An adaptive light control system for spherical glass curtain walls based on spectral regulation

Abstract

The present invention relates to the field of architectural engineering technology, and discloses an adaptive light control system for a spherical glass curtain wall based on spectral regulation. The system is composed of multiple 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 transmittance based on multiple factors; the spectral feature decomposition unit optimizes spectral regulation; the curved incident angle compensation unit accurately controls light on the curved surface of the curtain wall; and the multi-source data fusion unit makes decisions based on multi-source information and predicts the position of the sun. The system also has anomaly detection and redundant design to ensure stable operation. The system realizes intelligent and precise control of the lighting of the spherical glass curtain wall, effectively optimizes the indoor light environment, and improves energy efficiency.

Description

Technical Field

[0001] The present 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 booming construction industry, glass curtain walls, as a modern building envelope, are widely used in various buildings due to their unique aesthetics and excellent lighting performance. Spherical glass curtain walls, with their novel and distinctive appearance, have become the preferred choice for many landmark buildings, adding a unique charm to the urban landscape. However, spherical glass curtain walls face many challenges in practical application.

[0003] From a lighting perspective, its unique spherical curved surface creates an extremely complex light-incident situation. Unlike traditional flat glass curtain walls, the normal direction of each point on the sphere constantly changes, and the angle of incidence of light varies significantly. This results in extremely uneven distribution of light across the curtain wall surface. In some areas, excessive light concentration creates intense glare, which not only severely impacts the visual comfort of occupants but can also lead to vision loss if exposed to such conditions for extended periods. Furthermore, excessive solar radiation entering the room rapidly raises the indoor temperature, increasing the load on the air conditioning system and resulting in significant energy consumption. In other areas, insufficient light may be sufficient for normal indoor use, necessitating the use of additional artificial lighting, 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 a uniform and effective shading effect when faced with the complex curves of a spherical curtain wall. Although electrochromic glass has the function of changing light transmittance, it is difficult to ensure the uniformity of color change on a spherical curtain wall due to the influence of curvature changes, and traditional control methods cannot accurately adapt to the complex and changeable lighting conditions of the spherical curtain wall, resulting in poor light control effects.

[0005] Furthermore, with increasing demands for indoor building environmental quality and a growing global emphasis on energy conservation and emission reduction, higher requirements are being placed on intelligent light-control systems for building glass curtain walls. Existing light-control systems often fail to fully account for seasonal and time-of-day variations in light intensity, nor do they adequately incorporate factors such as indoor human activity and building functions for comprehensive regulation. For example, during intense summer sunlight, excessive heat cannot be effectively and promptly blocked from entering the building, significantly increasing energy consumption for indoor air conditioning. In areas with high human activity, appropriate light intensity cannot be provided based on actual needs. Furthermore, most light-control systems lack comprehensive collection and in-depth analysis of environmental data, resulting in unscientific and inaccurate light-control decisions, making it impossible to achieve efficient energy utilization and optimize the 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, interference from reflected light from nearby buildings, etc., the system is unable to respond quickly and accurately, which can easily cause drastic fluctuations in the indoor light environment. Moreover, the coordination between the various 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 spherical glass curtain walls, achieve precise, intelligent, and efficient light control, and at the same time have 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 objectives, the present invention provides the following technical solution: an adaptive light control system for a spherical glass curtain wall based on spectral regulation, the system comprising:

[0009] 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 based on historical climate data;

[0010] A 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 curtain wall curvature radius;

[0011] 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 bank based on the intrinsic transmittance of the glass material;

[0012] 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;

[0013] The multi-source data fusion unit is configured to utilize a Bayesian network to fuse ambient temperature, building orientation, and user activity data to construct a dynamic priority weight allocation mechanism.

[0014] 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.

[0015] 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.

[0016] Preferably, the non-negative matrix decomposition algorithm sets the constraint conditions as follows: the decomposed basis vectors must meet the coordinate range of the International Commission on Illumination standard chromaticity space, and the reconstruction error does not exceed the measurement accuracy threshold of the spectrometer.

[0017] Preferably, the incident angle compensation model introduces a surface deformation compensation factor , calculate the curvature change of glass panels under wind load through finite element analysis , establish the angle correction The quadratic surface equation of : ,in 、 、 are the coefficients obtained through experimental fitting.

[0018] 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. The prediction model formula is: , ,in is hidden state, is the hidden state at the previous moment, Input data includes ambient temperature, building orientation, 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.

