Spherical glass curtain wall self-adaptive light control system based on spectrum regulation and control

Abstract

The invention relates to the technical field of constructional engineering, and discloses a spherical glass curtain wall self-adaptive light control system based on spectrum regulation and control. The system is composed of a plurality of units such as a multispectral sensor array unit and a dynamic electrochromic film layer unit. The multispectral sensor array unit collects light data and predicts light intensity distribution; the dynamic electrochromic film layer unit intelligently adjusts the light transmittance according to various factors; the spectrum characteristic decomposition unit optimizes spectrum regulation and control; the curved surface incident angle compensation unit accurately controls light for the curtain wall curved surface; and the multi-source data fusion unit integrates the multi-source information decision and predicts the sun position. And the system also has anomaly detection and redundancy design, so that stable operation is ensured. The system realizes intelligent and accurate regulation and control of illumination of the spherical glass curtain wall, effectively optimizes the indoor light environment, and improves the energy utilization efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of building engineering, and in particular to a spherical glass curtain wall self-adaptive light control system based on spectrum regulation. Background Art

[0002] With the vigorous development of the construction industry, glass curtain walls, as a modern building envelope structure, are widely used in various buildings due to their unique aesthetic effects and good lighting performance. Spherical glass curtain walls have become the first choice for many landmark buildings with their novel and unique appearance, adding unique charm to the urban landscape. However, spherical glass curtain walls face many challenges in practical applications.

[0003] From the perspective of lighting, its special spherical curved surface structure makes the light incidence situation extremely complicated. Unlike traditional flat glass curtain walls, the normal direction of each point on the spherical surface is constantly changing, and the incident angle of light varies significantly, which makes the distribution of light on the curtain wall surface extremely uneven. In some areas, the light is overly concentrated, resulting in strong glare, which not only seriously affects the visual comfort of indoor personnel, but also may cause vision loss if they are in this environment for a long time; at the same time, too much solar radiation enters the room, which will cause the indoor temperature to rise rapidly, increase the load of the air-conditioning system, and cause a large amount of energy consumption. In other areas, due to insufficient light, it may not be able to meet normal indoor use needs, and additional artificial lighting equipment needs to be turned on, further increasing energy costs.

[0004] In terms of light control technology, traditional glass curtain wall light control methods are difficult to apply to spherical glass curtain walls. Common fixed shading devices, such as sun visors and blinds, are not only difficult and costly to install when installed on spherical curtain walls, but also cannot be flexibly adjusted according to real-time changes in light. Even some adjustable shading facilities find it difficult to achieve uniform and effective shading effects when faced with the complex curved surfaces of spherical curtain walls. Although electrochromic glass has the function of changing light transmittance, it is difficult to ensure the uniformity of color change on spherical curtain walls due to the influence of curvature changes, and traditional control methods cannot accurately adapt to the complex and changeable lighting conditions of spherical curtain walls, resulting in poor light control effects.

[0005] In addition, as people's requirements for the quality of indoor building environment are increasing, and the world's emphasis on energy conservation and emission reduction continues to deepen, higher requirements are placed on the intelligent light control system of building glass curtain walls. Existing light control systems often fail to fully consider the changing characteristics of light in different seasons and time periods, and it is difficult to combine indoor personnel activities, building functions and other factors for comprehensive regulation. For example, when the sun is strong in summer, it is impossible to effectively block excessive heat from entering the room in a timely manner, resulting in a significant increase in indoor air conditioning energy consumption; in areas where personnel activities are frequent, it is impossible to provide appropriate light intensity according to actual needs. At the same time, most light control systems lack comprehensive collection and in-depth analysis of environmental data, which makes light control decisions lack scientificity and accuracy, and cannot achieve efficient use of energy and optimization of indoor light environment.

[0006] In terms of system stability and reliability, the existing spherical glass curtain wall light control system also has obvious shortcomings. In the face of sudden abnormal lighting conditions, such as sudden changes in lighting caused by cloud cover and interference from reflected light from nearby buildings, the system cannot respond quickly and accurately, which can easily cause drastic fluctuations in the indoor light environment. Moreover, the coordination between the components of the system is poor, and the information transmission and processing efficiency is low. Once a component fails, it may cause the entire light control system to be paralyzed, seriously affecting the normal use of the building. Therefore, the development of a light control system that can adapt to the special structure of the spherical glass curtain wall, achieve precise, intelligent, and efficient light regulation, and has high stability and reliability has become a key issue that needs to be urgently addressed in the construction field. Summary of the invention

