaction and generating control parameters. According to the invention, the movable double air ports are adopted, and the natural wind feature learning and cooperative control algorithm of the LSTM neural network is combined, so that intelligent natural wind simulation control is realized, and the naturalness and comfort of spatial airflow distribution are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of natural wind simulation, and in particular to a double-outlet constant-air-volume natural wind simulation system based on a sliding rail air duct. BACKGROUND

[0002] With the continuous acceleration of urbanization, the development and utilization of deep enclosed spaces (such as underground stations, urban tunnels, underground commercial areas, etc.) are becoming increasingly widespread, and higher requirements are placed on the air flow and thermal-humidity environmental comfort of such spaces. Traditional mechanical ventilation systems usually adopt a fixed outlet, constant air speed supply mode, which faces many technical challenges in actual application. The fixed outlet and constant air speed of these systems result in a single air flow field structure, making it difficult to form air speed fluctuations and direction changes in the space, lacking dynamic disturbance mechanisms, and having low air renewal efficiency. Due to the lack of control ability over the spatial distribution of the air source, phenomena such as local heat island effect, humidity accumulation, and ventilation dead angles often occur, which in turn affect the overall environmental comfort.

[0003] Existing natural wind simulation technology mainly achieves air speed fluctuations by adjusting the speed of the fan, or achieves a certain degree of air direction changes using a guide vane with a limited swing angle. Although this technology has been applied in residential air conditioning, exhibition hall environmental regulation, etc., it still has obvious technical limitations in deep enclosed spaces. First, the wind feeling lacks naturalness, and the air speed changes mainly rely on motor speed control, resulting in a rigid wind feeling and a lack of wind levels and spatial dynamic changes. Second, the air direction changes are single, and existing air direction adjustment is mainly based on swing leaf mechanisms, with limited movement range, making it impossible to form multi-dimensional and multi-regional dynamic air direction changes. In addition, the existing equipment structure is mostly based on fixed outlets, lacking dynamic variability in spatial position, and cannot achieve the combined effect of air source movement and multi-point disturbance, making it difficult to effectively regulate the air flow path at the building scale.

[0004] Chinese patent CN115782508A discloses a control method, device, system and storage medium for an air outlet. This technology generates a three-dimensional user image of the user by receiving three-dimensional image data of the user, determines the target position of the air blowing according to the three-dimensional user image and the air blowing position selected by the user, and then determines the air blowing angle of the target air outlet to control the air outlet to blow accurately. However, it is based on angle adjustment of fixed outlet position, and cannot solve the problem of air flow distribution limitation caused by fixed outlet position, lacking dynamic optimization ability for the overall space air flow environment. SUMMARY

[0005] Therefore, the present application provides a double-outlet constant-air-volume natural wind simulation system based on a sliding rail air duct, which can solve the problem of dynamic optimization of space air flow in the prior art, realize intelligent natural wind simulation control, and improve the naturalness and comfort of space air flow distribution.

[0006] The technical scheme of the present application is implemented as follows: The present application provides a double-air-outlet constant air volume natural wind simulation system based on a sliding rail air duct, comprising: A double-air-outlet sliding rail structure, comprising a side wall air outlet, a top surface air outlet, a sliding rail assembly and a driving mechanism, the sliding rail assembly carries the side wall air outlet and the top surface air outlet to reciprocate along a preset track, and the driving mechanism is used for adjusting the air supply amount of the double air outlets; A sensor network module, comprising a wind speed and direction sensor, a temperature and humidity sensor and a CO2 concentration sensor, used for monitoring the wind speed and direction data and environmental parameters in the space in real time; An intelligent control module, comprising an edge computing controller and a PLC controller, the edge computing controller is used for data processing and inter-module coordination control, and the PLC controller drives the driving mechanism to adjust the air outlet position and air volume through a control signal; A feature learning module, used for extracting and analyzing the time sequence features of the wind speed and direction data and environmental parameters by using an LSTM neural network, and establishing a natural wind feature database; A cooperative control module, used for establishing a cooperative control algorithm based on the interaction of the air flows of the double air outlets, and generating control parameters; wherein the control parameters include air outlet position coordinates and air volume distribution ratios; The wind speed and direction data and environmental parameters are collected by the sensor network module, processed by the edge computing controller, input into the feature learning module for natural wind feature extraction and learning, the control parameters are generated by the cooperative control module based on the natural wind features, and the driving structure is driven by the PLC controller to adjust the position and air volume of the double air outlets according to the control parameters.