[0019] 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 120ms.

[0020] Preferably, the multispectral sensor array adopts a hexagonal close-packed layout, and 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 multispectral sensor array responds to.

[0021] Preferably, the system is provided with an abnormal spectrum detection module, which automatically triggers the Grubbs criterion to perform outlier analysis and start a backup control strategy when a sudden change in radiation intensity in a specific wavelength range is detected.

[0022] 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.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 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 reducing the stimulation of glare to the human eye, creating a comfortable and stable visual environment for indoor people, and improving the comfort of work and life.

[0025] The dynamic electrochromic film unit utilizes a fuzzy logic control algorithm, comprehensively considering nine input variables: UV 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 crowded places such as conference rooms with specific lighting requirements, when a meeting begins, the system automatically adjusts the transmittance based on the occupant density and building function type to provide appropriate light intensity. During high summer temperatures, the system combines infrared radiation flux and seasonal parameters to reduce transmittance, blocking heat from entering the room and improving indoor thermal comfort.

[0026] The spectral feature decomposition unit utilizes a non-negative matrix factorization algorithm to separate the energy distribution of the visible and infrared bands, and constructs a frequency domain filter bank based on the intrinsic transmittance of the glass material. By setting the decomposed basis vectors to meet the coordinate range of the International Commission on Illumination's standard chromaticity space, and ensuring that the reconstruction error does not exceed the spectrometer's measurement accuracy threshold, precise spectral control is ensured. This not only effectively regulates indoor light intensity but also optimizes light quality, reducing the entry of harmful light, such as filtering out excessive ultraviolet rays to protect the health of indoor occupants. At the same time, it rationally controls infrared radiation, reduces indoor heat load, and further improves indoor environmental quality.

[0027] The curved incident angle compensation unit establishes a light incident angle compensation model based on a spherical coordinate system, introduces a surface deformation compensation factor, and uses finite element analysis to calculate the curvature change of the glass panel under wind load, establishing a quadratic surface equation for the angle correction. This design fully considers the curved surface characteristics of the spherical glass curtain wall and the impact of external factors on the glass panel, accurately compensating for changes in light incident angle caused by the curved shape and deformation, thereby improving the accuracy of light control. Even in severe weather conditions such as strong winds, when the glass panel deforms, the system can still ensure the uniformity and stability of indoor lighting, unaffected by changes in the external environment.

[0028] The multi-source data fusion unit utilizes a Bayesian network to fuse multiple sources of information, including ambient temperature, building orientation, and user activity data, to establish a dynamic priority weighting mechanism, making lighting control decisions more scientific and rational. Furthermore, it integrates a weather forecast API and uses a long-short-term memory network to predict the sun's position over the next two hours, generating a pre-adjustment instruction queue. The system can proactively adjust the light transmittance of the glass curtain wall based on changes in the sun's position, enabling proactive intelligent light control. For example, during winter, when daylight hours are shorter, the system increases light transmittance based on forecasts, fully utilizing sunlight for indoor lighting and heating, reducing energy consumption for artificial lighting and heating. In the event of sudden weather changes, the system can also promptly adjust the light control strategy to maintain a stable indoor light environment.

[0029] To enhance system reliability, an abnormal spectrum detection module is implemented. When a sudden change in radiation intensity within a specific wavelength range is detected, it automatically triggers the Grubbs criterion for outlier analysis and activates a backup control strategy. This design enhances the system's ability to respond to abnormal lighting conditions and ensures a stable indoor lighting environment during emergencies. The central control unit utilizes a distributed edge computing architecture, with each zone controller equipped with redundant CAN bus interfaces, enabling millisecond-level state synchronization and fault isolation. This architecture significantly improves overall system reliability. Even if some components fail, the system remains stable, ensuring normal light control functions, reducing maintenance costs and enhancing building safety and reliability. The multispectral sensor array utilizes a dense hexagonal layout. Each sensor node contains four photodiodes with different cutoff wavelengths, covering the electromagnetic spectrum from 380nm to 2500nm. This improves the comprehensiveness and accuracy of light data collection, providing reliable data support for the system's precise control. This distributed edge computing architecture enables more efficient data processing, reduces data transmission latency, and improves system response speed, enabling timely adjustments to light changes, further enhancing system performance and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a working principle diagram of the spherical glass curtain wall adaptive light control system based on spectral regulation according to the present invention;

