Color-changing glass window self-adaptive control system based on environmental perception
By combining multispectral light sensors and prediction models, changes in light intensity can be collected and predicted in real time, solving the hysteresis and oscillation problems of existing photochromic glass control systems, and achieving more efficient optical state switching and extended equipment life.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-03-31
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing photochromic glass control systems rely on simple light sensors, which cannot distinguish between direct and diffused light, resulting in control lag and oscillations, affecting comfort and accelerating equipment aging.
Using a multispectral light sensor, a solar position calculation module, and a sky condition analysis module, combined with a prediction model, multidimensional environmental data is collected in real time to predict future changes in light intensity, and the state of the color-changing glass window is driven to switch through an adaptive control method.
It achieves forward-looking control, reduces control lag and oscillation, improves visual comfort, and extends equipment life.
Smart Images
Figure CN121763729A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of adaptive control technology, and more specifically, to an adaptive control system for a color-changing glass window based on environmental perception. Background Technology
[0002] With the development of industrial control systems and intelligent technologies, smart windows such as electrochromic glass are playing an increasingly important role in the light and heat management and comfort regulation of industrial environments. Currently, the environmental sensing and control strategies of such systems mostly rely on a limited number of environmental sensors. A common approach is to use one or more light sensors installed inside and outside the building to monitor light intensity (illuminance) or total solar irradiance, and to preset fixed trigger thresholds. Based on these point-like, instantaneous measurement data, the control system drives the electrochromic glass to switch between transparent and tinted states through classic feedback control logic (such as PID control) or simple if-then rules.
[0003] However, existing photochromic glass control systems based on this type of environmental perception still have significant limitations:
[0004] 1. Problem of limited perception: Most existing technologies rely on simple light sensors, which can only measure the total illuminance at the "current" time, and cannot distinguish between direct light and diffused light, let alone perceive the dynamic changes of clouds in the sky;
[0005] 2. Control lag and oscillation issues: Because feedback control is based solely on the current lighting conditions, the system only begins to function after environmental changes occur. For example, a sudden drop in light caused by passing clouds causes the glass to become transparent; when the clouds move away and strong light reappears, the glass immediately becomes tinted again. This "constantly running around" control mode leads to severe lag and frequent state oscillations, affecting comfort and accelerating equipment aging.
[0006] Therefore, an adaptive control system for color-changing glass windows based on environmental perception is designed. Summary of the Invention
[0007] The purpose of this invention is to provide an adaptive control system for color-changing glass windows based on environmental perception, so as to solve the problems mentioned in the background art.
[0008] To achieve the above objectives, the present invention aims to provide an adaptive control system for photochromic windows based on environmental perception, comprising:
[0009] An environmental sensing unit is used to collect multi-dimensional environmental data in real time.
[0010] The environmental sensing unit includes a multispectral illumination sensor, a solar position calculation module, and a sky condition analysis module;
[0011] Multispectral illumination sensors are used to measure and output total solar irradiance and visible spectrum illumination in real time;
[0012] The solar position calculation module is used to obtain solar trajectory data for the current and future times based on built-in latitude and longitude coordinates and a real-time clock, and to obtain the theoretical maximum solar irradiance received by the Earth's surface at the current solar position.
[0013] The Sky Condition Analysis module is used to capture hemispherical sky images and identify cloud cover density type and cloud motion vector field through image processing technology.
[0014] The prediction model unit is used to predict the illumination prediction sequence within a future target time window based on cloud cover density type and cloud motion vector field, combined with solar trajectory data.
[0015] A central processing unit is used to receive real-time multidimensional environmental data and future illumination prediction sequences, execute an adaptive control method, and generate a tint control strategy for controlling the optical state of the photochromic glass window.
[0016] A control execution unit is used to receive a chromaticity control strategy and drive the photochromic glass window to switch optical states.
[0017] As a further improvement to this technical solution, the multispectral illumination sensor can receive the current theoretical maximum solar irradiance from the solar position calculation module, compare the real-time total solar irradiance with the current theoretical maximum solar irradiance, and use spectral distribution characteristics analysis to obtain the values of the direct sunlight component and the sky diffuse light component in the current ambient illumination.
[0018] As a further improvement to this technical solution, the sky condition analysis module includes a cloud optical thickness classification submodule and a cloud motion vector field generation submodule.