[0007] On the basis of the above technical scheme, preferably, the side wall air outlet is installed on the side wall surface of the space and moves horizontally along the side wall surface through a horizontal sliding rail assembly; The top surface air outlet is installed on the top of the space and moves horizontally along the top surface through a horizontal sliding rail assembly; The driving mechanism comprises a servo motor and a variable frequency fan, the servo motor drives the sliding rail assembly to move, and the variable frequency fan adjusts the air supply amount of the air outlet.

[0008] On the basis of the above technical scheme, preferably, the wind speed and direction sensor is installed at the air outlet of the side wall air outlet and the top surface air outlet, the temperature and humidity sensor and the CO2 concentration sensor are distributed and arranged in the monitoring space, and the wind speed sensor is arranged in the air flow intersection area of the side wall air outlet and the top surface air outlet, used for monitoring the air flow superposition effect of the double air outlets.

[0009] On the basis of the above technical scheme, preferably, the steps of the intelligent control module are as follows: The edge computing controller receives the wind speed and direction data and the environmental parameters, and performs sliding average filtering and outlier detection processing on the wind speed and direction data and the environmental parameters to obtain processed data; The processed data are transmitted to a feature learning module and a cooperative control module respectively, and control parameters are generated through the feature learning module and the cooperative control module; The PLC controller receives the control parameters and converts them into execution instructions for the driving mechanism.

[0010] On the basis of the above technical scheme, preferably, the edge computing controller uses a PI controller for adaptive parameter adjustment, specifically including: The target wind speed determined based on the natural wind features is obtained from the feature learning module, and the measured wind speed in the space is obtained from the sensor network module; The deviation value of the target wind speed and the measured wind speed is calculated as an input error signal of the PI controller, and the output compensation signal of the PI controller is calculated according to the preset proportional gain coefficient and integral gain coefficient; The output compensation signal of the PI controller is converted into a wind port air volume fine tuning instruction and a position fine tuning instruction.

[0011] On the basis of the above technical scheme, preferably, the natural wind features include wind speed time sequence features, turbulence intensity features, power spectral density features, and wind direction change features.

[0012] On the basis of the above technical scheme, preferably, the LSTM neural network includes an input layer, two layers of LSTM units, an attention mechanism layer, and an output layer, the input layer receives time window data with a length of 120, each layer of LSTM units contains 64 neurons, and the output layer outputs a 32-dimensional natural wind feature vector.

[0013] On the basis of the above technical scheme, preferably, the cooperative control module specifically includes: The real-time wind port air volume of the double wind port is obtained from the driving mechanism, and the three-dimensional Euclidean distance between the center of the side wall wind port outlet and the center of the top surface wind port outlet is measured; A double wind port airflow velocity field model is established, wherein the double wind port airflow velocity field model includes velocity vector distribution and pressure field distribution; The air volume distribution ratio of the double wind port is calculated through a weight distribution algorithm according to the natural wind features provided by the feature learning module, the spatial coordinates and movement trajectory of the double wind port are determined through a position optimization algorithm according to the interaction intensity coefficient, and the control parameters are generated.

[0014] On the basis of the above technical scheme, preferably, the calculation formula of the interaction intensity coefficient is: ; wherein, denotes the interaction intensity coefficient, denotes the first air port air volume, denotes the second air port air volume, denotes the distance between the two air ports, denotes the mutual influence degree of the two air port air flows.

[0015] Further preferably, the logic of the air volume distribution ratio of the two air ports is as follows: ; ; wherein, denotes the air volume distribution weight coefficient of the first air port at time t, denotes the main variation frequency, denotes the time variable, denotes the random disturbance term, denotes the air volume distribution weight coefficient of the second air port at time t.