[0007] The object of the present invention is to provide a spherical glass curtain wall adaptive light control system based on spectrum regulation to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: a spherical glass curtain wall adaptive light control system based on spectrum regulation, the system comprising: A multispectral sensor array unit is configured to collect the wavelength distribution matrix of incident light in real time, build a light intensity distribution prediction model based on a convolutional neural network, and establish a three-dimensional radiation transfer equation in combination with historical climate data; The dynamic electrochromic film layer unit is configured to generate a transmittance adjustment matrix through a fuzzy logic control algorithm and establish a nonlinear mapping relationship between film thickness and voltage according to the radius of curvature of the curtain wall; The spectral feature decomposition unit is configured to use a non-negative matrix decomposition algorithm to separate the energy distribution of visible light and infrared bands, and to construct a frequency domain filter group in combination with the intrinsic transmittance of the glass material; The curved surface incident angle compensation unit is configured to establish a light incident angle compensation model based on a spherical coordinate system and use a particle swarm optimization algorithm to calculate the normal vector offset of each node; The multi-source data fusion unit is configured to utilize the Bayesian network to fuse the ambient temperature, building azimuth and user activity data to construct a dynamic priority weight allocation mechanism.

[0009] Preferably, the light intensity distribution prediction model adopts a transfer learning framework, takes the standard atmospheric radiation model parameters as initial weights, and performs online fine-tuning through real-time collected polarized light data. The model formula can be expressed as: ,in are the fine-tuned model parameters, is the initial weight, is the learning rate, is the polarization data pair collected in real time. , They are , The mean of is the number of data points.

[0010] Preferably, the fuzzy logic control algorithm is provided with 9-dimensional input variables, including ultraviolet intensity, visible light band proportion, infrared radiation flux, surface temperature gradient, user density distribution, time factor, seasonal parameter, visibility index and building function type.

[0011] Preferably, the non-negative matrix decomposition algorithm sets constraints as follows: the decomposed basis vectors must satisfy the International Commission on Illumination standard chromaticity space coordinate range, and the reconstruction error does not exceed the spectrometer measurement accuracy threshold.

[0012] Preferably, the incident angle compensation model introduces a surface deformation compensation factor , the curvature change of the glass panel under wind load is calculated by finite element analysis , establish the angle correction The quadratic surface equation is: ,in , , are the coefficients obtained through experimental fitting.

[0013] Preferably, the multi-source data fusion unit integrates a weather forecast API interface, uses a long short-term memory network to predict the sun's position trajectory within the next 2 hours, and generates a pre-adjustment instruction queue, and its prediction model formula is: , ,in is the hidden state, is the hidden state at the previous moment, The input data includes ambient temperature, building azimuth, user activity data and weather forecast data. , is the weight matrix, , is the bias vector, The output is the predicted value of the sun's position trajectory, based on which a pre-adjustment instruction queue is generated.

[0014] Preferably, the dynamic electrochromic film layer unit comprises a tungsten oxide nanowire array structure, uses pulse width modulation technology to control the ion migration rate, and sets the minimum response time window to 120 ms.

[0015] Preferably, the multispectral sensor array adopts a hexagonal close-packed layout, and each sensor node includes four photodiodes with different cut-off wavelengths, covering the electromagnetic spectrum range of 380nm to 2500nm. The cut-off wavelengths of the four photodiodes are respectively , , , , its photocurrent , , , The wavelength of incident light , light intensity The relationship satisfies: , ,in For the The spectral response function of a photodiode, when i=1, , represents the minimum wavelength that the multi-spectral sensor array responds to.

[0016] Preferably, the system is provided with an abnormal spectrum detection module, and when a sudden change in the radiation intensity in a specific wavelength interval is monitored, the Grubbs criterion is automatically triggered to perform outlier analysis and start a backup control strategy.

[0017] Preferably, the system also includes a central control unit, which adopts a distributed edge computing architecture, and each partition controller is equipped with a redundant CAN bus interface to achieve millisecond-level state synchronization and fault isolation functions.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The spherical glass curtain wall adaptive light control system based on spectral regulation proposed in the present invention has many significant beneficial effects. In terms of light environment optimization, the multi-spectral sensor array unit collects the wavelength distribution matrix of the incident light in real time, constructs a light intensity distribution prediction model through a convolutional neural network, and establishes a three-dimensional radiation transfer equation in combination with historical climate data, which can accurately predict the light intensity distribution. This allows the system to adjust the transmittance of the dynamic electrochromic film layer unit in advance according to the trend of light changes. For example, when the sunlight gradually becomes stronger, the transmittance is reduced in advance to avoid excessive brightness or overheating in the room, effectively reduce the stimulation of glare to the human eye, create a comfortable and stable visual environment for indoor personnel, and improve the comfort of work and life.