[0019] The cloud optical thickness classification submodule is based on hemispherical sky images and classifies clouds into at least three categories: high-transmittance clouds, medium-transmittance clouds, and low-transmittance clouds, and assigns a standard optical attenuation coefficient range to each category of clouds.
[0020] The cloud motion vector field generation submodule generates a two-dimensional cloud motion vector field covering the entire sky area by performing optical flow calculations on hemispherical sky images, thereby obtaining the average moving speed and direction of the dominant cloud layer.
[0021] As a further improvement to this technical solution, the specific steps for establishing the relative motion model of the sun and clouds in the prediction model unit are as follows:
[0022] S21. Based on the solar trajectory data provided by the solar position calculation module, the solar position at each moment within the future target time window is mapped onto the sky image plane in the same coordinate system output by the sky condition analysis module, generating a continuous future solar trajectory curve.
[0023] S22. Based on the cloud optical thickness classification results and cloud motion vector field output by the sky condition analysis module, perform the following operations on the sky image plane:
[0024] Identify and outline the contours of different types of cloud formations;
[0025] Each cloud cluster outline is assigned a corresponding cloud type attribute and a standard optical attenuation coefficient determined by the cloud optical thickness classification submodule.
[0026] Based on the average moving speed and direction of the dominant cloud layer, a dynamic cloud model is established for the outline of each cloud cluster to predict its future positional changes.
[0027] S23. Perform spatiotemporal overlay analysis on the future trajectory curve of the sun generated in S21 and the dynamic cloud model established in S22. By calculating the geometric relationship between the outline of the moving cloud and the stationary trajectory of the sun, predict the sequence of events in the future target time window where the sun is obscured by clouds and the sun is exposed.
[0028] S24. Sort and integrate all the events of the sun being blocked by clouds and the sun being exposed as predicted in S23 in chronological order to generate a structured future illumination prediction sequence.
[0029] As a further improvement to this technical solution, the prediction model unit connects to the external regional meteorological data stream interface, and performs data fusion and cross-validation between the information from the external regional meteorological data stream interface and the cloud cover density type and cloud motion vector field provided by the sky condition analysis module, thereby correcting the boundary conditions of the cloud motion vector field and extending the effective prediction duration of the future target time window.
[0030] As a further improvement to this technical solution, when the central processing unit executes the adaptive control method, it treats the direct sunlight component and the sky diffuse light component as independent input variables, and assigns different control weights to the direct sunlight component and the sky diffuse light component respectively, wherein the control weight of the direct sunlight component is higher than the control weight of the sky diffuse light component.
[0031] As a further improvement to this technical solution, the adaptive control method executed in the central processing unit includes a feedforward control decision mechanism, a lifecycle management mechanism, and a signal filtering and execution mechanism;
[0032] Among them, the feedforward control decision mechanism is used to identify strong light events based on the future illumination prediction sequence output by the prediction model unit, and generate advanced shading control instructions according to the strong light events.
[0033] The lifecycle management mechanism is used to quantify the loss caused by the switching of the optical state of the photochromic glass window into a cost function, and to prioritize the control strategy that minimizes this cost as the chromaticity control strategy in the decision-making process.
[0034] The signal filtering and execution mechanism is used to determine the final triggering of the state switching command for the color-changing glass window based on the dynamic change rate threshold and the minimum duration.
[0035] As a further improvement to this technical solution, the feedforward control decision mechanism is specifically as follows:
[0036] Continuously monitor the future illumination prediction sequence output by the prediction model unit;
[0037] When the sequence indicates that a strong direct sunlight event will occur in the first time in the future, a high-priority control command is triggered, which is to start the tinting state of the tinted glass window in advance or deepen it.
[0038] The threshold for determining strong direct sunlight incident events is based on the location of the tinted glass window, solar irradiance, and the functional requirements of the indoor space.
[0039] As a further improvement to this technical solution, the lifecycle management mechanism is implemented through a state transition cost function, and the specific steps are as follows:
[0040] S31. A state switching cost function is pre-constructed, which is used to quantitatively evaluate the negative impact of optical state switching of the photochromic window;
[0041] The state switching cost function takes at least the energy consumption caused by each state switch and the cumulative loss on the cycle life of the electrochromic material as calculation factors, and outputs a switching cost estimate.
[0042] S32. When generating the shading control strategy, the central processing unit takes maintaining indoor light and heat comfort, maximizing building energy-saving benefits and minimizing the total switching cost obtained from the state switching cost function as a comprehensive objective.
