Building skin regulation and control method and system based on multi-mode environment perception

Abstract

The invention provides a building skin regulation and control method and system based on multi-modal environment perception, and relates to the technical field of building skins. Structural feature data and environment feature data of a building skin are obtained, and the building skin is regulated and controlled based on the structural feature data and the environment feature data; and a coupling regulation model is adopted to predict a comprehensive regulation strategy of the building skin at the future moment, and an execution device is controlled to execute the comprehensive regulation strategy, so that the building comprehensive energy efficiency and the indoor environment standard reaching rate are improved, the service life of the device is prolonged, and self-adaptive building skin regulation of a'perception-decision-execution 'chain is realized.

Description

Technical Field

[0001] This invention relates to the field of building skin technology, and more specifically, to a method and system for regulating building skin based on multimodal environmental perception. Background Technology

[0002] In existing technologies, when data-driven control strategies are used to regulate the building skin of composite building envelopes, the control strategies optimize energy consumption or comfort in isolation, and the actuators lack a time-series coordination mechanism. For example, when the light transmittance adjustment of electrochromic glass overlaps with the opening and closing of louvers, the probability of indoor illuminance fluctuations exceeding ±300 lux reaches 41%, which seriously damages visual comfort; and the frequent triggering of control equipment reduces the lifespan of the control equipment. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a building skin control method and system based on multimodal environmental perception, so as to improve the overall energy efficiency of buildings, the indoor environmental compliance rate and extend the life of equipment.

[0004] Firstly, this application provides a building skin control method based on multimodal environment perception, including: Acquire structural and environmental feature data of the building facade; Based on structural and environmental feature data, a coupled control model is used to predict the comprehensive control strategy for the building facade at future times. The coupled control model includes a predictive model and a time-series model. The predictive model extracts features from the environmental feature data to obtain time-series and local feature data, and then calculates the probability of future fresh air demand for the building facade based on these data. The time-series model, based on the structural and environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building facade, determines the control parameters of each actuator required to regulate the building facade temperature at future times. These actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the following for each actuator: activation time, angle, and light transmittance. Control the execution equipment to implement a comprehensive control strategy.

[0005] Optionally, structural and environmental feature data of the building skin can be acquired, including: The internal and external temperatures of the building skin, collected by a temperature measuring array installed on the building skin, are used as structural feature data. The pollutant concentration, solar incident light intensity, and angle collected by environmental sensors installed on the building facade are used as environmental characteristic data.

[0006] Optionally, the building skin control method based on multimodal environment perception provided in this application further includes: Based on the internal temperature of the building skin, an internal temperature matrix is ​​obtained by processing the internal temperature using a spatiotemporal attention mechanism and a sliding time window. Based on the internal temperature matrix, a multi-scale Kalman filter algorithm is used to fuse the external temperature to obtain temperature field data as structural feature data of the building skin.

[0007] Optionally, the prediction model includes: Environmental feature data is received through the input layer; Feature extraction is performed on environmental feature data through a long short-term memory layer to obtain time-series feature data; Local feature data is obtained by extracting features from environmental feature data through temporal convolutional layers; The time-series feature data and local feature data are weighted, fused, and stitched together using a fusion layer. The probability of future fresh air demand for the building facade is obtained by performing feature compression and mapping on the weighted and fused temporal and local feature data through a fully connected layer. The output layer outputs the probability of future fresh air demand. Optionally, the weights and thresholds of the time series model can be optimized, including: A swarm of particles is randomly generated, along with the initial velocity and position of each particle in the swarm; each particle includes the parameter values ​​of each weight and each threshold of the initial time series model. Iterative optimization is performed based on particle swarm optimization until the termination condition is met. Then, iterative training is performed based on the weights and thresholds of the initial time series model optimized in the last iteration. The iterative optimization operation includes: The fitness function is used to calculate the objective function value of each particle in the particle swarm. Based on the calculated objective function values ​​of each particle in the particle swarm and the thermal inertia coefficient of the building skin, the global optimal position and the individual optimal position are updated. Based on the updated global optimal position and individual optimal position, optimize the weights and thresholds of the initial time series model, and update the current position and current velocity of each particle in the particle swarm.