[0016] The double air port constant air volume natural wind simulation system based on the sliding rail air duct of the present application has the following beneficial effects relative to the prior art: (1) By adopting a movable double air port, combining the natural wind feature learning and collaborative control algorithm of the LSTM neural network, the spatio-temporal variation characteristics of natural wind can be reproduced in space, and at the same time, through the collaborative movement and dynamic air volume distribution of the double air ports, intelligent natural wind simulation control is realized, and the naturalness and comfort of the space air flow distribution are improved; (2) The double air port sliding rail structure of the side wall air port and the top air port is adopted, the precise position control and collaborative movement of the air port are realized through the driving of the servo motor, and based on the interaction intensity coefficient, the collaborative strategy of the two air ports can be dynamically adjusted according to the real-time change of the air volume distribution between the air ports, so as to avoid the mutual interference of the air flows and realize the intelligent optimization of the air port movement trajectory; (3) By establishing a double air port air flow velocity field model, the interaction mechanism of the two air flows in the space intersection area is accurately described, the risk is dynamically adjusted through weight distribution, and the naturalness of the space air flow is improved. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 is a flow chart of a double air port constant air volume natural wind simulation system based on the sliding rail air duct of the present application; Figure 2 This is a schematic diagram of a dual-outlet constant air volume simulated natural wind system based on a sliding rail duct according to the present invention. Figure 3 This is a schematic diagram of the side wall air outlet and track of a dual-air outlet constant air volume simulated natural wind system based on a sliding rail air duct according to the present invention. Figure 4 This is a schematic diagram of the top and wall air vents and track of a dual-vent constant air volume simulated natural wind system based on a sliding rail duct according to the present invention. Figure 5 This is a schematic diagram of a dual-outlet constant air volume simulated natural wind system based on a sliding rail duct according to the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 and Figure 2 As shown, this invention provides a dual-outlet constant air volume simulated natural wind system based on a sliding rail duct, comprising: The dual-outlet slide rail structure includes a side wall air outlet, a top air outlet, a slide rail assembly, and a drive mechanism. The slide rail assembly carries the side wall air outlet and the top air outlet to reciprocate along a preset track, and the drive mechanism is used to adjust the air supply volume of the dual air outlets. The sensor network module includes wind speed and direction sensors, temperature and humidity sensors, and CO2 concentration sensors, which are used to monitor wind speed and direction data and environmental parameters in the space in real time. The intelligent control module includes an edge computing controller and a PLC controller. The edge computing controller is used for data processing and inter-module coordination control, and the PLC controller drives the drive mechanism to adjust the position and air volume of the air outlet through control signals. The feature learning module is used to extract and analyze the temporal features of wind speed and direction data and environmental parameters using LSTM neural networks, and to establish a natural wind feature database. The natural wind features include wind speed temporal features, turbulence intensity features, power spectral density features, and wind direction change features. The collaborative control module is used to establish a collaborative control algorithm based on the interaction of airflow between the two air outlets and generate control parameters; the control parameters include the coordinates of the air outlet positions and the airflow distribution ratio. The sensor network module collects wind speed and direction data and environmental parameters. The data is then processed by the edge computing controller. The processed data is input into the feature learning module for natural wind feature extraction and learning. The collaborative control module generates control parameters based on the natural wind features. The PLC controller drives the drive structure to adjust the position and air volume of the dual air outlets according to the control parameters.

[0021] Understandably, wind speed time-series characteristics reflect the basic laws governing wind speed changes over time. The average wind speed is calculated by taking the arithmetic mean of all wind speed data within a time window, and includes statistical parameters such as average wind speed, wind speed standard deviation, and wind speed trend. The wind speed standard deviation measures the dispersion of wind speed around the average value. The wind speed trend is determined through linear regression analysis; a positive slope indicates an upward trend, and a negative slope indicates a downward trend. The wind speed time-series variation is as follows: ,in, This represents the instantaneous wind speed at time t. This represents the average wind speed within the time window. Indicates the standard deviation of wind speed. It represents a standardized stochastic process, which is a random variable with a mean of 0 and a standard deviation of 1.