[0043] S33. When multiple feasible control strategies exist, the central processing unit is configured to: compare the switching cost estimates corresponding to the multiple feasible control strategies, and preferentially select the control strategy with the smallest switching cost estimate as the shading control strategy.
[0044] As a further improvement to this technical solution, the specific steps of the signal filtering and execution mechanism are as follows:
[0045] S34. Set a threshold for the dynamic rate of change of light intensity, which is dynamically adjusted based on the prediction results of the prediction model unit:
[0046] When stable sky conditions are predicted, the first value is used as the dynamic rate of change threshold.
[0047] When a dramatic change in sky conditions is predicted, a second value lower than the first value is used as the dynamic change rate threshold.
[0048] S35. Set a preset minimum duration determination window, with a time range of 30 to 60 seconds;
[0049] S36. Confirm and generate the switching command for the optical state of the photochromic window only if both of the following conditions are met simultaneously:
[0050] Condition 1: The rate of change of light intensity exceeds the currently set dynamic change rate threshold;
[0051] Condition 2: The predicted change in light intensity will continue for more than the minimum duration judgment window.
[0052] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0053] 1. In this adaptive control system for color-changing glass windows based on environmental perception, a forward-looking control mechanism is constructed by fusing perception through "multispectral analysis and sky imaging" and combining "solar trajectory and cloud movement" for joint prediction. This enables the system to issue control commands in advance before strong light reaches the window, fundamentally eliminating control lag and oscillation.
[0054] 2. In this environmentally-aware adaptive control system for photochromic windows, equipment lifespan management is incorporated into the core control objective. By introducing a state switching cost function and an intelligent filtering mechanism, the system can autonomously distinguish between primary and secondary issues related to lighting changes, prioritizing responses to critical changes in direct sunlight while intelligently filtering out transient interference and unnecessary minor adjustments. This not only significantly improves visual comfort but also reduces ineffective switching at the decision-making stage, directly extending the lifespan of the photochromic windows. Attached Figure Description
[0055] Figure 1 This is an overall flowchart of the present invention;
[0056] Figure 2 This is a system block diagram of the environmental sensing unit in this invention;
[0057] The meanings of the labels in the diagram are as follows:
[0058] 1. Environmental perception unit; 11. Multispectral illumination sensor; 12. Solar position calculation module; 13. Sky condition analysis module; 2. Prediction model unit; 3. Central processing unit; 4. Control execution unit. Detailed Implementation
[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0060] Example: Please refer to Figures 1-2 As shown, an adaptive control system for a color-changing glass window based on environmental perception is provided, including an environmental perception unit 1, a prediction model unit 2, a central processing unit 3, and a control execution unit 4.
[0061] Among them, the environmental sensing unit 1 is used to collect multi-dimensional environmental data in real time;
[0062] The environmental perception unit 1 includes a multispectral illumination sensor 11, a solar position calculation module 12, and a sky condition analysis module 13;
[0063] The multispectral light sensor 11 is used to measure and output the total solar irradiance and the illuminance in the visible spectrum in real time. The multispectral light sensor 11 is a module that integrates different functional sensors. The total solar irradiance is measured using a total radiometer, and the illuminance in the visible spectrum is measured using a photodiode with a V(λ) correction filter. When the system is working, the two devices are sampled synchronously, and the measured irradiance and illuminance values are sent to the central processing unit 3 in real time.
[0064] The multispectral illumination sensor 11 can receive the current theoretical maximum solar irradiance from the solar position calculation module 12, compare the real-time total solar irradiance with the current theoretical maximum solar irradiance, and use spectral distribution characteristics analysis to obtain the values of the direct sunlight component and the sky diffuse light component in the current ambient illumination.
[0065] The solar position calculation module 12 is used to obtain solar trajectory data for the current and future times based on the built-in latitude and longitude coordinates and real-time clock, and to obtain the theoretical maximum solar irradiance received by the Earth's surface at the current solar position.
[0066] The current theoretical maximum solar irradiance is calculated by the solar position calculation module 12 using an internally embedded astronomical and atmospheric physics model, as follows:
[0067] Under ideal conditions of clear skies, cloudless skies, and extremely clean atmosphere, the solar irradiance reaching the Earth's surface depends primarily on the path length of sunlight as it travels through the atmosphere. This path length is directly determined by the solar altitude angle, which can be precisely calculated using the module's built-in latitude and longitude coordinates and real-time clock.