[0008] Optionally, the building skin control method based on multimodal environment perception provided in this application also includes: The objective function value is determined based on the parameters and weights of each execution device.

[0009] Optionally, the building skin control method based on multimodal environment perception provided in this application also includes: Obtain a training dataset; the training dataset includes multiple training sample data; each training sample data includes structural feature data, environmental feature data, and actual comprehensive control strategies; Based on the training dataset, iterative training operations are performed on the initial coupling control model until the termination condition of the iterative training is met. Then, based on the weights and thresholds of the initial coupling control model updated during the last iteration, the coupling control model is obtained. The iterative training operations include: Select target training sample data from the training dataset; The structural feature data and environmental feature data from the target training sample data are input into the initial coupled control model so that the initial coupled control model can predict the building skin at future times based on the structural feature data and environmental feature data. Based on the prediction error between the predicted integrated control strategy and the actual integrated control strategy in the target training sample data, the weights and thresholds of the initial coupled control model are updated.

[0010] Secondly, this application provides a building skin control system based on multimodal environmental perception, comprising: The data acquisition module is used to acquire structural and environmental feature data of the building facade; The temperature control module is used to predict the comprehensive control strategy of the building facade at future times based on structural feature data and environmental feature data, using a coupled control model. The coupled control model includes a prediction model and a time-series model. The prediction model extracts features from the environmental feature data to obtain time-series feature data and local feature data, and then calculates the probability of future fresh air demand for the building facade based on these data. The time-series model, based on the structural feature data, environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building facade, determines the control parameters of each actuator required to regulate the building facade temperature at future times. These actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the following for each actuator: activation time, angle, and light transmittance. The control and execution module is used to control the execution equipment to implement comprehensive control strategies.

[0011] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described building skin control method based on multimodal environment perception.

[0012] Fourthly, this application provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the aforementioned building skin control method based on multimodal environment perception.

[0013] This invention provides a building skin control method and system based on multimodal environmental perception. By acquiring structural and environmental feature data of the building skin, and using a coupled control model to predict the comprehensive control strategy of the building skin at future moments, the invention controls the execution equipment to implement the comprehensive control strategy, thereby improving the building's overall energy efficiency, indoor environmental compliance rate, and extending equipment life. This achieves adaptive building skin control of the "perception-decision-execution" chain.

[0014] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 The flowchart illustrates a building skin control method based on multimodal environment perception provided by an embodiment of the present invention. Figure 2 This diagram illustrates a structural schematic of a building skin control system based on multimodal environmental perception provided in an embodiment of the present invention. Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0018] To facilitate a better understanding of this application by those skilled in the art, the technical terms used in this application will be briefly introduced below.

[0019] Structural characteristic data, used to represent the structural characteristics of the building skin, includes, in this application, the internal and external temperatures of the building skin.

[0020] Environmental characteristic data, which is used to represent the environmental characteristics of the building facade, includes pollutant concentration, solar incident light intensity and angle, etc. in this application.

[0021] Electrochromic glass modules trigger ion migration phase transitions based on transmittance commands, and a louver mechanism performs asymmetric opening and closing actions (such as a dragonfly-wing-like structure to reduce wind resistance). The cadmium telluride photovoltaic module adjusts its tilt angle in real time to maximize the reception of scattered radiation. It also employs dual-pulse drive technology (such as a forward pulse triggering Li+ embedding and a reverse pulse accelerating the uniform distribution of ions) to shorten the transmittance adjustment response time to ≤45 seconds. Through a transmittance-photovoltaic efficiency mapping table, the optimal transmittance range is dynamically selected based on the temperature of the cadmium telluride module (such as locking the transmittance to ≥40% when the module temperature is >60℃ to prevent hot spot effects).

[0022] The biomimetic louver component, based on an asymmetric airfoil aerodynamic model, calculates the wind pressure coefficient at different opening angles and generates a minimum wind resistance trajectory. It implements a graded anti-interference strategy: when the PM2.5 concentration is between 50-100 μg / m³, it activates the cyclone dust removal mode (louver high-frequency vibration + 30° tilt angle); when it is >100 μg / m³, it switches to the sealed mode.

