Indoor environment physical enhancement model establishment method based on optimization control of external window sunshade system

Abstract

The invention provides an indoor environment physical enhancement model establishment method based on optimization control of an external window sunshade system. The relation with the indoor environment is established in various functional forms by dividing various shielding types possibly occurring in the external window sunshade system. According to the method, the sparse recognition technology is introduced into the indoor environment quality model of the external window shading system, and the comprehensiveness of the model is improved by constructing the candidate function library and adding various functions influencing the indoor environment quality in the construction process. In the process of adding the function, each parameter of the function has clear physical significance, and the model can inherit the interpretability of the function. According to the indoor environment physical model established according to the model establishing method provided by the invention, the interpretability of the indoor environment model based on data driving is enhanced, and the model is simple, accurate and quick in response.

Description

Technical Field

[0001] The present application relates to the field of building energy-saving technology, and in particular to a method for establishing a physical enhancement model of an indoor environment based on optimized control of an exterior window shading system. Background Art

[0002] External window shading systems are one of the most fundamental building structures, fulfilling the crucial function of regulating the indoor environment. Specifically, the indoor thermal, light, acoustic, and air environments all rely on external window shading systems to regulate them. Therefore, the rational use of external window shading systems is key to reducing building energy consumption. Their flexibility allows for rapid response to user needs. However, how the indoor environment responds to user needs through adjustments made by external window shading systems requires a reliable model to describe.

[0003] Physically augmented models provide interpretability and generalizability through physical knowledge. Combining the advantages of data-driven models, such as high precision, fast computational speed, and strong ability to describe uncertainties, they have become a promising model for the green building industry. However, physical augmented models still have some shortcomings in market application. For example, the privacy of building automation operational data makes data acquisition more difficult, while the deployment of expensive sensors significantly increases operating costs. Establishing a physically augmented model requires designers to be proficient in both machine learning algorithms and building-related physical knowledge. However, accurate descriptions of building physics cannot be derived from universal laws, which increases the difficulty of model establishment. In terms of applicability, while physically augmented models are highly interpretable, their prediction accuracy is inferior to that of data-driven models, and they often lack the trust of managers.

[0004] Therefore, how to enhance the interpretability of data-driven models is of great significance to the digital development of green buildings today. Summary of the Invention

[0005] This application provides a method for establishing a physical enhancement model of an indoor environment based on the optimization control of an external window shading system, so as to enhance the interpretability of the data-driven indoor environment model and make the model concise, accurate and responsive.

[0006] The technical solution of this application is as follows: A method for establishing a physical enhancement model of an indoor environment based on the optimization control of an exterior window shading system. In this method, all windows in an exterior window frame and interior sunshades together constitute the exterior window shading system. The windows include fixed windows and movable windows. The method specifically includes the following steps: S1. Divide the window shading type of each area of ​​the exterior window system according to the shading ratio and the position of the movable window; The window shading types include no shading, single-pane glass shading, double-pane glass shading, sun shading only, single-pane glass sun shading and combined shading, and double-pane glass sun shading and combined shading; S2. Calculate the area occupied by each window shading type based on the shading ratio and the sliding distance of the movable window; S3. Obtaining climate environment data, control instruction data, and time series data of the building main body based on the exterior window shading system under ideal operating conditions; The ideal operating state is the operating state when the composition and physical equipment of the building remain unchanged; The climate environment data includes indoor and outdoor temperature, indoor and outdoor illumination, indoor and outdoor noise, and indoor and outdoor carbon dioxide concentration; The control instruction data includes the opening ratio of the sunshade curtain, the rotation angle of the sunshade blind and the sliding distance of the movable window; S4. Express the indoor environment model of the building as a discrete-time dynamic model: ; Where, Y Indicates the predicted value of indoor environmental quality; A library of candidate functions representing indoor environment models; represents a sparse matrix; Represents the set of functions selected from the candidate function library; X Represents building climate environment data; U Indicates control instruction data; S5. Create a sparse matrix The sparse identification expression of the sparse matrix is ​​represented by The coefficients of each function term in ; S6: Substitute the climate environment data, control instruction data and time series data obtained in step S3 into , based on the least squares method, the error between the measured value of indoor environmental quality and the predicted value of indoor environmental quality is minimized to calculate the optimal sparse matrix, and a penalty term is introduced to use regularization to penalize the sparse matrix; S7. Construct an indoor environmental quality prediction model for the building according to the optimal sparse matrix and candidate function library calculated in step S6.

[0007] Furthermore, the candidate function library in step S4 includes constant terms, state variables, control variables, and thermal environment functions, light environment functions, sound environment functions, air quality functions, comfort functions, and trigonometric function terms represented by the state variables and control variables, wherein the trigonometric function terms are used to represent the periodic changes of the state variables and control variables.