[0022] Turbulence intensity characteristics characterize the severity of wind speed fluctuations. By monitoring and analyzing turbulence intensity characteristics, the naturalness of the current airflow can be assessed, and control strategies can be adjusted accordingly. Turbulence intensity is... , Indicates turbulence intensity. This represents the standard deviation of wind speed fluctuations.

[0023] Power spectral density characteristics describe the energy distribution of wind speed fluctuations at different frequency components, reflecting the frequency domain characteristics of wind speed changes. By fitting the power spectral density curve, slope parameters and energy distribution characteristics are extracted to guide the formulation of natural wind mimicry control strategies. The power spectral density function is: , Represents the power spectral density. Represents a proportionality constant. Indicates frequency, The parameter representing the negative slope of the power spectrum.

[0024] The characteristics of wind direction change reflect the spatial fluidity and directional variability of natural wind.

[0025] This invention employs movable dual air outlets, combined with LSTM neural network-based natural wind feature learning and collaborative control algorithms, to reproduce the spatiotemporal variation characteristics of natural wind in space. Simultaneously, through the collaborative movement of the dual air outlets and dynamic airflow distribution, it eliminates the ventilation dead zones and local heat island effects commonly found in fixed air outlet systems, achieving intelligent natural wind mimicry control and enhancing the naturalness and comfort of spatial airflow distribution.

[0026] In one embodiment of the present invention, the side wall vent is installed on the side wall of the space and moves horizontally back and forth along the side wall via a horizontal slide rail assembly; The top air vent is installed at the top of the space and moves horizontally back and forth along the top plane via a horizontal slide rail assembly; The drive mechanism includes a servo motor and a variable frequency fan. The servo motor drives the slide rail assembly to move, and the variable frequency fan adjusts the air volume delivered by the air outlet.

[0027] Understandable, such as Figure 3 and Figure 5 As shown, the sidewall air vent includes an air vent body 1 and a matching slide rail assembly 2. The air vent body 1 is installed on the side wall of the space and moves horizontally back and forth along the side wall via the slide rail assembly 2. The slide rail assembly 2 adopts a mechanical guiding structure, including a track guide mechanism arranged horizontally along the wall and a slider unit fixedly installed on the back of the air vent body. The slider unit cooperates with the track guide mechanism through ball bearings, allowing the air vent body to slide smoothly left and right within a preset horizontal track range. The servo motor adopts a servo motor drive method and is connected to the slider unit through a transmission mechanism, enabling precise program control of the air vent position. The air vent body 1 is connected to a variable frequency fan system through a flexible air duct, and the variable frequency fan is responsible for providing an adjustable air volume.

[0028] like Figure 4 and Figure 5 As shown, the air vent body 3 of the top air vent is installed below the top structure of the space and is connected to the top panel structure through the slide rail assembly 4, realizing the function of sliding and moving horizontally along the ceiling plane. The slide rail assembly 4 includes an embedded guide rail groove structure and a sliding bracket system installed on the back of the air vent body. The guide rail groove is arranged in the top structure along a predetermined path, and the sliding bracket moves smoothly in the guide rail groove through a precision sliding mechanism.

[0029] Specifically, the side wall vents adopt a grille structure, accounting for 50% to 100% of the air volume. They are mainly used for directional air supply in areas with high foot traffic. By continuously varying the air volume, the local average wind speed can be flexibly adjusted to enhance the gentleness and comfort of the air supply. The ceiling vents adopt a vortex air supply structure, accounting for 0% to 50% of the air volume. They are mainly used for the diffusion and disturbance of the overall airflow, forming a large-scale air mixing, improving the uniformity of the space, and ensuring the control of the overall turbulence in the room. The two vents can operate independently or in coordination. By adjusting their position and air volume ratio, they can form regional alternating air supply, breaking the limitations of traditional single-point air supply methods.

[0030] Among them, the top vortex air outlet can balance the distribution of indoor airflow, avoid strong wind concentration, and enhance the gentleness of indoor airflow, providing a more natural wind experience; the side wall grille air outlet, through directional air supply, makes the wind speed and turbulence in specific areas more in line with comfort requirements, providing more precise air flow and adapting to the comfort needs of different activity areas.