[0023] For photovoltaic modules, a scattered radiation tracking algorithm was developed. By identifying cloud movement trends through hemispherical images of the sky, the optimal module tilt angle for the next 5 minutes was predicted. A maximum power point tracking (MPPT) compensation module was embedded. When changes in transmittance caused fluctuations in output power, a variable step size perturbation observation method was used to quickly lock the new operating point.

[0024] After introducing the technical terms used in this application, the technical solution provided in this application will be described in detail below.

[0025] This application provides a method for controlling building skin based on multimodal environment perception. (See attached document.) Figure 1 As shown in the embodiments of this application, the general flow of the building skin control method based on multimodal environment perception is as follows: Step 110: Obtain structural and environmental feature data of the building skin.

[0026] In this embodiment of the application, the structural feature data and environmental feature data of the building skin can be obtained through the following methods: The internal and external temperatures of the building skin, collected by a temperature measuring array installed on the building skin, are used as structural feature data. The pollutant concentration, solar incident light intensity, and angle collected by environmental sensors installed on the building facade are used as environmental characteristic data.

[0027] Furthermore, the temperature measurement array can be a spiral distributed fiber optic temperature measurement array laid inside the building skin layer, or it can be an infrared thermal imager array spliced ​​together by multiple infrared thermal imagers to collect the internal temperature of the building skin; the temperature measurement array can be multiple infrared thermal imagers set on the outside of the building skin to collect the external temperature of the building skin; the environmental sensor can be an aerosol particle size spectrum sensor or a laser scattering PM2.5 sensor set on the outside of the building skin to collect the PM2.5 pollutant concentration; the environmental sensor can be a solar radiation sensor set on the outside of the building skin to monitor the incident light intensity and angle.

[0028] Furthermore, after acquiring the structural and environmental feature data of the building skin, the process also includes: processing the internal temperature of the building skin using a spatiotemporal attention mechanism and a sliding time window to obtain an internal temperature matrix; and fusing the external temperature using a multi-scale Kalman filter algorithm based on the internal temperature matrix to obtain temperature field data as the structural feature data of the building skin.

[0029] In practice, based on the collected internal temperature of the building skin, the continuous internal temperature data is segmented into segments using a fixed sliding time window (30 seconds / time). The spatial and temporal feature data of the segmented internal temperature are then extracted using a spatiotemporal attention mechanism to obtain the internal temperature matrix. Then, the Kalman filter prediction and update iteration mechanism is used to correct the errors of the internal temperature matrix data and the external temperature data at multiple scales, eliminating spatial drift errors caused by environmental interference and equipment aging.

[0030] Step 120: Based on structural feature data and environmental feature data, a coupled control model is used to predict the comprehensive control strategy of the building facade at future times. The coupled control model includes a prediction model and a time-series model. The prediction model extracts features from the environmental feature data to obtain time-series feature data and local feature data, and obtains the probability of future fresh air demand for the building facade based on the time-series and local feature data. The time-series model, based on the structural feature data, environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building facade, determines the control parameters of each actuator required to regulate the building facade temperature at future times. The actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the following for each actuator: start-up time, angle, and light transmittance.