[0031] Figure 2 Flowchart for fine-tuning the light intensity distribution prediction model;

[0032] Figure 3 It is a step diagram of the non-negative matrix factorization algorithm;

[0033] Figure 4 This is the workflow diagram of the surface incident angle compensation model. DETAILED DESCRIPTION

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0035] See also Figure 1-4 This invention provides an adaptive light control system for spherical glass curtain walls based on spectral regulation. This system aims to achieve intelligent light control for spherical glass curtain walls, meeting 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:

[0036] 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 using a convolutional neural network, which effectively predicts this distribution. Furthermore, a three-dimensional radiation transfer equation is established in conjunction with historical climate data, providing basic data support for subsequent light control analysis.

[0037] The dynamic electrochromic film unit utilizes a fuzzy logic control algorithm to generate a transmittance adjustment matrix. A nonlinear mapping relationship between film thickness and voltage is established based on the curtain wall's curvature radius, enabling precise adjustment of the film's transmittance to accommodate varying lighting requirements.

[0038] The spectral feature decomposition unit uses a non-negative matrix factorization algorithm to separate the energy distribution of visible light and infrared bands. A frequency domain filter bank is constructed based on the intrinsic transmittance of the glass material to further optimize the spectral control effect.

[0039] 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 shape of the glass curtain wall, thereby improving the accuracy of light control.

[0040] The multi-source data fusion unit uses a Bayesian network to fuse ambient temperature, building orientation, and user activity data to create a dynamic priority weighting mechanism. This multi-source information makes lighting control decisions more scientific and rational, meeting the needs of different scenarios.

[0041] The implementation of the present invention will be further described below with reference to Examples 1 to 6.

[0042] Example 1:

[0043] This embodiment focuses on the construction and optimization of the light intensity distribution prediction model in the multispectral sensor array unit. Its unique role is to improve the accuracy and real-time performance of light intensity distribution prediction, so that the system can respond to lighting changes more accurately.

[0044] When constructing the light intensity distribution prediction model, a transfer learning framework was used. The parameters of the standard atmospheric radiation model were used as the initial weights. This is because the standard atmospheric radiation model is scientific and universal in describing the influence of the atmosphere on radiation transmission, and it can provide a relatively reasonable initial state for the model.

[0045] Online fine-tuning is implemented using real-time polarization data. In real-world applications, spherical glass curtain walls operate in complex and ever-changing environments. Polarization data can reveal various changes in light properties during propagation, such as scattering and reflection. Real-time collection of this data and its use in online model fine-tuning allows the model to better adapt to dynamic changes in the real environment.

[0046] The model formula is: .in, Represents the fine-tuned model parameters, which are the result of the model adjustment based on 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 how quickly the model learns new data during fine-tuning. If the learning rate is too high, the model may overfit to the current real-time data and ignore the prior knowledge contained in the initial weights. If the learning rate is too low, the model will converge very slowly and fail to respond to environmental changes in a timely manner. In practical applications, the appropriate learning rate is determined through multiple experiments and optimizations.

[0047] These are pairs of polarization data collected in real time. Each pair contains two related characteristic values of polarization light at a certain moment. These data pairs are the basis for fine-tuning the model. 、 They are 、 They are used to calculate the deviation of data pairs 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. A larger amount of data leads to higher accuracy, but also increases the computational effort. In actual system operation, the number of collected data points should be appropriately determined based on hardware performance and real-time requirements.

[0048] For example, in a spherical building, a multispectral sensor array collects polarization data in real time. Assume that within a certain period of time, 100 data pairs are collected and the learning rate After optimization, it is set to 0.01. Substituting these data into the model formula for calculation, the model parameters are continuously updated, 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.

[0049] Example 2:

[0050] This embodiment mainly focuses on the dynamic electrochromic film layer unit, and explains the specific application of the fuzzy logic control algorithm as well as the special structure and control technology of the film layer. Its function is to achieve intelligent and precise adjustment of the transmittance of the dynamic electrochromic film layer.

[0051] The fuzzy logic control algorithm plays a key role in this system. It employs nine input variables: UV 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 encompass information on ambient light, temperature, occupant activity, and building properties, comprehensively reflecting the various factors influencing the light transmission requirements of glass curtain walls.