[0068] The solar position calculation module 12 first calculates the current solar altitude angle based on the date and time. Then, it calls a pre-stored atmospheric top-level solar irradiance model (for example, referring to the standard solar spectral irradiance under the condition of "Air Mass (AM) 1.5" as defined in the ASTM standard) and performs interpolation or calculation in combination with the calculated atmospheric mass (AM=1 / sin(solar altitude angle)).
[0069] The model ultimately outputs the theoretically maximum total solar irradiance value achievable at the current geographical location and time. This value represents the sum of direct sunlight and some sky-scattered light under perfectly clear weather conditions. It serves as a dynamically changing theoretical benchmark value and is sent to the multispectral illumination sensor 11 for analysis.
[0070] Sky Condition Analysis Module 13 is used to capture hemispherical sky images and identify cloud cover density type and cloud motion vector field through image processing technology;
[0071] The hemispherical sky image is acquired by a sky imager; the hemispherical sky image contains at least the pixel brightness, color features, and relative position of the cloud region to the sun.
[0072] Sky Condition Analysis Module 13 includes a cloud optical thickness classification submodule and a cloud motion vector field generation submodule;
[0073] The cloud optical thickness classification submodule is based on hemispherical sky images and classifies clouds into at least three categories: high-transmittance clouds, medium-transmittance clouds, and low-transmittance clouds, and assigns a standard optical attenuation coefficient range to each category of clouds.
[0074] Specifically, if the overall image brightness increase is not significant, the sun's outline is still visible but slightly blurred, and the sky remains pale blue, and the brightness attenuation when passing through the sun's region is small, then the clouds are classified as high-transmittance clouds, with an optical attenuation coefficient range of 0.1-0.3 (i.e., transmittance 70%-90%). If the image brightness increases significantly, the sun's outline is blurred or completely obscured, but the surrounding area is not particularly dark, and the sky is white or grayish-white, then the clouds are classified as medium-transmittance clouds, with an optical attenuation coefficient range of 0.4-0.7 (i.e., transmittance 30%-60%). If the corresponding area in the image is very dark (especially when the cloud layer is extremely thick), dark gray in color, and the sun is completely obscured, with extremely low brightness in this area, then the clouds are classified as low-transmittance clouds, with an optical attenuation coefficient range of 0.8-0.95 (i.e., transmittance 5%-20%).
[0075] In this way, when a cloud is predicted to block the sun, the predictive model not only knows "when" the cloud will block the sun, but also estimates "how much" the irradiance will decrease after the cloud blockage, based on the type of cloud. This allows the system to determine in advance whether the blocking event is severe enough to require changing the state of the glass, thus making more accurate decisions.
[0076] The cloud motion vector field generation submodule generates a two-dimensional cloud motion vector field covering the entire sky area by performing optical flow calculations on hemispherical sky images, thereby obtaining the average moving speed and direction of the dominant cloud layer.
[0077] The cloud motion vector field generation submodule uses optical flow technology to convert continuous sky images into a dynamic vector field, accurately measuring the movement speed and direction of the clouds. The specific steps are as follows:
[0078] S1. Input a sequence of two or more consecutive preprocessed hemispherical sky images. and In the formula, For the first frame image at a specific time Located at coordinates The brightness value at that location; For the second frame image at the time interval Then, located at coordinates The brightness value at the location; convert the image to grayscale and perform Gaussian filtering to suppress noise.
[0079] S2. Assume that the brightness of a cloud pixel remains constant as it moves between consecutive frames. That is... .
[0080] Performing a first-order Taylor expansion on the right-hand side, we obtain the fundamental equation for optical flow:
[0081]
[0082] In the formula, For the image in Spatial gradient of direction; For the image in Spatial gradient of direction; It is the gradient of the image over time (i.e., the difference between the next frame and the previous frame). For pixels in Speed of motion in a certain direction; For pixels in Speed of motion in a certain direction;
[0083] S3. Using the Lucas-Kanade method, it is assumed that within a small image window (e.g., 5x5 pixels), all pixels have the same motion. .
[0084] For each pixel within the window, an equation can be derived based on the optical flow equation, resulting in an overdetermined system of equations. Solving this system of equations using the least squares method yields the motion vector at the center of the window. The solution is:
[0085]
[0086] in, The design matrix is composed of the spatial gradients of all pixels within a local window. From all pixels within the window Composition, vector From all pixels composition; The transposed matrix .