[0031] In this embodiment, the coupling control model includes a prediction model and a temporal model; the prediction model is a Long Short-Term Memory-Temporal Convolutional Network (LSTM-TCN). This application utilizes a hybrid network model (Convolutional Network) to address the shortcomings of single long short-term memory (LSM) models in capturing local mutation features and the weakness of single temporal convolutional networks in modeling long-term dependencies. By extracting both long-term dependencies and local fine-grained features, it achieves accurate probabilistic prediction of fresh air demand. The prediction model includes an input layer, two LSM layers, a temporal convolutional layer, a fusion layer, two fully connected layers, and an output layer. The input layer receives environmental feature data (such as PM2.5 pollutant concentration, CO2 concentration data, and the intensity and angle of solar incident light). The LSM layer captures long-term features of the environmental feature data through 256 neurons. The temporal convolutional layer captures local mutation features of the environmental feature data through convolutional kernels. The fusion layer uses weighted fusion to splice long-term features and local mutation features. The fully connected layer performs feature compression and mapping on the spliced ​​long-term features and local mutation features to obtain the probability of future fresh air demand for the building facade. Finally, the output layer outputs the probability of future fresh air demand. The time-series model, based on the MPC framework, employs an improved particle swarm optimization algorithm and a weighted adaptive decay strategy (i.e., the decay factor is positively correlated with the building skin's thermal inertia coefficient) to accelerate convergence to the Pareto optimal solution. Specifically, the time-series model processes structural characteristic data, environmental characteristic data, and the probability of future fresh air demand to determine the control parameters of each actuator required to regulate the building skin temperature in the future. This resolves the conflict between energy efficiency and comfort goals in traditional systems and improves the efficiency of the actuators. The specific processing steps of the time-series model are as follows: A swarm of particles is randomly generated, along with the initial velocity and position of each particle in the swarm; each particle includes the parameter values ​​of each weight and each threshold of the initial time series model. Iterative optimization is performed based on particle swarm optimization until the termination condition is met. Then, iterative training is performed based on the weights and thresholds of the initial time series model optimized in the last iteration. The iterative optimization operation includes: The fitness function is used to calculate the objective function value of each particle in the particle swarm. Based on the calculated objective function values ​​of each particle in the particle swarm and the thermal inertia coefficient of the building skin, the global optimal position and the individual optimal position are updated. Based on the updated global optimal position and individual optimal position, optimize the weights and thresholds of the initial time series model, and update the current position and current velocity of each particle in the particle swarm.

[0032] Furthermore, the termination condition for iterative optimization is that the number of iterations is not less than the threshold number.

[0033] Furthermore, the objective function value is obtained using the following formula:

[0034] In the formula, Calculate the value for the objective function. , and These are the weighting coefficients. For air conditioning cooling energy consumption, The power generation efficiency of photovoltaic modules. for The cost of controlling concentration data; The weighting coefficients are adjusted based on real-time operating conditions such as photovoltaic module temperature, grid electricity price during certain periods, and personnel density to balance the priorities of different objectives.

[0035] Furthermore, the time-series model also considers the heat transfer delay time of building walls. (i.e., the thermal inertia coefficient of the building skin) and each actuator, determine the timing supplement sequence of each actuator so that the opening and closing action of the bionic louver assembly is synchronized with the temperature wave propagation of the electrochromic glass assembly.

[0036] Step 130: Control the execution equipment to implement the comprehensive control strategy.

[0037] In this embodiment, the execution devices include, but are not limited to, an electrochromic glass assembly, a bionic louver assembly, and a photovoltaic assembly; wherein, dual-pulse drive technology is used to adjust the light transmittance of the electrochromic glass assembly; the bionic louver assembly is controlled to open or close to the angle specified in the control parameters; the photovoltaic assembly is controlled to adjust to the tilt angle specified in the control parameters; and each execution device is started according to the start time specified in the control parameters.

[0038] In this embodiment of the application, the building skin control method based on multimodal environment perception further includes: Obtain a training dataset; the training dataset includes multiple training sample data; each training sample data includes structural feature data, environmental feature data, and actual comprehensive control strategies; Based on the training dataset, iterative training operations are performed on the initial coupling control model until the termination condition of the iterative training is met. Then, based on the weights and thresholds of the initial coupling control model updated during the last iteration, the coupling control model is obtained. The iterative training operations include: Select target training sample data from the training dataset; The structural feature data and environmental feature data from the target training sample data are input into the initial coupled control model so that the initial coupled control model can predict the building skin at future times based on the structural feature data and environmental feature data. Based on the prediction error between the predicted integrated control strategy and the actual integrated control strategy in the target training sample data, the weights and thresholds of the initial coupled control model are updated.