[0052] UV intensity is directly related to the health of indoor users. Excessive UV intensity can cause damage to human skin and eyes. Therefore, the film transmittance needs to be adjusted according to the UV intensity to reduce UV penetration into the room. The visible light band ratio reflects the relative content of visible light in the current lighting. Different visible light ratios affect the brightness and visual effect of the room. Adjusting the film transmittance based on this ratio can create a suitable indoor lighting environment. Infrared radiation flux is closely related to indoor thermal comfort. Excessive infrared radiation can cause indoor temperatures to rise. By adjusting the film transmittance, the amount of infrared radiation entering the room can be controlled, thereby achieving indoor temperature regulation.

[0053] The surface temperature gradient reflects the temperature variations of a glass curtain wall's surface and is related to the glass's thermal stress and heat transfer processes. Large surface temperature gradients can affect the glass's structural stability and compromise the efficiency of heat exchange between indoor and outdoor spaces. By monitoring the surface temperature gradient and using it as an input variable in a fuzzy logic control algorithm, the system can adjust the film's transmittance based on the temperature gradient to optimize the glass curtain wall's thermal performance.

[0054] User density distribution reflects indoor occupancy patterns. Different areas have varying user densities, resulting in varying lighting and thermal comfort requirements. For example, densely populated areas may require more adequate lighting and improved cooling, while less crowded areas can appropriately reduce light transmittance to conserve energy. Time factors and seasonal parameters reflect the lighting characteristics and climatic conditions of different time periods and seasons. Light intensity and temperature vary significantly between daytime and nighttime, and between seasons. Adjusting the film's light transmittance based on these factors allows the system to better adapt to changes in the natural environment.

[0055] The visibility index affects both indoor and outdoor visual effects. In low visibility conditions, the film's light transmittance may need to be increased to ensure sufficient natural light indoors and enhance visibility. Different building types have different lighting requirements. For example, office buildings require bright, even lighting to meet office needs; exhibition halls require precise control of lighting intensity and angle based on the characteristics of the exhibits and display requirements.

[0056] The dynamic electrochromic film unit comprises a tungsten oxide nanowire array structure, a unique structure that endows the film with excellent electrochromic properties. Tungsten oxide nanowires have a large surface area and excellent ion transport pathways, which can increase the migration rate of ions in the film, thereby accelerating the film's color change response.

[0057] Pulse-width modulation technology is used to control the ion migration rate. Pulse-width modulation controls the average voltage applied to the film by adjusting the duty cycle of the pulse signal, thereby controlling the ion migration rate. To increase the film's transmittance, increasing the duty cycle of the pulse signal increases the average voltage applied to the film, promoting faster ion migration, resulting in a lighter film color and increased transmittance. Conversely, decreasing the duty cycle and average voltage slows the ion migration rate, darkens the film color, and reduces transmittance.

[0058] The minimum response time window is set to 120ms, ensuring the film can quickly respond to changes in lighting conditions. For example, when sunlight suddenly intensifies, the system detects changes in relevant input variables and, using a fuzzy logic control algorithm, calculates the need to reduce the film's transmittance. At this point, using pulse width modulation technology, the system adjusts the voltage applied to the tungsten oxide nanowire array structure within 120ms, enabling rapid ion migration and a rapid reduction in the film's transmittance. This effectively blocks excessive light from entering the room, providing a comfortable lighting environment for those inside.

[0059] Example 3:

[0060] This embodiment focuses on the non-negative matrix decomposition algorithm and its constraints used in the spectral feature decomposition unit, which is used to accurately separate the energy distribution of visible light and infrared bands and optimize the spectral control effect.

[0061] The spectral feature decomposition unit adopts the non-negative matrix decomposition algorithm. The core idea of this algorithm is to decompose a non-negative matrix into the product of two or more non-negative matrices. In this system, it is used to separate the energy distribution of visible light and infrared bands.

[0062] 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, it can ensure that the energy distribution of the separated visible light band is within a 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 appear natural and comfortable, avoid problems such as color distortion, and provide a good visual experience for indoor occupants.

[0063] 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 to which the decomposition algorithm can restore the original data. The spectrometer measurement accuracy threshold is the accuracy standard that the spectrometer can achieve during the measurement process. Requiring that the reconstruction error does not exceed the spectrometer measurement accuracy threshold 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 separated visible light and infrared band energy distribution. 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.

