A method and system for evaluating indoor environmental quality based on quantitative model

Abstract

The present invention discloses a method and system for evaluating indoor environmental quality based on a quantitative model, including obtaining indoor environmental quality data, preprocessing the indoor environmental quality data, inputting the indoor building characteristics into the indoor building scoring model to obtain a basic indoor environmental score, obtaining a first quality coefficient, a second quality coefficient and harmful characteristics, obtaining a first quality score and a second quality score, constructing an indoor environmental quality evaluation quantitative model according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and inputting the indoor environmental quality data to be evaluated into the indoor environmental quality evaluation quantitative model to obtain an indoor environmental quality score and evaluation. The method can not only improve the efficiency and accuracy of indoor environmental quality evaluation, but also has good interpretability, and can be directly applied to an indoor environmental quality evaluation system.

Description

Technical Field

[0001] The present invention relates to the field of quality evaluation, and in particular to a method and system for evaluating indoor environment quality based on a quantitative model. Background Art

[0002] Indoor environmental quality is a comprehensive evaluation indicator that covers all physical, chemical, biological and psychological factors in the indoor environment, and plays an important role in protecting human physical and mental health, improving comfort and work efficiency. Quality evaluation based on quantitative models is an analytical tool that applies mathematical statistics and machine learning to the field of quality evaluation. It collects, processes and analyzes data to evaluate the quality level of the target in a numerical way, providing a scientific basis for quality improvement and decision-making.

[0003] With the improvement of modern living standards, people's requirements for the quality of living and working environments are increasing. However, the traditional indoor environmental quality assessment method is characterized by relying on subjective feelings or a single indicator and lacking a comprehensive, objective and quantitative evaluation system, which also makes the evaluation results inaccurate and difficult to meet the needs of modern buildings for fine management of indoor environmental quality. With the rise of big data analysis and artificial intelligence algorithms, multi-source data fusion and quantitative models provide new ideas and methods for indoor environmental quality evaluation methods. By integrating indoor environmental data with indoor building data, a more comprehensive and accurate indoor environmental quality evaluation system is constructed. This method not only improves the objectivity and accuracy of the evaluation, but also provides a scientific basis for the optimization of the indoor environment. Therefore, it is of great significance to design an indoor environmental quality evaluation method and system based on a quantitative model to overcome the shortcomings of the existing evaluation model, to achieve a comprehensive and accurate evaluation of the indoor environment, and to provide a scientific basis for improving the indoor environment. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for evaluating indoor environmental quality based on a quantitative model.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Acquire indoor environment quality data and pre-process the indoor environment quality data; the indoor environment quality data includes indoor environment data and indoor building data;

[0008] Extracting features from the indoor building data to obtain indoor building features, constructing an indoor building scoring model, inputting the indoor building features into the indoor building scoring model to obtain a basic indoor environment score, and labeling the indoor environment basic score according to the indoor building data and the indoor environment data;

[0009] Classifying the indoor environmental data to obtain a harmful factor index, a sensory factor index, and an air factor index, and inputting the sensory factor index and the indoor building characteristics into a first quality function to obtain a first quality coefficient;

[0010] Inputting the air factor index into a second quality function to obtain a second quality coefficient, and determining a first quality score according to the second quality coefficient and the indoor environment basic score;

[0011] Screening the harmful factor index to obtain harmful characteristics, calculating a harmful damage coefficient according to the harmful characteristics, and determining a second quality score according to the harmful damage coefficient and the first quality score;

[0012] An indoor environment quality evaluation quantitative model is constructed according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and the indoor environment quality data to be evaluated is input into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation.

[0013] Furthermore, the method for obtaining the basic indoor environment score includes:

[0014] The principal component analysis method is used to extract and select indoor architectural data to obtain indoor architectural features, and the indoor architectural features are divided into training set, test set and evaluation set;

[0015] Constructing an indoor building scoring model, which includes an input layer, a base model layer, an integrated model layer and an output layer;

[0016] The training set enters the base model layer through the input layer. The base model layer outputs multiple prediction results to the integrated model layer. The base model layer merges the prediction results and outputs the basic indoor environment score through the output layer. The absolute error loss function is used to evaluate the difference between the prediction result and the true value. The gradient descent optimization algorithm is used to optimize the model parameters. The test set is used to evaluate the indoor building scoring model.

[0017] The evaluation set is input into the indoor building scoring model to obtain the basic score of the indoor environment.

[0018] Furthermore, the method for obtaining the first quality coefficient includes:

[0019] Extract and classify the indoor environmental data to obtain harmful factor index, sensory factor index and air factor index;

[0020] Indoor architectural features include window area A c 、Wall area A q 、Green plant rate ρ 1 , Floor quality score Q 1 And wall quality score Q2 , sensory factor indicators include light intensity I t , spectral deviation λ t and noise N t , substitute the indoor architectural characteristics and sensory factor index into the first quality function to obtain the first quality coefficient. The expression of the first quality function is:

[0021]

[0022] Where S 1 is the first quality coefficient, w 1 is the floor quality score weight coefficient, w 2 is the wall quality scoring weight coefficient, w 3 is the light intensity weight coefficient, w 4 is the spectral range weight coefficient, w 5 is the noise weight coefficient, I 0 is the appropriate light intensity, 0 For the appropriate spectral range, N 0 The first term represents the similarity between the indoor architectural feature vector x and the sensory factor index vector y, x = (x 1 , x 2 , x 3 ), y=(y 1 ,y 2 ,y 3 ), y 1 =ln(|I t -I 0 |+1),