[0039] In practical implementation, this application constructs a digital twin containing a coupled control model and real meteorological data on the virtual end, compares the measured data with the simulation results, dynamically corrects the MPC control weight parameters, and realizes online hot update of the control strategy through edge computing nodes; The twin construction employs multi-resolution modeling technology, achieving a mesh accuracy of 1mm in key areas (such as curtain wall joints), while using surrogate models to simplify calculations in non-critical areas. A random forest weather forecast corrector is embedded, and local weather station data is integrated in real-time to correct the virtual terminal's input boundary conditions. Strategy optimization utilizes a control parameter sensitivity analysis module, identifying key influencing factors in the MPC weight coefficients through Sobol sequence sampling. The Deep Deterministic Policy Gradient (DDPG) algorithm is used to train the control strategy in the virtual environment, aiming to minimize the weighted sum of energy consumption and equipment wear. Online updates employ an incremental strategy migration protocol, transmitting only the differential parameters of the control strategy (average data volume <50KB / update) to adapt to the low-bandwidth environment of edge computing nodes. A safe rollback mechanism automatically switches to the historically optimal version and triggers manual review when a new strategy causes a decrease in simulation metrics >5%.

[0040] The building skin control method based on multimodal environmental perception provided in this application achieves millimeter-level three-dimensional reconstruction of the building skin temperature field through a distributed fiber optic temperature measurement array (spatial resolution ≤ 5 cm) and a spatiotemporal attention mechanism algorithm. Compared with traditional single-point temperature measurement methods, the accuracy of identifying local overheated areas is improved to 98.7%. Combined with the adaptive sampling technology of an aerosol particle size spectrum sensor, the PM2.5 concentration monitoring response time is shortened to within 3 seconds, significantly improving the early warning capability of sudden pollution events. Furthermore, by combining a thermal inertia compensation algorithm with the thermal capacity characteristics of the building envelope, [the method further enhances the building skin control method]. Optimizing the timing of actuator actions has shown that the number of times the electrochromic glass is driven has decreased from an average of 127 times per day to 82 times (a reduction of 35%), the lifespan of the louver mechanism motor has been extended from 5 years to 7.3 years, and equipment maintenance costs have been reduced by 28%. The digital twin platform, through the DDPG algorithm and incremental policy migration protocol, enables online optimization and second-level updates of control strategies. Under sudden operating conditions such as sandstorms, the system adjustment delay has been reduced from 6.2 hours in the traditional scheme to <15 minutes, the duration of PM2.5 exceeding the standard has been reduced by 76%, and the error rate of virtual and real system parameters has been stabilized at <1.2%. This application provides a building skin control system based on multimodal environment perception. (See attached document.) Figure 2 As shown in the embodiments of this application, the building skin control system based on multimodal environment perception includes: Data acquisition module 210 is used to acquire structural feature data and environmental feature data of the building skin; The temperature control module 220 is used to predict the comprehensive control strategy of the building facade at future times based on structural feature data and environmental feature data, using a coupled control model. The coupled control model includes a prediction model and a time-series model. The prediction model extracts features from the environmental feature data to obtain time-series feature data and local feature data, and then obtains the probability of future fresh air demand for the building facade based on the time-series and local feature data. The time-series model, based on the structural feature data, environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building facade, determines the control parameters of each actuator required to adjust the building facade temperature at future times. The actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the following for each actuator: start-up time, angle, and light transmittance. The control and execution module 230 is used to control the execution equipment to execute the comprehensive control and execution strategy.

[0041] It should be noted that the principle of the building skin control system based on multimodal environment perception provided in this application embodiment to solve the technical problem is similar to the building skin control method based on multimodal environment perception provided in this application embodiment. Therefore, the implementation of the building skin control system based on multimodal environment perception provided in this application embodiment can refer to the implementation of the building skin control method based on multimodal environment perception provided in this application embodiment, and the repeated parts will not be described again.

[0042] After introducing the building skin control method and device based on multimodal environment perception provided in the embodiments of this application, the electronic device provided in the embodiments of this application will be briefly introduced next.

[0043] See Figure 3 As shown, the electronic device 500 provided in this application embodiment includes at least a processor 501, a memory 502, and a computer program stored in the memory 502 and executable on the processor 501. When the processor 501 executes the computer program, it implements the building skin control method based on multimodal environment perception provided in this application embodiment.