[0064] In practical applications, a spectral data matrix containing visible and infrared bands is first obtained and used as input for a non-negative matrix factorization algorithm. The algorithm iterates within the constraints outlined above, continuously adjusting the decomposed matrix until the reconstruction error meets the required standards. After multiple iterations, the energy distribution of the visible and infrared bands is successfully separated, providing an accurate data foundation for the subsequent construction of a frequency-domain filter bank based on the intrinsic transmittance of the glass material, thereby achieving precise spectral control.

[0065] Example 4:

[0066] This embodiment describes in detail the incident angle compensation model and related calculation method of the curved surface incident angle compensation unit, 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 enhance the accuracy of the light control system.

[0067] A light incident angle compensation model is established based on a spherical coordinate system. Due to the curved shape of a spherical glass curtain wall, the incident angle of light will vary depending on the location of the light. The spherical coordinate system accurately describes the position and direction of each point on the sphere, providing a suitable spatial framework for establishing the incident angle compensation model.

[0068] Introducing surface deformation compensation factors into the model This is because glass panels are subject to various external forces during actual use, such as wind loads and temperature changes. These forces can cause the glass panel to deform, which in turn affects the incident angle of light. Considering the surface deformation compensation factor can more accurately compensate for the changes in the incident angle caused by glass panel deformation.

[0069] Calculate the curvature change of glass panels under wind loads through finite element analysis Finite element analysis is a powerful engineering analysis method that discretizes glass panels 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 phase of the glass curtain wall of a spherical building, finite element analysis software was used to model the glass panel. Parameters such as the magnitude and direction of the wind load were input, and the curvature change of the glass panel under specific wind load conditions was calculated.

[0070] Establishing Angle Correction The quadratic surface equation of : .in, 、 、 These coefficients are obtained through experimental fitting. These coefficients are derived through extensive experiments in the laboratory or in actual building environments, measuring the angle corrections for varying curvature variations, and then applying mathematical fitting methods. They reflect the quantitative relationship between curvature variation and angle correction.

[0071] For example, on an experimental platform, we simulate wind loads of different sizes acting on a spherical glass panel and measure the corresponding curvature change and angle correction. By fitting and analyzing multiple sets of experimental data, we can determine the coefficients. 、 、 In actual application, 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 coefficients 、 、 , calculate the angle correction Based on the calculated angle correction, the system can compensate for the incident angle of light and adjust subsequent light control strategies 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.

[0072] Example 5:

[0073] This embodiment mainly describes the working process of the multi-source data fusion unit, including the data fusion method, the prediction of the sun's position trajectory, and the generation of the pre-adjustment instruction queue. Its function is to integrate multi-source information, predict the change of the sun's position in advance, and achieve more intelligent light control.

[0074] The multi-source data fusion unit uses a Bayesian network to fuse ambient temperature, building orientation, and user activity data. A Bayesian network is a graphical model based on probabilistic reasoning that effectively handles uncertainty. It fuses data from different sources and, based on the relationships between the data, calculates probability distributions under different conditions.

[0075] 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 transmittance of glass curtain walls to reduce solar radiation entering the room, thereby lowering the indoor temperature. The building's azimuth determines the angle at which sunlight strikes the glass curtain wall at different times of the day. Different azimuths result in differences in light intensity and distribution across 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 stronger direct sunlight in the morning or evening. User activity data reflects the behavioral habits and needs of indoor occupants, such as the frequency of activity in different areas and lighting preferences. Fusion of this data through a Bayesian network provides a more comprehensive understanding of the building's environmental conditions and user needs, providing a more accurate basis for light control decisions.

[0076] This unit integrates a weather forecast API to access weather forecast data, such as weather conditions and cloud thickness. This weather data affects the intensity and angle of solar radiation reaching the ground. For example, on cloudy days, thick clouds reflect and scatter solar radiation, reducing the intensity of light reaching the glass curtain wall. On sunny days, solar radiation is stronger. Combined with weather forecast data, future changes in the sun's position and intensity can be more accurately predicted.

[0077] The long short-term memory network is used to predict the sun's position trajectory in the next 2 hours. The prediction model formula is: , .in, is 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 includes ambient temperature, building orientation, user activity data and weather forecast data, which provide the model with the information needed for prediction.

[0078] 、 The weight matrix determines the importance of input data and hidden states in the model calculation. Different weight matrix settings affect the model's ability to learn and utilize different data features. By optimizing the weight matrix through a large amount of training data, 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 calculation, helping the model to better converge and learn complex functional relationships.