[0008] Furthermore, the constant term in the candidate function library is used to characterize the baseline influence of the constant influencing factor, and its value is 1; The state variables are direct characteristic quantities of the indoor environment, including indoor temperature, outdoor temperature, indoor illumination, outdoor illumination, indoor carbon dioxide concentration, outdoor carbon dioxide concentration, indoor noise, and outdoor noise; The control variables include the sunshade curtain opening and closing ratio, the sunshade louver rotation angle, and the moving distance of the movable window.

[0009] Furthermore, the thermal environment function in the candidate function library is determined based on the physical relationship between the indoor temperature and the outdoor temperature under the control of the exterior window system. The physical relationship between the indoor temperature and the outdoor temperature is obtained based on the heat balance equation under steady-state conditions. The heat balance equation is as follows: ; Where, T in 、 T out Indicates indoor temperature and outdoor temperature respectively. U t represents the comprehensive heat transfer coefficient of the exterior window shading system, A Indicates the window area. I represents the solar radiation intensity (W / m2), g t Indicates the comprehensive solar transmittance of the exterior window shading system. c p represents the specific heat capacity, r represents the air density, G represents the volume flow of gas through the window, G’ and Q’ They represent the wind and heat gained from external disturbances of the environment respectively; t Indicates time; Calculate the comprehensive heat transfer coefficient of the exterior window shading system based on the area occupied by each window shading type: ; Where, Indicates the area blocked by a single piece of glass; It represents the area of ​​the area blocked by the composite sunshade of a single piece of glass; Indicates the area of ​​the area that is only shaded; Indicates the area blocked by double pane glass; It represents the area of ​​the double-pane glass sunshade composite shielding; 、 、 、 、 Respectively represent the heat transfer coefficient corresponding to each window shading type; Calculate the comprehensive solar transmittance of the exterior window shading system based on the area occupied by each window shading type: ; Where, represents the area of ​​the unobstructed region, 、 、 、 、 、 Respectively represent the solar transmittance coefficient corresponding to each window shading type; Calculate the gas volume flow through the window based on the window shading type of no shading and only shading: ; Where, v r Indicates wind speed, H 、 W Respectively represent the height and width of the outer window frame, Indicates the rotation angle of the sunshade louver; l Indicates the moving distance of the movable window; Remove the fixed physical parameters and extract the related items with the state variables and control variables to obtain the thermal environment function sub-library of the candidate function library: ; Where, Represents the thermal environment function sub-library of the candidate function library.

[0010] Furthermore, the light environment functions in the candidate function library include light environment comfort evaluation and the physical relationship between indoor illuminance and outdoor illuminance under the control of the external window system; The evaluation model of light environment comfort is as follows: ; Where, L index Indicates the comfort level of the light environment; Extracted from the evaluation model of light environment comfort ln ( ln(E in ))、 ln 2 ( ln(E in )) as a candidate function; The physical relationship between indoor illuminance and outdoor illuminance is defined as the ratio of outdoor light passing through the exterior window shading system, as follows: ; Where, t t It represents the comprehensive light transmittance of the exterior window shading system and can be quantified as: ; Where, 、 、 、 、 、 Respectively represent the light transmittance corresponding to each window shading type; After removing fixed physical parameters and extracting items related to state variables and control variables, we can obtain the light environment function sub-library of the candidate function library: ; Where, Represents the light environment function sub-library of the candidate function library.

[0011] Furthermore, the acoustic environment function in the candidate function library is determined based on the physical relationship between indoor noise and outdoor noise, as follows: ; Where, L’ represents the external noise of the environment, R s It represents the comprehensive sound insulation of the maintenance structure, which is characterized by the average sound transmission coefficient, as follows: ; Where, t s represents the average sound transmission coefficient, A w Indicates the wall area, t s,w is the wall sound transmission coefficient, t s,t is the sound transmission coefficient of the exterior window shading system, and t s,t The calculation formula is as follows: ; Where, R Indicates the sound insulation of glass; After extracting the items related to the state variables and control variables, the acoustic environment function sub-library of the candidate function library is obtained: ; Where, The acoustic environment function sub-library representing the candidate function library; The air quality functions in the candidate function library include air quality comfort and the physical relationship between indoor and outdoor carbon dioxide concentrations under the control of the external window shading system; The air quality comfort model is: ; Where, IAQ index Indicates air quality comfort; Extract in the above formula As a candidate function; The physical relationship between indoor and outdoor carbon dioxide concentrations under the control of the exterior window shading system is as follows: ; Where, V Indicates the volume of the room; In the above formula, after removing the related items of fixed physical parameter extraction, state variables and control variables, the air quality function sub-library of the candidate function library is obtained by combining the selected candidate functions: ; Where, Represents the air quality function sub-library.