[0019] The dynamic electrochromic film unit adopts fuzzy logic control algorithm, comprehensively considering 9-dimensional input variables such as ultraviolet intensity, visible light band proportion, infrared radiation flux, surface temperature gradient, user density distribution, time factor, seasonal parameters, visibility index and building function type to generate a transmittance adjustment matrix. This multi-factor fusion control method enables the system to intelligently and flexibly adjust the transmittance of the glass curtain wall according to different scenarios and needs. For example, in places such as conference rooms where there are dense crowds and specific requirements for lighting, when the meeting starts, the system automatically adjusts the transmittance according to the density of people and the type of building function to provide appropriate light intensity; in high temperatures in summer, combined with infrared radiation flux and seasonal parameters, the transmittance is reduced to block heat from entering the room and improve indoor thermal comfort.

[0020] The spectral feature decomposition unit uses a non-negative matrix decomposition algorithm to separate the energy distribution of visible light and infrared bands, and constructs a frequency domain filter group in combination with the intrinsic transmittance of the glass material. By setting the decomposed basis vector to meet the standard chromaticity space coordinate range of the International Commission on Illumination, and the reconstruction error does not exceed the measurement accuracy threshold of the spectrometer, accurate control of the spectrum is ensured. This can not only effectively adjust the indoor light intensity, but also optimize the quality of light and reduce the entry of harmful light, such as filtering excessive ultraviolet rays to protect the health of indoor personnel; at the same time, it can reasonably control infrared radiation, reduce indoor heat load, and further improve the quality of the indoor environment.

[0021] The curved surface incident angle compensation unit establishes a light incident angle compensation model based on the spherical coordinate system, introduces a curved surface deformation compensation factor, calculates the curvature change of the glass panel under wind load through finite element analysis, and establishes a quadratic surface equation for the angle correction amount. This design fully considers the curved surface characteristics of the spherical glass curtain wall and the influence of external factors on the glass panel, accurately compensates for the change in light incident angle caused by the curved surface shape and deformation, and improves the accuracy of light control. Even in severe weather conditions such as strong winds, when the glass panel is deformed, the system can still ensure the uniformity and stability of indoor lighting without being disturbed by changes in the external environment.

[0022] The multi-source data fusion unit uses the Bayesian network to fuse multi-source information such as ambient temperature, building azimuth, and user activity data, and constructs a dynamic priority weight allocation mechanism to make light control decisions more scientific and reasonable. At the same time, the weather forecast API interface is integrated, and the long short-term memory network is used to predict the trajectory of the sun's position in the next 2 hours to generate a pre-adjustment instruction queue. The system can adjust the light transmittance of the glass curtain wall in advance according to changes in the sun's position to achieve active intelligent light control. For example, in winter when the sunshine time is short, the system will increase the light transmittance in advance according to the forecast, make full use of sunlight for indoor lighting and heating, and reduce the energy consumption of artificial lighting and heating; when the weather changes suddenly, the light control strategy can be adjusted in time to maintain the stability of the indoor light environment.

[0023] In terms of system reliability, an abnormal spectrum detection module is set up. When a sudden change in radiation intensity in a specific wavelength range is detected, the Grubbs criterion is automatically triggered for outlier analysis and the backup control strategy is started. This design enhances the system's ability to respond to abnormal lighting conditions and ensures the stability of the indoor light environment in emergencies. The central control unit adopts a distributed edge computing architecture, and each partition controller is equipped with a redundant CAN bus interface to achieve millisecond-level state synchronization and fault isolation functions. This architectural design greatly improves the overall reliability of the system. Even if some components fail, the system can still maintain stable operation, ensure the normal realization of the light control function, reduce maintenance costs, and improve the safety and reliability of the building. The multispectral sensor array adopts a hexagonal close-packed layout. Each sensor node contains 4 photodiodes with different cutoff wavelengths, covering the electromagnetic spectrum range of 380nm to 2500nm, which improves the comprehensiveness and accuracy of light data collection and provides reliable data support for the precise control of the system. The distributed edge computing architecture makes data processing more efficient, reduces data transmission delays, improves the response speed of the system, and can make timely adjustments according to light changes, further improving the performance and practicality of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a working principle diagram of the spherical glass curtain wall adaptive light control system based on spectrum regulation according to the present invention; Figure 2 Flowchart for fine-tuning the light intensity distribution prediction model; Figure 3 It is a step diagram of the non-negative matrix factorization algorithm; Figure 4 This is the workflow diagram of the surface incident angle compensation model. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] See also Figure 1-4 The present invention provides a spherical glass curtain wall adaptive light control system based on spectrum regulation, which aims to realize intelligent light control of spherical glass curtain walls to meet the lighting and energy-saving requirements of buildings in different environments and usage requirements. The system mainly consists of the following key units working together: The multispectral sensor array unit is responsible for collecting the wavelength distribution matrix of the incident light in real time. A light intensity distribution prediction model is constructed through a convolutional neural network, which can effectively predict the light intensity distribution. At the same time, a three-dimensional radiation transfer equation is established in combination with historical climate data to provide basic data support for subsequent light control analysis.