[0087] Specifically: For a person with A window of pixels, a matrix It is A matrix of the form of ,vector It is The column vectors are of the form...
[0088] S4. By calculating thousands of such windows in the image, the system can generate a two-dimensional cloud motion vector field describing the cloud motion. Then, through cluster analysis, the average moving speed and direction of the dominant cloud are obtained. By combining this with the sun's trajectory, the situation of the sun being blocked in the next few minutes can be predicted.
[0089] Prediction model unit 2 is used to predict the illumination prediction sequence within the future target time window based on cloud cover density type and cloud motion vector field, combined with solar trajectory data;
[0090] In prediction model unit 2, the specific steps for establishing the relative motion model of the sun and clouds are as follows:
[0091] S21. Based on the solar trajectory data provided by the solar position calculation module 12, the solar position at each moment within the future target time window is mapped onto the sky image plane in the same coordinate system output by the sky condition analysis module 13, generating a continuous future solar trajectory curve; the solar trajectory data includes at least the solar altitude angle and azimuth angle.
[0092] S22. Based on the cloud optical thickness classification results and cloud motion vector field output by the sky condition analysis module 13, the following operations are performed on the sky image plane:
[0093] Identify and outline the contours of different types of cloud formations;
[0094] Each cloud cluster outline is assigned a corresponding cloud type attribute and a standard optical attenuation coefficient determined by the cloud optical thickness classification submodule.
[0095] Based on the average moving speed and direction of the dominant cloud layer, a dynamic cloud model is established for the outline of each cloud cluster to predict its future positional changes.
[0096] S23. Perform spatiotemporal overlay analysis on the future trajectory curve of the sun generated in S21 and the dynamic cloud model established in S22. By calculating the geometric relationship between the outline of the moving cloud and the stationary trajectory of the sun, predict the sequence of events in the future target time window where the sun is obscured by clouds and the sun is exposed.
[0097] The forecast specifically includes:
[0098] Start time prediction: Calculate the moment when the leading edge of each cloud profile intersects with the solar trajectory, and use this as the start time of the occlusion event;
[0099] Duration prediction: Calculate the chord length of the sun's trajectory across the cloud's outline, and combine this with the cloud's movement speed to calculate the duration of the obstruction;
[0100] Irradiance attenuation prediction: Based on the cloud type corresponding to the cloud that caused this shading and its standard optical attenuation coefficient, the degree of solar irradiance attenuation caused by this shading event is determined.
[0101] S24. Sort and integrate all the events of solar obscuration and solar exposure predicted in S23 in chronological order to generate a structured future illumination prediction sequence. This sequence clearly lists the start time, duration and corresponding irradiance attenuation of the sun being obscured or exposed by different cloud types within the future target time window.
[0102] Prediction model unit 2 is connected to the external regional meteorological data stream interface, which is used to receive real-time weather forecast data, satellite cloud images and radar echo images from professional meteorological service agencies;
[0103] The prediction model unit 2 integrates and cross-validates information from the external regional meteorological data stream interface with the cloud cover density type and cloud motion vector field provided by the sky condition analysis module 13, thereby correcting the boundary conditions of the cloud motion vector field and extending the effective prediction duration of the future target time window, thus improving the reliability of predictions for long-term, large-area cloudy or clear weather caused by large-scale weather systems (such as frontal passage).
[0104] The central processing unit 3 is used to receive real-time multi-dimensional environmental data and future illumination prediction sequences, execute adaptive control methods, and generate a tint control strategy for controlling the optical state of the photochromic glass window.
[0105] When executing the adaptive control method, the central processing unit 3 treats the direct sunlight component and the sky diffuse light component as independent input variables, and assigns different control weights to the direct sunlight component and the sky diffuse light component respectively, wherein the control weight of the direct sunlight component is higher than the control weight of the sky diffuse light component.
[0106] The weighting is based on the different effects of the two light components on the indoor environment and building energy consumption. A pre-defined strategy in the control algorithm of the central processing unit 3 is used: a weighting coefficient is set so that the control algorithm's sensitivity to the direct sunlight component is 3 to 5 times that of the diffused light component. In this embodiment, the direct sunlight weighting factor is set to 0.8; the sky diffused light weighting factor is set to 0.2.
[0107] The reason for setting the control weights in this way is that direct sunlight is the main culprit for indoor heat gain (the main source of cooling load) and severe glare (the primary factor affecting visual comfort and health), while skylight is usually softer and its thermal effect and glare-causing ability are far lower than that of direct sunlight.