[0012] Furthermore, the comfort function in the candidate function library is determined based on the variable weight model of thermal environment comfort, light environment comfort, acoustic environment comfort, air quality comfort model, and indoor comprehensive environment comfort, wherein the thermal environment comfort model is as follows: ; Where, TC index represents the thermal environment comfort model; The light comfort model is as follows: ; Where, L index Represents the light environment comfort model; The acoustic environment comfort model is as follows: ; Where, AC index represents the acoustic environment comfort model; The air quality comfort model is as follows: ; Where, IAQ index represents the air quality comfort model; The variable weight model of indoor comprehensive environmental comfort is as follows: ; Where, IEQ index represents the comprehensive indoor environmental comfort function; m i ( i =1, 2, 3, 4) represents the weight of the corresponding comfort model, and: ; Where, x represents the calculated value of a single comfort level, m 0 is the initial constant weight obtained after linear regression, a i is the upper limit of the comfort index prediction model, b i is the weight adjustment parameter; After removing the related items of fixed physical parameter extraction, state variables and control variables from the above formula, the comfort function sub-library of the candidate function library is obtained: ; Where, Represents the comfort function sub-library; The trigonometric functions in the candidate function library are: ; Where, Represents the set of trigonometric function items in the candidate function library; The constant terms, state variables, trigonometric function terms and each function sub-library of the candidate function library are integrated to obtain the candidate function library: .

[0013] Furthermore, the calculation method of the sparse matrix in step S6 is as follows: ; Where, represents the sparse matrix solution, Show the general k The state variables and control variables at the moment are input into the dynamic model and calculated k+ The predicted value of indoor environmental quality at moment 1, Y k+1 express k+ The measured value of indoor environmental quality at moment 1, represents the penalty term coefficient; For the above formula: The data obtained from S3 is divided into training set and validation set according to the time series, and the The preset value set is extracted in turn. After substituting the preset value and the training set data into the above formula, the least squares method and l 1 Regularization calculation, get each The final sparse matrix corresponding to the preset values X f , determine the optimal penalty coefficient based on usage requirements And the optimal sparse matrix solution corresponding to the coefficient Xb .

[0014] Due to the adoption of the above technical solution, the beneficial effects of this application are as follows: 1. This application innovatively introduces sparse recognition technology into the indoor environmental quality model of the exterior window shading system, breaking through the contradiction between the interpretability and accuracy of the data-driven model. In the technical solution of this application, the candidate function library is the basis of modeling. The candidate function library can be constructed based on prior experience or known physical parameters of the building body. During the construction process, the comprehensiveness of the model is improved by adding various functions that affect the indoor environmental quality. In the process of adding functions, the various parameters of the function have clear physical meanings, and the function is interpretable, so the model finally established can inherit interpretability. Finally, relying on a large amount of measured data for function calculation, the final model is interpretable while also having the accuracy brought by data-driven.

[0015] 2. The computational efficiency of the technical solution of this application is greatly improved. In the sparse recognition algorithm, the measured data is input into a sparse matrix, and the matrix is ​​then mapped into a function library. By utilizing the sparsity of the dynamic system model itself, its dynamic terms are framed in a candidate function library established based on relevant physical knowledge and prior knowledge, thus completing the dimensionality reduction of the data and avoiding mechanical combinatorial search. At the same time, with the advantages of data-driven algorithms, the complexity of the operation is greatly reduced and the operation efficiency is improved.

[0016] 3. This application places the regulation mechanism for indoor environmental quality within the exterior window shading system. By enumerating the various types of shading that can occur, the application details the various functional forms of the exterior window shading system and establishes a connection between these functional forms and the indoor environment. The regulation of the exterior window shading system is environmentally friendly, economical, and real-time, which also makes the resulting model have broad application prospects.

[0017] 4. The model building results of this application can be adjusted according to demand. In this application, the coefficients of each function in the function database are controlled by penalty items, thereby controlling the accuracy and simplicity of the model. For example, if the subsequent optimization control of the building body requires a simple model, a model corresponding to a relatively high λ value can be selected, but the model accuracy is low. If the subsequent optimization control of the building body requires a high-precision model, a model corresponding to a relatively low λ value can be selected. In this case, more candidate function items are activated, and the model accuracy is high but relatively complex. In addition, in the indoor environment of the building body, when the interaction effect of indoor and outdoor environmental elements is significant, the system will have a collinearity problem. This application also adopts l 1 regularization to penalize sparse coefficients. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0019] Figure 1 This is a schematic diagram of the exterior window shading system in this application; Figure 2 This application provides a flow chart of a method for establishing a physical enhancement prediction model for an indoor environment based on optimized control of an exterior window shading system. DETAILED DESCRIPTION

[0020] Based on the shortcomings of the physical enhancement model proposed in the background technology, as shown in the attached Figure 1 and attached Figure 2 As shown, the present application provides a method for establishing a physical enhancement prediction model of an indoor environment based on the optimization control of an external window shading system. In this method, all windows in the external window frame and indoor sunshades together constitute an external window shading system, and the windows include fixed windows and movable windows.