[0027] The dynamic electrochromic film unit generates a transmittance adjustment matrix with the help of fuzzy logic control algorithm. The nonlinear mapping relationship between film thickness and voltage is established according to the curvature radius of the curtain wall, so as to achieve precise adjustment of the film transmittance to meet different lighting requirements.

[0028] The spectral feature decomposition unit uses a non-negative matrix decomposition algorithm to separate the energy distribution of visible light and infrared bands. A frequency domain filter group is constructed in combination with the intrinsic transmittance of the glass material to further optimize the spectral control effect.

[0029] The curved surface incident angle compensation unit establishes a light incident angle compensation model based on the spherical coordinate system, and uses the particle swarm optimization algorithm to calculate the normal vector offset of each node to compensate for the change in light incident angle caused by the curved surface shape of the glass curtain wall, thereby improving the accuracy of light control.

[0030] The multi-source data fusion unit uses the Bayesian network to fuse ambient temperature, building azimuth and user activity data to build a dynamic priority weight allocation mechanism. The integration of multi-source information makes the light control decision more scientific and reasonable to meet the needs of different scenarios.

[0031] The implementation of the present invention is further described below in conjunction with Examples 1 to 6.

[0032] Embodiment 1: This embodiment focuses on the construction and optimization of the light intensity distribution prediction model in the multi-spectral sensor array unit. Its unique role is to improve the accuracy and real-time performance of the light intensity distribution prediction, so that the system can respond to light changes more accurately.

[0033] When constructing the light intensity distribution prediction model, the transfer learning framework is used. The parameters of the standard atmospheric radiation model are used as the initial weights. This is because the standard atmospheric radiation model has certain scientificity and universality in describing the influence of the atmosphere on radiation transmission, and can provide a more reasonable initial state for the model.

[0034] Online fine-tuning is performed through real-time collected polarized light data. In actual application scenarios, the environment in which the spherical glass curtain wall is located is complex and changeable. Polarized light data can reflect the changes in various characteristics of light during propagation, such as light scattering and reflection. Real-time collection of this data and use of it to perform online fine-tuning of the model can make the model better adapt to the dynamic changes of the actual environment.

[0035] The model formula is: .in, Represents the fine-tuned model parameters, which are the results of the model adjustment combined with real-time data, reflecting more accurate light intensity distribution prediction parameters in the current environment; is the initial weight, provided by the standard atmospheric radiation model; As the learning rate, it controls the speed at which the model learns new data during fine-tuning. If the learning rate is too large, the model may overfit the current real-time data and ignore the prior knowledge contained in the initial weights; if the learning rate is too small, the model will converge very slowly and will not be able to respond to environmental changes in a timely manner. In practical applications, the appropriate learning rate value is determined through multiple experiments and optimizations.

[0036] These are pairs of polarized light data collected in real time. Each pair of data contains two related characteristic values ​​of polarized light at a certain moment. These data pairs are the basis for fine-tuning the model. , They are , , which are used to calculate the deviation of a data pair from the mean to measure the degree of variation in the data. The number of data points reflects the amount of data involved in model fine-tuning. The larger the amount of data, the higher the accuracy of model fine-tuning, but the amount of calculation will also increase accordingly. In actual system operation, the number of collected data points should be reasonably determined based on hardware performance and real-time requirements.

[0037] For example, in a spherical building, a multispectral sensor array collects polarized light data in real time. Assume that 100 data pairs are collected in a certain period of time, and the learning rate After optimization, it is set to 0.01. Substitute these data into the model formula for calculation, and continuously update the model parameters, so that the light intensity distribution prediction model can more accurately predict the light intensity distribution in the current building environment, providing reliable data support for subsequent light control operations.

[0038] Embodiment 2: This embodiment mainly revolves around the dynamic electrochromic film layer unit, and explains the specific application of the fuzzy logic control algorithm and the special structure and control technology of the film layer, which is used to achieve intelligent and precise adjustment of the transmittance of the dynamic electrochromic film layer.

[0039] The fuzzy logic control algorithm plays a key role in this system. It has 9-dimensional input variables, namely, ultraviolet intensity, visible light band proportion, infrared radiation flux, surface temperature gradient, user density distribution, time factor, seasonal parameter, visibility index and building function type. These input variables cover information on environmental lighting, temperature, human activities and building properties, and can fully reflect various factors that affect the light transmission requirements of glass curtain walls.

[0040] The intensity of ultraviolet light is directly related to the health of indoor personnel. Excessive ultraviolet light intensity may cause damage to human skin and eyes. Therefore, it is necessary to adjust the transmittance of the film layer according to the ultraviolet light intensity to reduce the ultraviolet light entering the room. The proportion of the visible light band reflects the relative content of visible light in the current light. Different proportions of visible light will affect the brightness and visual effect of the room. Adjusting the transmittance of the film layer according to the proportion can create a suitable indoor light environment. Infrared radiation flux is closely related to indoor thermal comfort. Excessive infrared radiation will cause the indoor temperature to rise. By adjusting the transmittance of the film layer, the amount of infrared radiation entering the room can be controlled to achieve indoor temperature regulation.