[0108] This allows the system to respond more sensitively to direct light that actually raises room temperature, enabling it to activate shading more promptly and effectively to block heat and achieve more precise energy consumption control. It avoids malfunctions caused by increased total illuminance (primarily harmless diffused light), saving energy. Furthermore, it can take proactive measures based on the predicted direct light component before strong glare that might cause discomfort to the user occurs, greatly improving visual comfort.
[0109] The adaptive control methods implemented in the central processing unit 3 include a feedforward control decision-making mechanism, a lifecycle management mechanism, and a signal filtering and execution mechanism.
[0110] Among them, the feedforward control decision mechanism is used to identify strong light events based on the future illumination prediction sequence output by the prediction model unit 2, and generate advanced shading control instructions according to the strong light events.
[0111] The lifecycle management mechanism is used to quantify the loss caused by the switching of the optical state of the photochromic glass window into a cost function, and to prioritize the control strategy that minimizes this cost as the chromaticity control strategy in the decision-making process.
[0112] The signal filtering and execution mechanism is used to determine the final triggering of the state switching command for the photochromic glass window based on the dynamic change rate threshold and the minimum duration.
[0113] The feedforward control decision-making mechanism is as follows:
[0114] Continuously monitor the future illumination prediction sequence output by the prediction model unit 2;
[0115] When the sequence indicates that a strong direct sunlight event will occur in the first time period in the future, where the value of the first time period is between 5 minutes and 15 minutes, a high-priority control command is triggered. The high-priority control command is to start the tinting state of the tinted glass window in advance or deepen it.
[0116] The threshold for judging strong direct sunlight incident events is determined based on the location of the tinted glass window, solar irradiance, and the functional requirements of the indoor space.
[0117] The lifecycle management mechanism is implemented through a state transition cost function, and the specific steps are as follows:
[0118] S31. A state switching cost function is pre-constructed, which is used to quantitatively evaluate the negative impact of optical state switching of the photochromic window;
[0119] The state switching cost function takes at least the energy consumption caused by each state switch and the cumulative loss on the cycle life of the electrochromic material as calculation factors, and outputs a switching cost estimate.
[0120] As a preferred implementation, the state transition cost function Represented as:
[0121] ;
[0122] in, This is the change in chromaticity. The coefficient is used to characterize the loss of material life. Weighted by lifetime loss; This is the estimated energy consumption for this switching action, and its coefficient is... Energy consumption weighting. The estimated switching cost is... The calculation results.
[0123] The process of estimating the cost of output switching is based on the control weights of the direct sunlight component and the sky diffuse light component. The direct light component (high weight) is the main cause of discomfort and energy consumption, so switching the state in response to its changes is "highly profitable and highly necessary," even at the cost of a certain lifespan.
[0124] The impact of changes in the scattered light component (low weight) is relatively small. If glass switching is triggered only in response to small fluctuations in the scattered light, it is considered a "low-return, high-cost" behavior and will be subject to a high penalty in the cost function.
[0125] Therefore, distinguishing and weighting these two components is the prerequisite and computational basis for the lifecycle management mechanism to intelligently and selectively reduce unnecessary switching.
[0126] S32. When generating the shading control strategy, the central processing unit 3 takes maintaining indoor light and heat comfort, maximizing building energy-saving benefits and minimizing the total switching cost obtained from the state switching cost function as a comprehensive objective.
[0127] S33. When multiple feasible control strategies exist, the central processing unit 3 is configured to: compare the switching cost estimates corresponding to the multiple feasible control strategies, and preferentially select the control strategy with the lowest switching cost estimate as the shading control strategy. The feasible control strategy can meet the preset indoor light and heat comfort and basic energy-saving requirements.
[0128] The specific steps of the signal filtering and execution mechanism are as follows:
[0129] S34. Set a threshold for the dynamic rate of change of light intensity, which is dynamically adjusted based on the prediction results of prediction model unit 2:
[0130] When stable sky conditions are predicted, the first value is used as the dynamic rate of change threshold.
[0131] When a dramatic change in sky conditions is predicted, a second value lower than the first value is used as the dynamic change rate threshold.
[0132] The first value (steady-state threshold) can be set as the rate of change of total irradiance within 1 minute exceeding [a certain threshold]. The second value (threshold during drastic changes) can be set when the rate of change exceeds... The specific values can be calibrated within the stated range based on building sensitivity.