[0021] The outer window frame contains several windows and sunshades. The blinds of the sunshades adjust the amount of light entering by rotating, and the sunshades control the amount of air and light entering by moving up and down.

[0022] As attached Figure 1 As shown, in an embodiment of the present application, the window of the external window shading system includes a fixed window at the top and a movable window at the bottom. The entire external window frame contains two layers of windows, the upper layer being fixed single-layer windows and the lower layer being movable single-layer windows. The window shading types of each area of ​​the external window system are divided according to the shading ratio of the external window shading system and the position of the movable windows in this embodiment. By moving the lower layer windows, this application forms a total of 6 window shading types. It should be noted that for other forms of external window shading systems, the window shading types in this embodiment can be deleted. The window shading types formed in this embodiment include no shading, single-pane glass shading, double-pane glass shading, shading only, single-pane glass shading combined shading, and double-pane glass shading combined shading. It should be noted that no shading means that air enters the room directly, double-pane glass shading means that one window overlaps with another window, and shading only means that only the sunshade blocks the empty window.

[0023] S2. Calculate the area occupied by each window shading type based on the shading ratio and the sliding distance of the movable window.

[0024] In this embodiment, as shown in the attached Figure 1 As shown, the upper fixed window occupies 1 / 4 of the outer window frame. The area and heat transfer coefficient corresponding to the shielding type will affect the indoor environmental quality. The area calculation method for windows of each shielding type is as follows: The exterior window shading system was quantified. The windows were divided into six sections based on the effects of window sliding and sunshade movement on the windows: 1. No shading; 2. Single-pane glazing; 3. Single-pane glazing, sunshading, and combined shading; 4. Sunshading only; 5. Double-pane glazing; 6. Double-pane glazing, sunshading, and combined shading. The area and corresponding heat transfer coefficient of each of the six sections are different, so the sunshade opening and closing ratio ( e )、Sunshade louver rotation angle( β ) and the distance the movable window moves ( l=l 1 +l 2 ) as the control variable, the area of ​​these six parts is quantified. Since 1 / 4 of the area above the window is fixed glass, when e When ≤0.25, w 1 = w 5 =0.75 l , w 2 =0.75·(1-2 l )+(0.25- e ), w 3 = e , w 4 = w 6 =0; when e >0.25 w 1 = w 5 = l (1-ε), w 2 =(1-2 l )·(1- e ), w 3 =(1-2 l )· e+ 0.25, w 4 = w 6 =( e- 0.25)· l .

[0025] S3. Obtain climate environment data, control instruction data, and time series data of the building main body based on the exterior window shading system under ideal operating conditions.

[0026] The test equipment selected for measurement is not limited to weather recorders, air quality measurements, sound level meters, and global radiation sensors, and the sampling method can be automatic sampling or manual sampling.

[0027] The ideal operating state refers to the state in which the building's composition and physical equipment remain unchanged. This refers to the functional layout, spatial divisions, and structural system, while the physical equipment remains unchanged, meaning the mechanical and electrical systems, facilities, and equipment installed within the building remain unchanged. This ideal operating state forms the foundation of building simulation software modeling and is used to evaluate the initial performance of design solutions.

[0028] The climate environment data includes indoor and outdoor temperature, indoor and outdoor illumination, indoor and outdoor noise, and indoor and outdoor carbon dioxide concentration; The control instruction data includes the opening ratio of the sunshade curtain, the rotation angle of the sunshade blinds and the sliding distance of the movable window; the parameters in the model derived from the control instruction data are control variables, and the control variables include the opening and closing ratio of the sunshade curtain, the rotation angle of the sunshade blinds and the moving distance of the movable window.

[0029] S4. Express the indoor environment model of the building as a discrete-time dynamic model: ; Where, Y Indicates the predicted value of indoor environmental quality; A library of candidate functions representing indoor environment models; represents a sparse matrix; Represents the set of functions selected from the candidate function library; X Represents building climate environment data; U Indicates control instruction data.

[0030] In the dynamic model, the candidate function library is composed of a series of nonlinear functions, and the candidate function library is established based on building-related physical knowledge and prior knowledge.

[0031] In this embodiment, the candidate function library includes constant terms, state variables, control variables, and thermal environment functions, light environment functions, sound environment functions, air quality functions, comfort functions, and trigonometric function terms represented by state variables and control variables, wherein the trigonometric function terms are used to represent the periodic changes of each variable.

[0032] In this embodiment, the constant term in the candidate function library represents the baseline impact of the constant influencing factor, and its value is 1. State variables are direct characteristic quantities of the indoor environment, including indoor temperature, outdoor temperature, indoor illuminance, outdoor illuminance, indoor carbon dioxide concentration, outdoor carbon dioxide concentration, indoor noise, and outdoor noise. Control variables include the sunshade opening / closing ratio, sunshade blind rotation angle, and movable window travel distance.