[0041] The surface temperature gradient reflects the change of the surface temperature of the glass curtain wall, which is related to the thermal stress and heat transfer process of the glass. When the surface temperature gradient is large, it may affect the structural stability of the glass and also affect the heat exchange efficiency between indoor and outdoor. By monitoring the surface temperature gradient and using it as one of the input variables of the fuzzy logic control algorithm, the system can adjust the light transmittance of the film layer according to the change of the temperature gradient to optimize the thermal performance of the glass curtain wall.

[0042] The user density distribution reflects the activities of indoor personnel. Different areas have different user densities, and their needs for lighting and thermal comfort are also different. For example, densely populated areas may require more sufficient lighting and better heat dissipation, while areas with fewer people can appropriately reduce the transmittance to save energy. The time factor and seasonal parameters reflect the lighting characteristics and climatic conditions of different time periods and seasons. The light intensity and temperature vary greatly between day and night and in different seasons. Adjusting the light transmittance of the film layer according to these factors can make the system better adapt to changes in the natural environment.

[0043] The visibility index affects the visual effects indoors and outdoors. In the case of low visibility, it may be necessary to increase the light transmittance of the film layer to ensure sufficient natural light indoors and enhance the visibility between indoors and outdoors. Different types of building functions have different requirements for the light environment. For example, office buildings need to provide bright and uniform lighting to meet office needs; exhibition halls need to precisely control the intensity and angle of light according to the characteristics of exhibits and display requirements.

[0044] The dynamic electrochromic film layer unit contains a tungsten oxide nanowire array structure, and this special structure endows the film layer with excellent electrochromic performance. Tungsten oxide nanowires have a large specific surface area and good ion transport channels, which can improve the migration rate of ions in the film layer, thereby accelerating the color change response speed of the film layer.

[0045] The pulse width modulation technology is used to control the ion migration rate. The pulse width modulation technology controls the average voltage applied to the film layer by adjusting the duty cycle of the pulse signal, and then controls the ion migration rate. When it is necessary to increase the light transmittance of the film layer, the duty cycle of the pulse signal is increased, so that the average voltage applied to the film layer rises, promoting the rapid migration of ions, realizing the film layer color becoming lighter and the light transmittance increasing; conversely, when the duty cycle is reduced and the average voltage is lowered, the ion migration rate slows down, the film layer color becomes darker, and the light transmittance decreases.

[0046] The minimum response time window is set to 120 ms, which ensures that the film layer can quickly respond to the light change requirements. For example, when the sunlight suddenly intensifies, the system detects the change of relevant input variables, and calculates that it is necessary to reduce the light transmittance of the film layer through the fuzzy logic control algorithm. At this time, based on the pulse width modulation technology, the system adjusts the voltage applied to the tungsten oxide nanowire array structure film layer within 120 ms, enabling the rapid migration of ions, realizing the rapid reduction of the film layer light transmittance, effectively blocking excessive light from entering the room, and providing a comfortable light environment for indoor personnel.

[0047] Example 3: This example focuses on introducing the non-negative matrix factorization algorithm and its constraint conditions adopted in the spectral feature decomposition unit, and its function is to accurately separate the energy distributions of the visible light and infrared bands and optimize the spectral regulation effect.

[0048] The spectral feature decomposition unit adopts the non-negative matrix factorization algorithm. The core idea of this algorithm is to decompose a non-negative matrix into the product of two or more non-negative matrices, which is used in this system to separate the energy distributions of the visible light and infrared bands.

[0049] In order to ensure the accuracy and effectiveness of the decomposition results, specific constraints are set. The decomposed basis vectors must meet the coordinate range of the International Commission on Illumination standard color space. The International Commission on Illumination standard color space is a widely recognized spatial coordinate system for describing color. Within this space range, it can ensure that the energy distribution of the separated visible light band is within the color range that conforms to the laws of human visual perception. For example, in this standard color space, different coordinate values ​​correspond to different color perceptions. Ensuring that the decomposed visible light basis vectors are within this range can make the visible light passing through the glass curtain wall present a natural and comfortable color, avoid color distortion and other problems, and provide a good visual experience for indoor people.

[0050] The reconstruction error does not exceed the measurement accuracy threshold of the spectrometer. The reconstruction error refers to the error generated when reconstructing the original matrix through the decomposed matrix. This error reflects the degree of restoration of the original data by the decomposition algorithm. The measurement accuracy threshold of the spectrometer is the accuracy standard that the spectrometer can achieve during the measurement process. Requiring that the reconstruction error does not exceed the measurement accuracy threshold of the spectrometer means that the non-negative matrix decomposition algorithm can accurately restore the original spectral data within the measurement accuracy range of the spectrometer, ensuring the accuracy of the energy distribution of the separated visible light and infrared bands. If the reconstruction error is too large, it means that the decomposition algorithm may have lost some important spectral information, resulting in deviations in the regulation of the spectrum and failure to achieve the expected light control effect.