[0031] This invention employs a dual-air vent sliding rail structure with sidewall and top surface air vents. A servo motor drives precise position control and coordinated movement of the air vents. Based on the interaction strength coefficient, the invention dynamically adjusts the coordination strategy of the two air vents according to real-time changes in airflow distribution between them, avoiding mutual airflow interference and achieving intelligent optimization of the air vent movement trajectory. In one embodiment of the present invention, the wind speed and direction sensor is installed at the air outlet of the side wall air outlet and the top air outlet, the temperature and humidity sensor and the CO2 concentration sensor are distributed in the monitoring space, and the wind speed sensor is set in the airflow convergence area of ​​the side wall air outlet and the top air outlet to monitor the superposition effect of the airflow from the two air outlets.

[0032] The number of sensors is determined by the available space.

[0033] Understandably, the wind speed and direction sensors, employing ultrasonic measurement principles, are installed at the air outlets of the side wall and roof vents. Temperature and humidity sensors are distributed at key locations within the monitored space, while the CO2 concentration sensor utilizes non-dispersive infrared detection. The determination of the airflow convergence area between the side wall and roof vents is based on computational fluid dynamics simulation analysis, typically located near the midpoint of the line connecting the two vents. All sensors are connected via a wireless sensor network using the ZigBee wireless communication protocol, with a star topology. The data aggregation node is connected to the intelligent control module via a wired connection.

[0034] In one embodiment of the present invention, the steps of the intelligent control module are as follows: The edge computing controller receives wind speed and direction data and environmental parameters, and performs moving average filtering and outlier detection on the wind speed and direction data and environmental parameters to obtain the processed data; The processed data is transmitted to the feature learning module and the collaborative control module respectively, and control parameters are generated through the feature learning module and the collaborative control module. The PLC controller receives control parameters and converts them into execution instructions for the drive mechanism.

[0035] Understandably, the intelligent control module operates according to timing control. The edge computing controller receives wind speed and direction data and environmental parameters collected by the sensor network module. The received data first undergoes a data integrity check, eliminating obviously erroneous data points. Then, the wind speed and direction data and environmental parameters are processed using a moving average filter with a filter window length of 10 sampling points. Outlier detection uses the 3σ criterion; when a data point deviates from the historical average by more than three times the standard deviation, that data point is marked as an outlier and replaced with valid data from the previous time step. After data preprocessing, the edge computing controller transmits the processed data to the feature learning module and the collaborative control module respectively. The edge computing controller is responsible for coordinating the data flow and control timing between the modules, ensuring that the feature learning module and the collaborative control module can obtain the required input data according to the predetermined timing.

[0036] The PLC controller converts the received control parameters into execution instructions for the drive mechanism through its internal program logic. For position control, the PLC controller calculates the deviation between the current position and the target position and generates a corresponding pulse sequence to send to the servo motor driver. For airflow control, the PLC controller calculates the target airflow for each air outlet based on the airflow distribution ratio and then converts it into an analog signal output to the frequency converter.

[0037] Specifically, the edge computing controller uses a PI controller for adaptive parameter adjustment, including: The target wind speed, determined based on natural wind features, is obtained from the feature learning module, and the measured wind speed in space is obtained from the sensor network module. The deviation between the target wind speed and the measured wind speed is calculated as the input error signal of the PI controller, and the output compensation signal of the PI controller is calculated based on the preset proportional gain coefficient and integral gain coefficient. The output compensation signal of the PI controller is converted into fine-tuning commands for airflow and position at the vent.

[0038] Understandably, real-time wind speed data is acquired from multiple wind speed sensors distributed throughout the monitoring space. The measured wind speed value at each control point is calculated using a spatial interpolation algorithm, employing the inverse distance weighting method. When the target wind speed is greater than the measured wind speed, the deviation of the PI controller is positive, indicating that the airflow needs to be increased; when the target wind speed is less than the measured wind speed, the deviation of the PI controller is negative, indicating that the airflow needs to be reduced.

[0039] In one embodiment of the present invention, the proportional gain coefficient of the PI controller is set to 0.5, and the integral gain coefficient is set to 0.1.