[0033] In this embodiment, the thermal environment function in the candidate function library is determined based on the physical relationship between the indoor temperature and the outdoor temperature under the control of the exterior window system. The physical relationship between the indoor temperature and the outdoor temperature is obtained based on the heat balance equation under steady-state conditions. The heat balance equation is as follows: ; Where, T in 、 T out Indicates indoor temperature and outdoor temperature respectively. U t represents the comprehensive heat transfer coefficient of the exterior window shading system, A Indicates the window area. I represents the solar radiation intensity, g t Indicates the comprehensive solar transmittance of the exterior window shading system. c p represents the specific heat capacity, r represents the air density, G represents the volume flow of gas through the window, G’ and Q’ They represent the wind and heat gained from external disturbances of the environment respectively; t Indicates time.

[0034] Calculate the comprehensive heat transfer coefficient of the exterior window shading system based on the area occupied by each window shading type: ; Where, Indicates the area blocked by a single piece of glass; It represents the area of ​​the area blocked by the composite sunshade of a single piece of glass; Indicates the area of ​​the area that is only shaded; Indicates the area blocked by double pane glass; It represents the area of ​​the double-pane glass sunshade composite shielding; 、 、 、 、 Respectively represent the heat transfer coefficient corresponding to each window shading type.

[0035] Calculate the comprehensive solar transmittance of the exterior window shading system based on the area occupied by each window shading type: ; Where, represents the area of ​​the unobstructed region, 、 、 、 、 、 Respectively represent the solar transmittance coefficient corresponding to each window shading type.

[0036] Calculate the gas volume flow through the window based on the window shading type of no shading and only shading: ; Where, v r Indicates wind speed, H 、 W Respectively represent the height and width of the outer window frame, Indicates the rotation angle of the sunshade louver; l Indicates the moving distance of the movable window.

[0037] In the sparse recognition algorithm, data-driven methods will automatically generate fixed physical parameters in the building environment through measured data. Therefore, the fixed physical parameters can be removed, and only the physical quantities related to the state variables and control variables can be extracted to construct a function sub-library. The fixed physical parameters include the height and width of the external window frame, the transmittance of visible light, the heat transfer coefficient, etc.

[0038] Remove the fixed physical parameters and extract the related items with the state variables and control variables to obtain the thermal environment function sub-library of the candidate function library: .

[0039] Where, Represents the thermal environment function sub-library of the candidate function library.

[0040] In this embodiment, the light environment functions in the candidate function library include light environment comfort evaluation and the physical relationship between indoor illuminance and outdoor illuminance under the control of the exterior window system; The evaluation model of light environment comfort is as follows: ; Where, L index Indicates the comfort level of the light environment; Extracted from the evaluation model of light environment comfort ln ( ln(E in ))、 ln 2 ( ln(E in )) as a candidate function; The physical relationship between indoor illuminance and outdoor illuminance is defined as the ratio of outdoor light passing through the exterior window shading system, as follows: ; Where, t tIt represents the comprehensive light transmittance of the exterior window shading system and can be quantified as: ; Where, 、 、 、 、 、 Respectively represent the light transmittance corresponding to each window shading type; After removing fixed physical parameters and extracting items related to state variables and control variables, we can obtain the light environment function sub-library of the candidate function library: ; Where, Represents the light environment function sub-library of the candidate function library.

[0041] In this embodiment, the acoustic environment function in the candidate function library is determined based on the physical relationship between indoor noise and outdoor noise, specifically as follows: ; Where, L’ represents the external noise of the environment, R s It represents the comprehensive sound insulation of the maintenance structure, which is characterized by the average sound transmission coefficient, as follows: ; Where, t s represents the average sound transmission coefficient, A w Indicates the wall area, t s,w is the wall sound transmission coefficient, t s,t is the sound transmission coefficient of the exterior window shading system, and t s,t The calculation formula is as follows: ; Where, R Indicates the sound insulation of glass; After extracting the items related to the state variables and control variables, the acoustic environment function sub-library of the candidate function library is obtained: ; Where, The acoustic environment function sub-library representing the candidate function library; In this embodiment, the air quality functions in the candidate function library include the physical relationship between air quality comfort and indoor carbon dioxide concentration and outdoor carbon dioxide concentration under the control of the external window shading system; The air quality comfort model is: ; Where, IAQ index Indicates air quality comfort; Extract in the above formula As a candidate function; In this embodiment, the physical relationship between the indoor carbon dioxide concentration and the outdoor carbon dioxide concentration under the control of the exterior window shading system is as follows: ; Where, V Indicates the volume of the room; In the above formula, after removing the related items of fixed physical parameter extraction, state variables and control variables, the air quality function sub-library of the candidate function library is obtained by combining the selected candidate functions: ; Where, Represents the air quality function sub-library.