[0133] S35. Set a preset minimum duration determination window, with a time range of 30 to 60 seconds;
[0134] S36. Confirm and generate the switching command for the optical state of the photochromic window only if both of the following conditions are met simultaneously:
[0135] Condition 1: The rate of change of light intensity exceeds the currently set dynamic change rate threshold;
[0136] Condition 2: The predicted change in light intensity will continue for more than the minimum duration judgment window.
[0137] This mechanism is used to effectively filter out brief, minute fluctuations in illumination caused by birds, fleetingly passing thin clouds, or sensor noise, thereby working in conjunction with the state switching cost function to further suppress oscillations in the control system.
[0138] The control execution unit 4 is used to receive the chromaticity control strategy and drive the photochromic glass window to switch optical states.
[0139] The control execution unit 4 drives the photochromic glass window to adaptively switch its optical state in a high-precision, stepless, and gradual manner, specifically as follows:
[0140] The control strategy generated by the central processing unit 3 includes an optimal chromaticity change trajectory curve based on the predicted event timeline. This curve defines the functional relationship of chromaticity changing smoothly over time. The control execution unit 4, based on this trajectory curve, precisely and smoothly controls the optical state of the chromatic glass window to evolve along the trajectory, thereby achieving a comfortable transition of natural lighting environment that is imperceptible to the user and completely avoiding the visual abruptness and discomfort caused by the step-like switching of chromaticity.
[0141] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.
Claims
1. An environmentally aware self-adaptive control system for a variable tint glazing, characterized in that, Comprise: An environment perception unit (1) for collecting multi-dimensional environment data in real time; The environment perception unit (1) comprises a multi-spectral light sensor (11), a sun position calculation module (12) and a sky condition analysis module (13); The multi-spectral light sensor (11) is used for measuring and outputting the total solar irradiance and the visible light spectrum illumination intensity in real time; The sun position calculation module (12) is used for obtaining the sun trajectory data at the current and future time based on the built-in latitude and longitude coordinates and the real-time clock, and obtaining the current solar theoretical maximum irradiance received by the ground surface under the current sun position; The sky condition analysis module (13) is used for capturing a hemispherical sky image, identifying the cloud cover density type and cloud motion vector field through image processing technology; A prediction model unit (2) for predicting a light prediction sequence in a future target time window based on the cloud cover density type and the cloud motion vector field, and combining the sun trajectory data; A central processing unit (3) for receiving real-time multi-dimensional environment data and future light prediction sequence, executing an adaptive control method, and generating a shading degree control strategy for controlling the optical state of the variable color glass window; A control execution unit (4) for receiving the shading degree control strategy and driving the variable color glass window to switch the optical state.
2. The environmental perception based adaptive control system for a chromic glazing according to claim 1, wherein: The multi-spectral light sensor (11) can receive the current solar theoretical maximum irradiance from the sun position calculation module (12), compare the real-time total solar irradiance with the current solar theoretical maximum irradiance, and obtain the values of direct sunlight component and sky scattered light component in the current environmental light by analyzing the spectral distribution characteristics.
3. The environmental perception based adaptive control system of the variable tint glass window according to claim 2, wherein: The sky condition analysis module (13) comprises a cloud optical thickness classification submodule and a cloud motion vector field generation submodule; The cloud optical thickness classification submodule classifies the cloud into at least three categories of high light transmittance cloud, medium light transmittance cloud and low light transmittance cloud based on the hemispherical sky image, and assigns each type of cloud a standard optical attenuation coefficient range; The cloud motion vector field generation submodule generates a two-dimensional cloud motion vector field covering the entire sky area by calculating the optical flow of the hemispherical sky image, and then obtains the average moving speed and direction of the dominant cloud.
4. The environmentally aware, self-adapting control system for a chromic glazing according to claim 3, wherein: In the prediction model unit (2), the specific steps of establishing a sun-cloud relative motion model are as follows: S21, based on the sun trajectory data provided by the sun position calculation module (12), the sun position at each time in the future target time window is mapped to the sky image plane in the same coordinate system output by the sky condition analysis module (13), generating a continuous sun future trajectory curve; S22, based on the cloud optical thickness classification result and the cloud motion vector field output by the sky condition analysis module (13), the following operations are performed on the sky image plane: Identify and outline the contours of different types of cloud clusters; Assign each cloud cluster contour its corresponding cloud type attribute and the standard optical attenuation coefficient determined by the cloud optical thickness classification submodule; A dynamic cloud model is established for each cloud profile based on the average moving speed and direction of the dominant cloud layer, and the future position change of the cloud is predicted; S23, the future trajectory curve of the sun generated by S21 is spatiotemporally superimposed with the dynamic cloud model established by S22, the geometric relationship between the moving cloud profile and the static sun trajectory is calculated, and the sequence of events of the sun being obscured by the cloud and the sun being exposed in the future target time window is predicted; S24, all the events of the sun being obscured by the cloud and the sun being exposed predicted by S23 are sorted and integrated in time sequence to generate a structured future light prediction sequence.