[0040] In one embodiment of the present invention, the LSTM neural network includes an input layer, two LSTM units, an attention mechanism layer, and an output layer. The input layer receives time window data of length 120, each LSTM unit contains 64 neurons, and the output layer outputs a 32-dimensional natural wind feature vector.

[0041] Understandably, the first LSTM layer is responsible for extracting short-term temporal patterns, while the second LSTM layer further extracts long-term temporal patterns based on the output of the first layer. Each LSTM layer contains three gating mechanisms: a forget gate, an input gate, and an output gate. The attention mechanism layer weights the outputs of the two LSTM layers according to their importance. When changes in wind speed or direction at certain moments have a significant impact on the current control strategy, the attention mechanism assigns higher weights to these moments, thereby improving the accuracy and specificity of feature extraction. The output layer is a fully connected neural network layer that maps the outputs of the LSTM layer and the attention layer into a 32-dimensional natural wind feature vector, which includes statistical parameters of various temporal features, frequency domain feature parameters, spatial correlation parameters, etc.

[0042] In one embodiment of the present invention, the loss function of the LSTM neural network adopts a combination of reconstruction error and feature regularization, and the calculation formula is as follows: in, This represents the total loss function value. This indicates the number of wind speed time-series data samples used in the training. Indicates the sample index. This represents the original input data for the i-th sample. This represents the reconstructed output data for the i-th sample. Describing the L2 norm, This represents the regularization coefficient, which is 0.0001. This represents the 32-dimensional natural wind feature vector corresponding to the i-th sample. This represents the L1 norm.

[0043] Understandable, regularization term By using the L1 norm penalty mechanism, the network is prompted to learn sparse feature representations, which can prevent overfitting and improve generalization ability.

[0044] Furthermore, the feature learning module learns natural wind feature vectors through LSTM neural networks, which include turbulence intensity features and power spectral density features. The collaborative control module generates control parameters based on these features, enabling the system to maintain turbulence intensity within the range of 0.2 to 0.7 and the double logarithmic power spectral index after wind speed Fourier transform within the range of 1.1 to 2.0, thereby simulating the fluctuation of natural wind and improving the comfort of the space.

[0045] The extraction of turbulence intensity and power spectral density features is achieved through time-series analysis using an LSTM neural network. The network automatically identifies the time-domain and frequency-domain patterns of wind speed fluctuations and encodes these patterns into corresponding dimensions in the feature vector. Upon receiving the feature vector, the collaborative control module maps the turbulence intensity features to a dynamic adjustment strategy for the air outlet location and airflow, and converts the power spectral density features into a time-series control mode for air outlet movement. This ensures that the airflow characteristics in the personnel activity area always meet the control requirements for simulated natural wind.

[0046] In one embodiment of the present invention, the collaborative control module specifically includes: The real-time airflow of the dual air outlets is obtained from the drive mechanism, and the three-dimensional Euclidean distance between the center of the side wall air outlet and the center of the top surface air outlet is measured. A velocity field model for airflow from two vents is established, which includes velocity vector distribution and pressure field distribution. Based on the natural wind features provided by the feature learning module, the airflow distribution ratio between the two air outlets is calculated using a weighted allocation algorithm. Based on the interaction strength coefficient, the spatial coordinates and movement trajectory of the two air outlets are determined using a location optimization algorithm, generating control parameters. The formula for calculating the interaction strength coefficient is as follows: ; in, Represents the interaction strength coefficient. This indicates the air volume at the first air inlet. This indicates the air volume at the second air inlet. Indicates the distance between the two air outlets. This indicates the degree of mutual influence between the airflow from the two air outlets.

[0047] Specifically, the logic behind the airflow distribution ratio of the dual air outlets is as follows: ; ; in, This represents the weighting coefficient for airflow distribution at the first air outlet at time t. Indicates the frequency of major changes. Represents a time variable. Represents the random disturbance term. This represents the weighting coefficient for airflow distribution at the second air outlet at time t.