[0044] The electronic device 500 provided in this application embodiment may further include a bus 503 connecting different components (including processor 501 and memory 502). The bus 503 represents one or more types of bus structures, including memory bus, peripheral bus, local area bus, etc.

[0045] Memory 502 may include a readable storage medium in the form of volatile memory, such as random access memory (RAM) 5021 and / or cache memory 5022, and may further include read-only memory (ROM) 5023. Memory 502 may also include a program tool 5025 having a set (at least one) of program modules 5024, including but not limited to an operating subsystem, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0046] Processor 501 can be a single processing element or a collective term for multiple processing elements. For example, processor 501 can be a central processing unit (CPU), or one or more integrated circuits configured to implement the building skin control method based on multimodal environment perception provided in the embodiments of this application. Specifically, processor 501 can be a general-purpose processor, including but not limited to CPUs, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0047] Electronic device 500 can communicate with one or more external devices 504 (e.g., keyboard, remote control, etc.), and also with one or more devices that enable a user to interact with electronic device 500 (e.g., mobile phone, computer, etc.), and / or with devices that enable electronic device 500 to communicate with one or more other electronic devices 500 (e.g., router, modem, etc.). This communication can be performed through input / output (I / O) interface 505. Furthermore, electronic device 500 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) through network adapter 506. Figure 3 As shown, network adapter 506 communicates with other modules of electronic device 500 via bus 503. It should be understood that, although... Figure 3As not shown, other hardware and / or software modules may be used in conjunction with the electronic device 500, including but not limited to microcode, device drivers, redundant processors, external disk drive arrays, Redundant Arrays of Independent Disks (RAID) subsystems, tape drives, and data backup storage subsystems.

[0048] It should be noted that, Figure 3 The electronic device 500 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0049] The computer-readable storage medium provided in the embodiments of this application is described below. The computer-readable storage medium provided in the embodiments of this application stores computer instructions, which, when executed by a processor, implement the building skin control method based on multimodal environment perception provided in the embodiments of this application. Specifically, the computer instructions can be built into or installed in the processor, so that the processor can implement the building skin control method based on multimodal environment perception provided in the embodiments of this application by executing the built-in or installed computer instructions.

[0050] In addition, the building skin control method based on multimodal environment perception provided in this application embodiment can also be implemented as a computer program product. The computer program product includes program code, which implements the building skin control method based on multimodal environment perception provided in this application embodiment when running on a processor.

[0051] The computer program product provided in this application embodiment may employ one or more computer-readable storage media, which may be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. Specifically, more specific examples (a non-exhaustive list) of computer-readable storage media include electrical connections with one or more wires, portable disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0052] The computer program product provided in this application embodiment can be a CD-ROM and include program code, and can also run on electronic devices such as computers. However, the computer program product provided in this application embodiment is not limited thereto. In this application embodiment, the computer-readable storage medium can be any tangible medium that contains or stores program code, which can be used by or in conjunction with an instruction execution system, device, or apparatus.

[0053] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0054] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0055] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0056] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for controlling building skin based on multimodal environmental perception, characterized in that, include: Acquire structural and environmental feature data of the building facade; Based on the structural feature data and the environmental feature data, a coupled control model is used to predict the comprehensive control strategy of the building skin at future times. The coupled control model includes a prediction model and a time-series model. The prediction model extracts features from the environmental feature data to obtain time-series feature data and local feature data, and obtains the probability of future fresh air demand for the building skin based on the time-series feature data and the local feature data. The time-series model, based on the structural feature data, the environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building skin, determines the control parameters of each actuator required to adjust the temperature of the building skin at future times. The actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the activation time, angle, and light transmittance of each actuator. Control the execution device to execute the comprehensive control strategy.

2. The building skin control method based on multimodal environment perception according to claim 1, characterized in that, Obtain structural and environmental feature data of the building facade, including: The internal and external temperatures of the building skin, collected by a temperature measuring array installed on the building skin, are obtained as structural feature data. The pollutant concentration, solar incident light intensity, and angle collected by environmental sensors installed on the building's surface are used as environmental characteristic data.