[0051] In practical applications, the spectral data matrix containing visible light and infrared bands is first obtained and used as the input of the non-negative matrix decomposition algorithm. The algorithm performs iterative calculations while satisfying the above constraints, and continuously adjusts the decomposed matrix until the reconstruction error meets the requirements. After multiple iterative calculations, the energy distribution of visible light and infrared bands is successfully separated, providing an accurate data basis for the subsequent construction of a frequency domain filter group combined with the intrinsic transmittance of the glass material, thereby achieving precise control of the spectrum.

[0052] Embodiment 4: This embodiment describes in detail the incident angle compensation model of the curved surface incident angle compensation unit and the related calculation method, which is used to improve the compensation accuracy of the light incident angle change caused by the curved surface shape of the glass curtain wall and improve the accuracy of the light control system.

[0053] The light incident angle compensation model is established based on the spherical coordinate system. Since the spherical glass curtain wall has a curved surface, when the light is incident on the curtain wall, the incident angle will change due to different positions. The spherical coordinate system can accurately describe the position and direction of each point on the sphere, providing a suitable spatial framework for establishing the incident angle compensation model.

[0054] Introducing surface deformation compensation factors into the model This is because the glass panel will be subjected to various external forces during actual use, such as wind load, temperature change, etc. These external forces will cause the glass panel to deform, thus affecting the incident angle of light. Considering the surface deformation compensation factor can more accurately compensate for the change in incident angle caused by the deformation of the glass panel.

[0055] Calculate the curvature change of glass panels under wind load through finite element analysis Finite element analysis is a powerful engineering analysis method. It discretizes the glass panel into multiple finite-sized units. By performing mechanical analysis on each unit, the stress and strain distribution of the glass panel under wind load is solved, and the curvature change is obtained. For example, during the design stage of the glass curtain wall of a spherical building, the glass panel is modeled using finite element analysis software, and the parameters such as the size and direction of the wind load are input. After calculation, the curvature change of the glass panel under specific wind load conditions is obtained.

[0056] Establishing Angle Correction The quadratic surface equation is: .in, , , are coefficients obtained through experimental fitting. These coefficients are obtained by conducting a large number of experiments in the laboratory or actual building environment, measuring the angle correction under different curvature changes, and then using mathematical fitting methods. They reflect the quantitative relationship between the curvature change and the angle correction.

[0057] For example, on an experimental platform, wind loads of different magnitudes are simulated on a spherical glass panel, and the corresponding curvature changes and angle corrections are measured. By fitting and analyzing multiple sets of experimental data, the coefficients , , In actual applications, when the system detects that the glass panel is subjected to wind load, the curvature change is first calculated through finite element analysis. , and then substitute it into the quadratic surface equation of the angle correction, combined with the known surface deformation compensation factor and coefficient , , , calculate the angle correction Based on the calculated angle correction, the system can compensate for the incident angle of light and adjust the subsequent light control strategy to ensure that the incident angles of light at different positions can be accurately processed, thereby improving the overall accuracy of the light control system.

[0058] Embodiment 5: This embodiment mainly describes the working process of the multi-source data fusion unit, including data fusion method, sun position trajectory prediction and generation of pre-adjustment instruction queue, which is used to integrate multi-source information, predict the change of sun position in advance and realize more intelligent light control.

[0059] The multi-source data fusion unit uses the Bayesian network to fuse the ambient temperature, building azimuth and user activity data. The Bayesian network is a graphical model based on probabilistic reasoning. It can handle uncertainty information well, fuse data from different sources, and calculate the probability distribution under different conditions based on the correlation between the data.

[0060] Ambient temperature affects the thermal comfort of indoor occupants and is also related to the heat transfer process of glass curtain walls. Higher ambient temperatures may require reducing the light transmittance of glass curtain walls to reduce solar radiation entering the room, thereby reducing indoor temperature. The building azimuth determines the angle at which sunlight hits the glass curtain wall at different times. Different azimuths lead to differences in light intensity and distribution on the glass curtain wall. For example, south-facing glass curtain walls receive more sunlight during the day, while east- or west-facing glass curtain walls receive strong direct sunlight in the morning or evening. User activity data reflects the behavioral habits and needs of indoor occupants, such as the frequency of activities in different areas and preferences for light. Fusion of these data through a Bayesian network can provide a more comprehensive understanding of the building's environmental status and user needs, providing a more accurate basis for light control decisions.