[0048] Understandably, dual air outlets refer to the first and second air outlets, namely the side wall outlet and the roof outlet. These two outlets can operate independently or work collaboratively simultaneously at specific times, based on real-time environmental monitoring results and natural wind characteristic learning results, to optimize and regulate the airflow environment in a local area or the entire space. The interaction intensity coefficient quantifies the degree of mutual influence between the airflows from the two outlets in the spatial convergence area. A larger value indicates strong interaction between the airflows at the two vents, requiring a strongly coupled collaborative control strategy; when... A smaller value indicates a weaker interaction, and a relatively independent control strategy can be adopted.

[0049] This invention establishes a dual-ventilation airflow velocity field model, which accurately describes the interaction mechanism of two airflows in the spatial convergence area. It solves the problem that traditional multi-ventilation systems lack consideration of airflow interaction and improves the naturalness of spatial airflow by dynamically adjusting the risk through weight allocation.

[0050] In one embodiment of the present invention, the moving trajectory of the dual-outlet sliding rail structure and the air volume output change curve are arbitrarily combined and dynamically adjusted through an intelligent algorithm, as follows: in, This represents the horizontal coordinate of the side wall vent at time t. Indicates the coordinates of the center position of the side wall air vent. Parameters representing the movement range of the side wall air vents. Indicates the angular frequency of the movement. This represents the horizontal coordinate of the top air vent at time t. Indicates the coordinates of the center position of the top air vent. This parameter represents the range of movement of the top-mounted air vent. This indicates the phase difference between the movements of the two air outlets.

[0051] Phase difference Based on the aforementioned interaction strength coefficient Dynamic adjustment ensures that when the interaction strength between the two air outlets is high, the two air outlets tend to move synchronously (with a small phase difference), and when the interaction strength is low, the two air outlets move more independently (with a larger phase difference).

[0052] Furthermore, this invention, through intelligent learning of the LSTM neural network and algorithm optimization of the collaborative control module, can accurately reproduce the spatiotemporal variation characteristics of natural wind. In actual operation, the system monitors and dynamically adjusts airflow parameters in the area where people are active in real time. Through intelligent movement of the air outlet position and dynamic distribution of air volume, the system can effectively simulate the spatiotemporal variation characteristics of natural wind, creating a comfortable air supply experience close to natural wind.

[0053] During system operation, the dual air outlets, based on the natural wind characteristics extracted by the feature learning module, and through the control parameters generated by the collaborative control module, achieve intelligent combination and adjustment of the position movement trajectory and air volume output curve, forming a dynamic, non-directional, and non-constant airflow distribution state. This overcomes the problems of insufficient airflow disturbance and monotonous wind feel in traditional fixed air outlet systems, achieving a technological breakthrough in simulating natural wind effects.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dual-outlet constant air volume simulated natural wind system based on a sliding rail duct, characterized in that: include: The dual-outlet slide rail structure includes a side wall air outlet, a top air outlet, a slide rail assembly, and a drive mechanism. The slide rail assembly carries the side wall air outlet and the top air outlet to reciprocate along a preset track, and the drive mechanism is used to adjust the air supply volume of the dual air outlets. The sensor network module includes wind speed and direction sensors, temperature and humidity sensors, and CO2 concentration sensors, which are used to monitor wind speed and direction data and environmental parameters in the space in real time. The intelligent control module includes an edge computing controller and a PLC controller. The edge computing controller is used for data processing and inter-module coordination control, and the PLC controller drives the drive mechanism to adjust the position and air volume of the air outlet through control signals. The feature learning module is used to extract and analyze time-series features of wind speed and direction data and environmental parameters using LSTM neural networks, and to establish a natural wind feature database. The collaborative control module is used to establish a collaborative control algorithm based on the interaction of airflow between the two air outlets and generate control parameters; the control parameters include the coordinates of the air outlet positions and the airflow distribution ratio. The sensor network module collects wind speed and direction data and environmental parameters. The data is then processed by the edge computing controller. The processed data is input into the feature learning module for natural wind feature extraction and learning. The collaborative control module generates control parameters based on the natural wind features. The PLC controller drives the drive structure to adjust the position and air volume of the dual air outlets according to the control parameters.