5. The environmentally aware, self-adapting control system for a chromic glazing according to claim 4, wherein: The prediction model unit (2) accesses an external regional meteorological data stream interface, fuses and cross- validates the information from the external regional meteorological data stream interface with the cloud coverage density type and the cloud motion vector field provided by the sky condition analysis module (13), thereby correcting the boundary conditions of the cloud motion vector field and extending the effective prediction length of the future target time window.
6. The environmentally aware, self-adapting control system for a chromic glazing according to claim 5, wherein: The central processing unit (3) takes the direct sunlight component and the sky scattered light component as independent input variables when executing the adaptive control method, and gives different control weights to the direct sunlight component and the sky scattered light component, respectively, wherein the control weight of the direct sunlight component is higher than that of the sky scattered light component.
7. The environmentally aware, self-adapting control system for a chromic glazing according to claim 6, wherein: The adaptive control method executed in the central processing unit (3) includes a feedforward control decision mechanism, a life cycle management mechanism, and a signal filtering and execution mechanism; The feedforward control decision mechanism is used to identify strong light events based on the future light prediction sequence output by the prediction model unit (2), and to generate advanced tint control instructions according to the strong light events; The life cycle management mechanism quantifies the loss caused by the switching of the optical state of the variable color glass window as a cost function, and preferentially selects the control strategy that minimizes the cost as the tint control strategy in decision-making; The signal filtering and execution mechanism is used to determine the final triggering of the variable color glass window state switching instruction according to the dynamic change rate threshold and the minimum duration.
8. The environmentally aware, self-adapting control system for a chromic glazing according to claim 7, wherein: The feedforward control decision mechanism specifically includes: Continuously monitor the future light prediction sequence output by the prediction model unit (2); When the sequence indicates that a strong direct sunlight incident event will occur within a first future time, a high-priority control instruction is triggered, which is to start or deepen the tinting state of the variable color glass window in advance; The judgment threshold of the strong direct sunlight incident event is determined based on the position of the variable color glass window, the solar irradiance, and the functional requirements of the indoor space.
9. The environmentally aware, self-adapting control system for a chromic glazing according to claim 8, wherein: The life cycle management mechanism is realized through a state switching cost function, and the specific steps are as follows: S31, a state switching cost function is constructed in advance, which is used to quantitatively evaluate the negative effects of the optical state switching of the variable color glass window; The state switching cost function at least takes the energy consumption caused by each state switching and the cumulative damage to the cycle life of the electrochromic material as calculation factors, and outputs a switching cost estimate; S32, in generating the tint control strategy, the central processing unit (3) takes maintaining indoor light thermal comfort, maximizing building energy saving benefits and minimizing the total switching cost obtained by the state switching cost function as the comprehensive goal; S33, when there are multiple feasible control strategies, the central processing unit (3) is configured to compare the switching cost estimates corresponding to the multiple feasible control strategies, and preferentially select the control strategy with the minimum switching cost estimate as the tint control strategy.
10. The environmentally aware, self-adapting control system for a chromic glazing according to claim 9, wherein: The specific steps of the signal filtering and execution mechanism are as follows: S34, set a dynamic change rate threshold for the change of light intensity, which is dynamically adjusted according to the prediction results of the prediction model unit (2): When it is predicted that the sky condition is stable, a first value is used as the dynamic change rate threshold; When it is predicted that the sky condition will soon change dramatically, a second value lower than the first value is used as the dynamic change rate threshold; S35, set a preset minimum duration determination window, which has a time range of 30 seconds to 60 seconds; S36, only when the following two conditions are met at the same time, the switching instruction for the optical state of the variable color glass window is confirmed and generated: Condition one: the light intensity change rate exceeds the currently set dynamic change rate threshold; Condition two: the light intensity change state prediction will last more than the minimum duration determination window.