[0042] In this embodiment, the comfort function in the candidate function library is determined based on the variable weight model of thermal environment comfort, light environment comfort, acoustic environment comfort, air quality comfort model, and indoor comprehensive environment comfort, wherein the thermal environment comfort model is as follows: ; Where, TC index represents the thermal environment comfort model; The light comfort model is as follows: ; Where, L index Represents the light environment comfort model; The acoustic environment comfort model is as follows: ; Where, AC index represents the acoustic environment comfort model; The air quality comfort model is as follows: ; Where, LAQ index represents the air quality comfort model; The variable weight model of indoor comprehensive environmental comfort is as follows: ; Where, IEQ index represents the comprehensive indoor environmental comfort function; m i (i =1, 2, 3, 4) represents the weight of the corresponding comfort model, and: ; Where, m 0 is the initial constant weight obtained after linear regression, and the values ​​of each item are as follows: m 1 0 =0.31, m 2 0 =0.12, m 3 0 =0.24, m 4 0 =0.33; a i is the upper limit of the comfort index prediction model, and the values ​​of each item are as follows: a 1 =23.75, a 2 =21.25, a 3 =25, a 4 =25; b i is a weight adjustment parameter, representing the penalty level, used to adjust the magnitude of the penalty interval weight change. The values ​​are as follows: b 1 =29.49, b 2 =14.74, b 3 =12.67, b 4 =9.68; x represents the calculated value of a single comfort level, After removing the related items of fixed physical parameter extraction, state variables and control variables from the above formula, the comfort function sub-library of the candidate function library is obtained: ; Where, Represents the comfort function sub-library; Trigonometric functions are often used to represent periodic changes in variables. For example, temperature and illumination vary regularly over a daily cycle with the rising and setting of the sun and the changing of seasons. Similarly, the height and angle of louvers vary regularly over the course of a day or night or seasonal cycle.

[0043] In this embodiment, the trigonometric function items in the candidate function library are: ; Where, Represents the set of trigonometric function items in the candidate function library; The constant terms, state variables, trigonometric function terms and each function sub-library of the candidate function library are integrated to obtain the candidate function library: .

[0044] The candidate function library in this embodiment covers the relevant physical knowledge of heat, light, sound and air quality in indoor environments. On the one hand, the function search space is greatly narrowed through the constraints of physical knowledge, so that the coefficient function does not need to traverse all possible nonlinear combinations, which greatly reduces the computational complexity. On the other hand, the embedding of physical knowledge gives each function a clear physical meaning, providing an explainable basis for subsequent model identification and optimization control.

[0045] S5. Create a sparse matrix The sparse identification expression of the sparse matrix is ​​represented by The coefficients of each function term in .

[0046] The sparse expression of a sparse matrix is ​​as follows: ; Where, Represents the coefficient of each function item in the candidate function library.

[0047] S6: Substitute the climate environment data, control instruction data and time series data obtained in step S3 into , based on the least squares method, the error between the measured value of indoor environmental quality and the predicted value of indoor environmental quality is minimized to calculate the optimal sparse matrix, and a penalty term is introduced to use regularization to penalize the sparse matrix.

[0048] The calculation method of sparse matrix is ​​as follows: ; Where, represents the sparse matrix solution, Show the general k The state variables and control variables at the moment are input into the dynamic model and calculated k+ The predicted value of indoor environmental quality at moment 1, Y k+1 express k+ The measured value of indoor environmental quality at moment 1, Represents the penalty coefficient.

[0049] For the above sparse matrix calculation formula: The data obtained from S3 is divided into training set and validation set according to the time series, and the The preset value set is extracted in turn. After substituting the preset value and the training set data into the above formula, the least squares method and l 1 Regularization calculation, get each The final sparse matrix corresponding to the preset values X f , determine the optimal penalty coefficient based on usage requirements And the sparse matrix solution corresponding to the coefficient X b .

[0050] In this embodiment, The preset value set is Set each of the preset values ​​in the Substitute the value into the sparse matrix calculation formula, and pass l 1 Regularization compresses unimportant coefficients to zero to eliminate redundant function items in the candidate function library, and finally obtains The sparse matrix corresponding to the values.

[0051] S7. Construct an indoor environmental quality prediction model for the building according to the optimal sparse matrix and candidate function library calculated in step S6.

[0052] According to the calculation results, the The final sparse matrix corresponding to the preset values X f , determine the optimal penalty coefficient based on usage requirements And the sparse matrix solution corresponding to the coefficient X b .

[0053] According to the use requirements, the coefficients of each function in the function database are controlled by penalty items, thereby controlling the accuracy and simplicity of the model. For example, if the subsequent optimization control of the building body requires a simple model, a model corresponding to a relatively high λ value can be selected, but the model accuracy is low. If the subsequent optimization control of the building body requires a high-precision model, a model corresponding to a relatively low λ value can be selected. In this case, more candidate function items are activated, and the model accuracy is high but relatively complex. In addition, in the indoor environment of the building body, when the interaction effect of indoor and outdoor environmental factors is significant, the system will have a collinearity problem. This application also adopts l 1 regularization to penalize sparse coefficients.