2. The dual-outlet constant air volume simulated natural wind system based on sliding rail duct as described in claim 1, characterized in that: The side wall vents are installed on the side wall of the space and move horizontally back and forth along the side wall via a horizontal slide rail assembly; The top air vent is installed at the top of the space and moves horizontally back and forth along the top plane via a horizontal slide rail assembly; The drive mechanism includes a servo motor and a variable frequency fan. The servo motor drives the slide rail assembly to move, and the variable frequency fan adjusts the air volume delivered by the air outlet.

3. The dual-outlet constant air volume simulated natural wind system based on sliding rail duct as described in claim 1, characterized in that: The wind speed and direction sensors are installed at the air outlets of the side wall air vents and the top air vents. The temperature and humidity sensors and the CO2 concentration sensors are distributed in the monitoring space. The wind speed sensor is set in the airflow convergence area of ​​the side wall air vents and the top air vents to monitor the superposition effect of the airflow from the two air vents.

4. The dual-outlet constant air volume simulated natural wind system based on sliding rail duct as described in claim 1, characterized in that: The steps of the intelligent control module are as follows: The edge computing controller receives wind speed and direction data and environmental parameters, and performs moving average filtering and outlier detection on the wind speed and direction data and environmental parameters to obtain the processed data; The processed data is transmitted to the feature learning module and the collaborative control module respectively, and control parameters are generated through the feature learning module and the collaborative control module. The PLC controller receives control parameters and converts them into execution instructions for the drive mechanism.

5. A dual-outlet constant-volume simulated natural wind system based on a sliding rail duct as described in claim 4, characterized in that: The edge computing controller uses a PI controller for adaptive parameter adjustment, specifically including: The target wind speed, determined based on natural wind features, is obtained from the feature learning module, and the measured wind speed in space is obtained from the sensor network module. The deviation between the target wind speed and the measured wind speed is calculated as the input error signal of the PI controller, and the output compensation signal of the PI controller is calculated based on the preset proportional gain coefficient and integral gain coefficient. The output compensation signal of the PI controller is converted into fine-tuning commands for airflow and position at the vent.

6. The dual-outlet constant air volume simulated natural wind system based on sliding rail duct as described in claim 1, characterized in that: The natural wind characteristics include wind speed temporal characteristics, turbulence intensity characteristics, power spectral density characteristics, and wind direction variation characteristics.

7. The dual-outlet constant air volume simulated natural wind system based on sliding rail duct as described in claim 1, characterized in that: The LSTM neural network includes an input layer, two LSTM units, an attention mechanism layer, and an output layer. The input layer receives time window data of length 120, each LSTM unit contains 64 neurons, and the output layer outputs a 32-dimensional natural wind feature vector.

8. A dual-outlet constant-volume simulated natural wind system based on a sliding rail duct as described in claim 1, characterized in that: The collaborative control module specifically includes: The real-time airflow of the dual air outlets is obtained from the drive mechanism, and the three-dimensional Euclidean distance between the center of the side wall air outlet and the center of the top surface air outlet is measured. A velocity field model for airflow from two vents is established, which includes velocity vector distribution and pressure field distribution. Based on the natural wind features provided by the feature learning module, the air volume distribution ratio of the two air outlets is calculated through a weight allocation algorithm. Based on the interaction intensity coefficient, the spatial coordinates and movement trajectory of the two air outlets are determined through a position optimization algorithm, and control parameters are generated.

9. A dual-outlet constant-volume simulated natural wind system based on a sliding rail duct as described in claim 1, characterized in that: The formula for calculating the interaction strength coefficient is as follows: ； in, Represents the interaction strength coefficient. This indicates the air volume at the first air inlet. This indicates the air volume at the second air inlet. Indicates the distance between the two air outlets. This indicates the degree of mutual influence between the airflow from the two air outlets.

10. A dual-outlet constant-volume simulated natural wind system based on a sliding rail duct as described in claim 1, characterized in that: The logic for the airflow distribution ratio of the dual air outlets is as follows: ； ； in, This represents the weighting coefficient for airflow distribution at the first air outlet at time t. Indicates the frequency of major changes. Represents a time variable. Represents the random disturbance term. This represents the weighting coefficient for airflow distribution at the second air outlet at time t.

Images