[0079] The predicted value of the sun's position trajectory obtained based on the long short-term memory network , generating a pre-adjustment command queue. For example, if it's predicted that the sun will gradually rise and light intensity will increase over the next period of time, the system will pre-generate a command to reduce the transmittance of the dynamic electrochromic film and add it to the pre-adjustment command queue in chronological order. Based on the commands in the command queue, the light control system will pre-adjust the transmittance of the glass curtain wall to prevent discomfort caused by sudden changes in indoor light intensity. This also enables more efficient energy use and intelligent light control.

[0080] Example 6:

[0081] This embodiment covers the layout and operating principle of the multispectral sensor array, the operating mechanism of the abnormal spectrum detection module, and the architectural features 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 improve the ability to respond to abnormal situations.

[0082] The multispectral sensor array adopts a hexagonal close-packed layout, which has high space utilization and can arrange more sensor nodes in a limited area, thereby improving the detection accuracy of light. Each sensor node contains four photodiodes with different cutoff wavelengths, covering the electromagnetic spectrum range of 380nm to 2500nm. The cutoff wavelengths of the four photodiodes are 、 、 、 .

[0083] 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 to which the multispectral sensor array responds. This means that photodiodes with different cutoff wavelengths integrate the incident light within a specific wavelength range based on their respective spectral response functions, converting the optical signal into an electrical signal to obtain light intensity information in different bands.

[0084] 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 in (Assuming the starting wavelength is )arrive The light within 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 photocurrents generated by the four photodiodes, the energy distribution of the incident light in different wavelength bands can be obtained, providing accurate data support for subsequent light intensity distribution prediction and spectral control.

[0085] The system incorporates an outlier spectrum detection module. When a sudden change in radiation intensity within a specific wavelength range is detected, it automatically triggers the Grubbs criterion for outlier analysis. The Grubbs criterion is a commonly used method for determining whether data is an outlier. It determines outliers by calculating the statistical characteristics of the data. In this system, the outlier spectrum detection module activates the Grubbs criterion when the multispectral sensor array detects a change in radiation intensity within a specific wavelength range that exceeds the normal range.

[0086] For example, under normal circumstances, the radiation intensity in a specific wavelength range 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, organizes 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 range is abnormal. Once it is determined to be abnormal, the system immediately activates the backup control strategy, such as switching to the 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.

[0087] The central control unit utilizes a distributed edge computing architecture, which distributes computing tasks to various edge nodes for processing, reducing data transmission latency and improving system response speed. Each partition controller is equipped with redundant CAN bus interfaces. CAN bus is a serial communication bus widely used in industrial control, offering advantages such as high reliability and strong real-time performance. The redundant CAN bus interfaces further enhance system reliability. If one bus fails, the other bus can immediately take over, achieving millisecond-level state synchronization and fault isolation.

[0088] In actual operation, each zone controller is responsible for collecting data from devices such as the multispectral sensor array and dynamic electrochromic film unit in its area and performing preliminary local processing. For example, the collected light data is preprocessed to extract key information. The processed data is then 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 zone controller via the CAN bus, achieving unified management and control of the entire system. At a certain moment, if a CAN bus between a zone controller and the central control unit fails, the redundant CAN bus will detect the failure within milliseconds and automatically take over data transmission. This ensures uninterrupted communication between all parts of the system, maintains stable operation, and ensures that the light control system can continuously and reliably provide excellent light environment control services for the building.

[0089] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0090] While 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 these embodiments without departing from the principles and spirit of the invention, and that the scope of the 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 based on historical climate data; A 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 curtain wall curvature radius; 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 bank based on 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 a Bayesian network to fuse ambient temperature, building orientation, 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 parameters of the standard atmospheric radiation model as the initial weights, and performs online fine-tuning through real-time collected polarized light data. The model formula is 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 equipped with nine 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 , calculate the curvature change of glass panels under wind load through finite element analysis , establish the angle correction The quadratic surface equation of : ,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 hidden state, is the hidden state at the previous moment, Input data includes ambient temperature, building orientation, 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 multispectral sensor array adopts a hexagonal close-packed layout. Each sensor node contains four photodiodes with different cutoff wavelengths, covering the electromagnetic spectrum range from 380nm to 2500nm. The cutoff 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 multispectral 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. 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