3. The building skin control method based on multimodal environment perception according to claim 2, characterized in that, Also includes: Based on the internal temperature of the building skin, an internal temperature matrix is ​​obtained by processing the internal temperature using a spatiotemporal attention mechanism and a sliding time window. Based on the internal temperature matrix, the external temperature is fused using a multi-scale Kalman filter algorithm to obtain temperature field data as the structural feature data of the building skin.

4. The building skin control method based on multimodal environment perception according to claim 1, characterized in that, The prediction model includes: Environmental feature data is received through the input layer; The environmental feature data is extracted by using a long short-term memory layer to obtain time-series feature data; Local feature data is obtained by extracting features from the environmental feature data through a temporal convolutional layer; The time-series feature data and the local feature data are weighted, fused, and spliced ​​together using a fusion layer. The probability of future fresh air demand for the building skin is obtained by performing feature compression and mapping on the weighted fused temporal feature data and the local feature data through a fully connected layer; The probability of the required fresh air volume at the next future moment is output through the output layer.

5. The building skin control method based on multimodal environment perception according to claim 1, characterized in that, The weights and thresholds of the time series model are optimized, including: A particle swarm is randomly generated, along with the initial velocity and initial position of each particle in the swarm; wherein each particle includes the parameter values ​​of each weight and each threshold of the initial temporal model. Iterative optimization is performed based on the particle swarm optimization until the iterative optimization termination condition is met. Then, iterative training is performed iteratively based on the weights and thresholds of the initial time series model optimized during the last iteration of the optimization operation. The iterative optimization operation includes: The objective function value of each particle in the particle swarm is calculated using the fitness function. Based on the calculated objective function values ​​of each particle in the particle swarm and the thermal inertia coefficient of the building skin, the global optimal position and the individual optimal position are updated. Based on the updated global optimal position and the individual optimal position, the weights and thresholds of the initial time series model are optimized, and the current position and current velocity of each particle in the particle swarm are updated.

6. The building skin control method based on multimodal environment perception according to claim 5, characterized in that, Also includes: The objective function value is determined based on the parameters and weights of each execution device.

7. The building skin control method based on multimodal environment perception according to claim 1, characterized in that, Also includes: Obtain a training data set; wherein the training data set includes multiple training sample data; each training sample data includes structural feature data, environmental feature data, and actual comprehensive control strategy; Based on the training dataset, iterative training operations are performed on the initial coupling control model until the iterative training termination condition is met. Then, based on the weights and thresholds of the initial coupling control model updated during the last execution of the iterative training operation, the coupling control model is obtained; wherein, the iterative training operation includes: Select target training sample data from the training dataset; The structural feature data and environmental feature data from the target training sample data are input into the initial coupled control model so that the initial coupled control model can predict the comprehensive control strategy of the building skin at future times based on the structural feature data and environmental feature data. Based on the prediction error between the predicted integrated control strategy and the actual integrated control strategy in the target training sample data, the weights and thresholds of the initial coupled control model are updated.

8. A building skin control system based on multimodal environmental perception, characterized in that, include: The data acquisition module is used to acquire structural and environmental feature data of the building facade; A temperature control module is used to predict the comprehensive control strategy of the building facade at future times based on the structural feature data and the environmental feature data using a coupled control model. The coupled control model includes a prediction model and a time-series model. The prediction model extracts features from the environmental feature data to obtain time-series feature data and local feature data, and obtains the probability of future fresh air demand for the building facade based on the time-series feature data and the local feature data. The time-series model determines the control parameters of each actuator required to adjust the temperature of the building facade at future times based on the structural feature data, the environmental feature data, the probability of future fresh air demand, and the thermal inertia of the building facade. The actuators include electrochromic glass components, biomimetic louver components, and photovoltaic components. The control parameters include at least one of the activation time, angle, and light transmittance of each actuator. The control and execution module is used to control the execution device to execute the comprehensive control strategy.

9. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the building skin control method based on multimodal environment perception as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the building skin control method based on multimodal environment perception as described in any one of claims 1 to 7.

Images