[0061] This unit integrates the weather forecast API interface to obtain weather forecast data, such as weather conditions, cloud thickness and other information. These weather data will affect the intensity and angle of solar radiation reaching the ground. For example, on cloudy days, the clouds are thicker, and the solar radiation will be reflected and scattered by the clouds, and the light intensity reaching the glass curtain wall will be weakened; on sunny days, the solar radiation intensity is stronger. Combined with weather forecast data, future changes in the sun's position and light intensity can be predicted more accurately.

[0062] The long short-term memory network is used to predict the sun's position trajectory within the next 2 hours. The prediction model formula is: , .in, is the hidden state, It is the hidden state of the previous moment, which saves the intermediate calculation results of the model at different time steps and contains the information of the previous input data, which is used to capture the long-term dependencies in the time series. The input data include ambient temperature, building orientation, user activity data and weather forecast data, which provide the model with the information needed for prediction.

[0063] , The weight matrix determines the importance of input data and hidden states in the model calculation. Different weight matrix settings will affect the model's ability to learn and utilize different data features. The weight matrix is ​​optimized and adjusted through a large amount of training data, so that the model can better fit the changing pattern of the sun's position trajectory. , They are bias vectors, which add a fixed offset to the model’s calculations, helping the model to better converge and learn complex functional relationships.

[0064] The predicted value of the sun's position trajectory obtained based on the long short-term memory network , generating a pre-adjustment instruction queue. For example, when it is predicted that the sun will gradually rise and the light intensity will increase in the future, the system will generate an instruction to reduce the transmittance of the dynamic electrochromic film layer in advance, and add it to the pre-adjustment instruction queue in chronological order. The light control system will adjust the transmittance of the glass curtain wall in advance according to the instructions in the instruction queue to avoid discomfort to personnel caused by sudden changes in indoor light intensity, while also making more efficient use of energy and realizing intelligent light control.

[0065] Embodiment 6: This embodiment covers the layout and working principle of the multi-spectral sensor array, the operating mechanism of the abnormal spectrum detection module and the architectural characteristics of the central control unit. Its role is to ensure the stable and reliable operation of the system, improve the accuracy of light data collection and the ability to deal with abnormal situations.

[0066] The multispectral sensor array adopts a hexagonal close-packed layout, which has a high space utilization rate and can arrange more sensor nodes in a limited area to improve the detection accuracy of light. Each sensor node contains 4 photodiodes with different cut-off wavelengths, covering the electromagnetic spectrum range of 380nm to 2500nm. The cut-off wavelengths of the 4 photodiodes are , , , .

[0067] Its photocurrent , , , The wavelength of incident light , light intensity The relationship satisfies: , .in, For the The spectral response function of a photodiode, when i=1, , which represents the minimum wavelength that the multi-spectral sensor array responds to. This means that photodiodes with different cut-off wavelengths integrate the incident light within a specific wavelength range according to their respective spectral response functions, convert the optical signal into an electrical signal, and thus obtain light intensity information in different bands.

[0068] For example, when a beam of light containing multiple wavelengths is irradiated onto a sensor node, the cutoff wavelength is The photodiode will detect wavelengths (Assuming the starting wavelength is )arrive The light in the range is responded to, and the corresponding photocurrent is generated by integrating the spectral response function and the incident light intensity. By analyzing the photocurrent generated by the four photodiodes, the energy distribution of the incident light in different bands can be obtained, providing accurate data support for subsequent light intensity distribution prediction and spectral regulation.

[0069] The system sets up an abnormal spectrum detection module. When a sudden change in the radiation intensity in a specific wavelength range is detected, the Grubbs criterion is automatically triggered for outlier analysis. The Grubbs criterion is a commonly used method to determine whether data is an outlier. It determines the outlier by calculating the statistical characteristics of the data. In this system, when the multi-spectral sensor array detects that the radiation intensity change in a certain wavelength range exceeds the normal range, the abnormal spectrum detection module activates the Grubbs criterion.

[0070] For example, under normal circumstances, the radiation intensity in a specific wavelength interval should vary within a certain fluctuation range. If at a certain moment, the radiation intensity in this interval suddenly rises or falls sharply, the system detects this mutation, sorts out the radiation intensity data in this interval, and calculates its mean and standard deviation. According to the Grubbs criterion, set an appropriate significance level (such as 0.05) and calculate the Grubbs statistic. If the statistic exceeds the corresponding critical value, the data point is determined to be an outlier, that is, the change in radiation intensity in this wavelength interval is abnormal. Once it is determined to be abnormal, the system immediately starts the backup control strategy, such as switching to a preset fixed transmittance mode or adopting other safe light control strategies to ensure the stability and safety of the indoor light environment and avoid adverse effects on indoor personnel and equipment due to abnormal lighting.