[0054] Anything not described in this application can be achieved by adopting or drawing on existing technologies.

[0055] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for establishing a physical enhancement model of an indoor environment based on the optimization control of an exterior window shading system. In this method, all windows in an exterior window frame and interior sunshades together constitute the exterior window shading system. The windows include fixed windows and movable windows. The method is characterized in that: This method specifically comprises the following steps: S1. Divide the window shading type of each area of ​​the exterior window system according to the shading ratio and the position of the movable window; The window shading types include no shading, single-pane glass shading, double-pane glass shading, sun shading only, single-pane glass sun shading and combined shading, and double-pane glass sun shading and combined shading; S2. Calculate the area occupied by each window shading type based on the shading ratio and the sliding distance of the movable window; S3. Obtaining climate environment data, control instruction data, and time series data of the building main body based on the exterior window shading system under ideal operating conditions; The ideal operating state is the operating state when the composition and physical equipment of the building remain unchanged; The climate environment data includes indoor and outdoor temperature, indoor and outdoor illumination, indoor and outdoor noise, and indoor and outdoor carbon dioxide concentration; The control instruction data includes the opening ratio of the sunshade curtain, the rotation angle of the sunshade blind and the sliding distance of the movable window; S4. Express the indoor environment model of the building as a discrete-time dynamic model: ； Where, Y Indicates the predicted value of indoor environmental quality; A library of candidate functions representing indoor environment models; represents a sparse matrix; Represents the set of functions selected from the candidate function library; X Represents building climate environment data; U Indicates control instruction data; S5. Create a sparse matrix The sparse identification expression of the sparse matrix is ​​represented by The coefficients of each function term in ; S6: Substitute the climate environment data, control instruction data and time series data obtained in step S3 into , based on the least squares method, the error between the measured value of indoor environmental quality and the predicted value of indoor environmental quality is minimized to calculate the optimal sparse matrix, and a penalty term is introduced to use regularization to penalize the sparse matrix; S7. Construct an indoor environmental quality prediction model for the building according to the optimal sparse matrix and candidate function library calculated in step S6.

2. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 1 is characterized in that: The candidate function library in step S4 includes constant terms, state variables, control variables, and thermal environment functions, light environment functions, sound environment functions, air quality functions, comfort functions, and trigonometric function terms represented by state variables and control variables, wherein the trigonometric function terms are used to represent the periodic changes of state variables and control variables.

3. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 2, characterized in that: The constant term in the candidate function library is used to characterize the baseline impact of the constant influencing factor, and its value is 1; The state variables are direct characteristic quantities of the indoor environment, including indoor temperature, outdoor temperature, indoor illumination, outdoor illumination, indoor carbon dioxide concentration, outdoor carbon dioxide concentration, indoor noise, and outdoor noise; The control variables include the sunshade curtain opening and closing ratio, the sunshade louver rotation angle, and the moving distance of the movable window.

4. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 3 is characterized in that: The thermal environment function in the candidate function library is determined based on the physical relationship between the indoor temperature and the outdoor temperature under the control of the external window system. The physical relationship between the indoor temperature and the outdoor temperature is obtained based on the heat balance equation under steady-state conditions. The heat balance equation is as follows: ； Where, T in 、 T out Indicates indoor temperature and outdoor temperature respectively. U t represents the comprehensive heat transfer coefficient of the exterior window shading system, A Indicates the window area. I represents the solar radiation intensity (W / m2), g t Indicates the comprehensive solar transmittance of the exterior window shading system. c p represents the specific heat capacity, ρ represents the air density, G represents the volume flow of gas through the window, G’ and Q’ They represent the wind and heat gained from external disturbances of the environment respectively; t Indicates time; Calculate the comprehensive heat transfer coefficient of the exterior window shading system based on the area occupied by each window shading type: ； Where, Indicates the area blocked by a single piece of glass; It represents the area of ​​the area blocked by the composite sunshade of a single piece of glass; Indicates the area of ​​the area that is only shaded; Indicates the area blocked by double pane glass; It represents the area of ​​the double-pane glass sunshade composite shielding; 、 、 、 、 Respectively represent the heat transfer coefficient corresponding to each window shading type; Calculate the comprehensive solar transmittance of the exterior window shading system based on the area occupied by each window shading type: ； Where, represents the area of ​​the unobstructed region, 、 、 、 、 、 Respectively represent the solar transmittance coefficient corresponding to each window shading type; Calculate the gas volume flow through the window based on the window shading type of no shading and only shading: ； Where, v r Indicates wind speed, H 、 W Respectively represent the height and width of the outer window frame, Indicates the rotation angle of the sunshade louver; l Indicates the moving distance of the movable window; Remove the fixed physical parameters and extract the related items with the state variables and control variables to obtain the thermal environment function sub-library of the candidate function library: ； Where, Represents the thermal environment function sub-library of the candidate function library.

5. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 4, characterized in that: The light environment functions in the candidate function library include light environment comfort evaluation and the physical relationship between indoor and outdoor illumination under the control of the external window system; The evaluation model of light environment comfort is as follows: ； Where, L index Indicates the comfort level of the light environment; Extracted from the evaluation model of light environment comfort ln ( ln(E in ))、 ln 2 ( ln(E in )) as a candidate function; The physical relationship between indoor illuminance and outdoor illuminance is defined as the ratio of outdoor light passing through the exterior window shading system, as follows: ； Where, τ t It represents the comprehensive light transmittance of the exterior window shading system and can be quantified as: ； Where, 、 、 、 、 、 Respectively represent the light transmittance corresponding to each window shading type; After removing fixed physical parameters and extracting items related to state variables and control variables, the light environment function sub-library of the candidate function library is obtained: ； Where, Represents the light environment function sub-library of the candidate function library.

6. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 5, characterized in that: The acoustic environment functions in the candidate function library are determined based on the physical relationship between indoor noise and outdoor noise, as follows: ； Where, L’ represents the external noise of the environment, R s It represents the comprehensive sound insulation of the maintenance structure, which is characterized by the average sound transmission coefficient, as follows: ； Where, τ s represents the average sound transmission coefficient, A w Indicates the wall area, τ s,w is the wall sound transmission coefficient, τ s,t is the sound transmission coefficient of the exterior window shading system, and τ s,t The calculation formula is as follows: ； Where, R Indicates the sound insulation of glass; After extracting the items related to the state variables and control variables, the acoustic environment function sub-library of the candidate function library is obtained: ； Where, The acoustic environment function sub-library representing the candidate function library; The air quality functions in the candidate function library include air quality comfort and the physical relationship between indoor and outdoor carbon dioxide concentrations under the control of the external window shading system; The air quality comfort model is: ； Where, IAQ index Indicates air quality comfort; Extract in the above formula As a candidate function; The physical relationship between indoor and outdoor carbon dioxide concentrations under the control of the exterior window shading system is as follows: ； Where, V Indicates the volume of the room; In the above formula, after removing the related items of fixed physical parameter extraction, state variables and control variables, the air quality function sub-library of the candidate function library is obtained by combining the selected candidate functions: ； Where, Represents the air quality function sub-library.

7. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 6, characterized in that: The comfort functions in the candidate function library are determined based on the variable weight model of thermal environment comfort, light environment comfort, acoustic environment comfort, air quality comfort model, and indoor comprehensive environment comfort. The thermal environment comfort model is as follows: ； Where, TC index represents the thermal environment comfort model; The light comfort model is as follows: ； Where, L index Represents the light environment comfort model; The acoustic environment comfort model is as follows: ； Where, AC index represents the acoustic environment comfort model; The air quality comfort model is as follows: ； Where, IAQ index represents the air quality comfort model; The variable weight model of indoor comprehensive environmental comfort is as follows: ； Where, IEQ index represents the comprehensive indoor environmental comfort function; m i ( i =1, 2, 3, 4) represents the weight of the corresponding comfort model, and: ； Where, x represents the calculated value of a single comfort level, m 0 is the initial constant weight obtained after linear regression, a i is the upper limit of the comfort index prediction model, b i is the weight adjustment parameter; After removing the related items of fixed physical parameter extraction, state variables and control variables from the above formula, the comfort function sub-library of the candidate function library is obtained: ； Where, Represents the comfort function sub-library; The trigonometric functions in the candidate function library are: ； Where, Represents the set of trigonometric function items in the candidate function library; The constant terms, state variables, trigonometric function terms and each function sub-library of the candidate function library are integrated to obtain the candidate function library: 。 8. The method for establishing a physical enhancement prediction model for indoor environment based on the optimization control of the exterior window shading system according to claim 7, characterized in that: The calculation method of the sparse matrix in step S6 is as follows: ； Where, represents the sparse matrix solution, Show the general k The state variables and control variables at the moment are input into the dynamic model and calculated k+ The predicted value of indoor environmental quality at moment 1, Y k+1 express k+ The measured value of indoor environmental quality at moment 1, represents the penalty term coefficient; For the above formula: The data obtained from S3 is divided into training set and validation set according to the time series, and the The preset value set is extracted in turn. After substituting the preset value and the training set data into the above formula, the least squares method and l 1 Regularization calculation, get each The final sparse matrix corresponding to the preset values Ξ f , determine the optimal penalty coefficient based on usage requirements And the optimal sparse matrix solution corresponding to the coefficient Ξ b .

Images