[0071] The central control unit adopts a distributed edge computing architecture, which distributes computing tasks to various edge nodes for processing, reducing data transmission delays and improving the system's response speed. Each partition controller is equipped with a redundant CAN bus interface. The CAN bus is a serial communication bus widely used in the field of industrial control, with the advantages of high reliability and strong real-time performance. The setting of redundant CAN bus interfaces further improves the reliability of the system. When one bus fails, the other bus can immediately take over the work, realizing millisecond-level state synchronization and fault isolation functions.

[0072] In actual operation, each partition controller is responsible for collecting data from devices such as multi-spectral sensor arrays and dynamic electrochromic film units in the area, and performing preliminary processing locally. For example, the collected light data is preprocessed to extract key information. Then, the processed data is transmitted to the central control unit via the CAN bus. At the same time, the central control unit can also send control instructions to each partition controller via the CAN bus to achieve unified management and control of the entire system. At a certain moment, if a CAN bus between a partition controller and the central control unit fails, the other redundant CAN bus will detect the fault within milliseconds and automatically take over the data transmission work, ensuring uninterrupted communication between the various parts of the system, maintaining the stable operation of the system, and ensuring that the light control system can continuously and reliably provide good light environment control services for the building.

[0073] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0074] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A spherical glass curtain wall adaptive light control system based on spectrum regulation, characterized in that: include: A multispectral sensor array unit is configured to collect the wavelength distribution matrix of incident light in real time, build a light intensity distribution prediction model based on a convolutional neural network, and establish a three-dimensional radiation transfer equation in combination with historical climate data; The dynamic electrochromic film layer unit is configured to generate a transmittance adjustment matrix through a fuzzy logic control algorithm and establish a nonlinear mapping relationship between film thickness and voltage according to the radius of curvature of the curtain wall; The spectral feature decomposition unit is configured to use a non-negative matrix decomposition algorithm to separate the energy distribution of visible light and infrared bands, and to construct a frequency domain filter group in combination with the intrinsic transmittance of the glass material; The curved surface incident angle compensation unit is configured to establish a light incident angle compensation model based on a spherical coordinate system and use a particle swarm optimization algorithm to calculate the normal vector offset of each node; The multi-source data fusion unit is configured to utilize the Bayesian network to fuse the ambient temperature, building azimuth and user activity data to construct a dynamic priority weight allocation mechanism.

2. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The light intensity distribution prediction model adopts a transfer learning framework, takes the standard atmospheric radiation model parameters as the initial weights, and performs online fine-tuning through real-time collected polarized light data. The model formula can be expressed as: ,in are the fine-tuned model parameters, is the initial weight, is the learning rate, is the polarization data pair collected in real time. , They are , The mean of is the number of data points.

3. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The fuzzy logic control algorithm is provided with 9-dimensional input variables, including ultraviolet intensity, visible light band proportion, infrared radiation flux, surface temperature gradient, user density distribution, time factor, seasonal parameter, visibility index and building function type.

4. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The constraints set by the non-negative matrix decomposition algorithm are: the decomposed basis vectors must meet the coordinate range of the International Commission on Illumination standard chromaticity space, and the reconstruction error must not exceed the measurement accuracy threshold of the spectrometer.

5. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The incident angle compensation model introduces a surface deformation compensation factor , the curvature change of the glass panel under wind load is calculated by finite element analysis , establish the angle correction The quadratic surface equation is: ,in , , are the coefficients obtained through experimental fitting.

6. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The multi-source data fusion unit integrates the weather forecast API interface, uses the long short-term memory network to predict the sun's position trajectory within the next 2 hours, and generates a pre-adjustment instruction queue. The prediction model formula is: , ,in is the hidden state, is the hidden state at the previous moment, The input data includes ambient temperature, building azimuth, user activity data and weather forecast data. , is the weight matrix, , is the bias vector, The output is the predicted value of the sun's position trajectory, based on which a pre-adjustment instruction queue is generated.

7. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The dynamic electrochromic film layer unit includes a tungsten oxide nanowire array structure, uses pulse width modulation technology to control the ion migration rate, and sets the minimum response time window to 120ms.

8. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The multi-spectral sensor array adopts a hexagonal close-packed layout. Each sensor node contains four photodiodes with different cut-off wavelengths, covering the electromagnetic spectrum range of 380nm to 2500nm. The cut-off wavelengths of the four photodiodes are , , , , its photocurrent , , , The wavelength of incident light , light intensity The relationship satisfies: , ,in For the The spectral response function of a photodiode, when i=1, , represents the minimum wavelength that the multi-spectral sensor array responds to.

9. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The system is provided with an abnormal spectrum detection module, and when a sudden change in the radiation intensity in a specific wavelength range is detected, the Grubbs criterion is automatically triggered to perform outlier analysis and start a backup control strategy.

10. The spherical glass curtain wall adaptive light control system according to claim 1, characterized in that: The system also includes a central control unit, which adopts a distributed edge computing architecture. Each partition controller is equipped with a redundant CAN bus interface to achieve millisecond-level state synchronization and fault isolation functions.

Images