[0023]

[0024] Furthermore, the method for determining the first quality score includes:

[0025] The air factor index is input into the second mass function to obtain the second mass coefficient. The expression of the second mass function is:

[0026]

[0027] Where S 2 is the second quality coefficient, k 1 is the oxygen concentration influence coefficient, k 2 is the carbon dioxide concentration influence coefficient, k 3 is the temperature influence coefficient, k 4 is the humidity influence coefficient, C o2 is the oxygen concentration, C co2 is the carbon dioxide concentration, c 1 is the baseline concentration of oxygen in the atmosphere, c2 is the baseline concentration of carbon dioxide in the atmosphere, r 1 is the oxygen response coefficient, r 2 is the response coefficient of carbon dioxide, C o ' 2 is the change in oxygen content, C c ' o2 is the change value of carbon dioxide content, ΔH is the humidity deviation, which indicates the deviation between the current air humidity and the ideal indoor humidity, and ΔT is the temperature deviation, which indicates the deviation between the current air temperature and the ideal indoor temperature;

[0028] The first quality score is determined according to the second quality coefficient and the basic score of the indoor environment, and the expression is:

[0029]

[0030] Where P 1 is the first quality score, P 0 is the basic score of indoor environment, k 5 It is the basic scoring weight coefficient of indoor environment.

[0031] Furthermore, the method of screening the harmful factor index to obtain the harmful characteristics includes:

[0032] The harmful factor indicators are screened to obtain harmful characteristics. The harmful factor indicators are screened using the equal volume hypersphere screening method. The data meets the following requirements:

[0033]

[0034] Where θ is the element of the array, s is the dimension of the point, is the optimized equivalent radius bounding coefficient, λ is the hypercube [-λ,λ] S Side length, n sel is the number of valid representative points that satisfy the above formula;

[0035] The effective representative point set deviation is calculated by the EVHS method, and the expression is:

[0036]

[0037] in is the effective representative point set deviation, is a set of valid representative points, The radius is Hypersphere The number of valid representative points in n is the total number of representative points. The equivalent radius is Hypersphere The volume, V(B(r c,s)) is the equivalent radius r c The hypersphere B(r c ,s) volume;

[0038] Searching for the smallest effective representative point set deviation obtains the effective representative point set with the best uniformity, and the expression is:

[0039]

[0040] in For the effective representative point set with the best uniformity, the data in the point set is classified to obtain harmful features.

[0041] Further, the method for determining the second quality score includes:

[0042] The harmful characteristics screened out include PM2.5, total volatile organic compounds TVOC and formaldehyde. The harmful loss coefficient is calculated based on the harmful characteristics and the health benchmark values ​​corresponding to the harmful characteristics. The expression is:

[0043]

[0044] Where S 3 is the harmful damage coefficient, p i is the damage factor corresponding to the i-th harmful feature, w n is the nth category weight corresponding to the i-th harmful feature, a i0 is the health benchmark value corresponding to the harmful characteristic, a i It is the harmful characteristic concentration;

[0045] The second quality score is determined according to the harmful loss coefficient, the first quality score and the basic score of the indoor environment. The expression is:

[0046] P 2 =P 1 -ln|P 0 ·S 3 |

[0047] Where P 2 is the second quality score, the second quality score is negatively correlated with the harmful loss coefficient, and the second quality score is positively correlated with the second quality coefficient.

[0048] Furthermore, the method for obtaining the indoor environment quality score includes:

[0049] The second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient form a comprehensive data set, and the comprehensive data set is divided into a training set and a test set;

[0050] Construct a quantitative model for indoor environmental quality evaluation, the specific structure of which includes input layer, base model layer, strategy layer and output layer;

[0051] The training set is input into the base model layer through the input layer. The base model layer connects the multi-layer perceptron model, support vector machine model and deep neural network model in parallel to make predictions and output the prediction results to the strategy layer. The strategy layer uses the stacking method to synthesize the final quantitative score and input it into the output layer. The two fully connected layers of the input layer generate quantitative scores and extract labels, and output the indoor environment quality score and evaluation.

[0052] The mean square error is used to calculate the difference between the predicted value and the true value, the adaptive moment estimation is used to adjust the model weight, and the test set is used to evaluate the model to output the quantitative model of indoor environmental quality evaluation;

[0053] The indoor environment quality data to be evaluated is input into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation.

[0054] In the second aspect, an indoor environment quality evaluation system based on a quantitative model includes:

[0055] Data acquisition module: used to obtain indoor environment quality data and pre-process the indoor environment quality data;

[0056] Data processing module: used for extracting features and classifying the indoor environment quality data, for obtaining the first quality coefficient, the second quality coefficient and the harmful damage coefficient, for determining the first quality score and the second quality score, and for transmitting data to the basic scoring module and the quality scoring module;

[0057] Basic scoring module: used to construct an indoor building scoring model according to indoor building characteristics to obtain a basic indoor environment score, and transmit the indoor environment basic score to the data module;

[0058] Evaluation model module: used for constructing an indoor environment quality evaluation quantitative model according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and inputting the indoor environment quality data to be evaluated into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation;

[0059] Intelligent supervision module: used to store, view and manage the indoor environment quality data, the indoor environment quality score and evaluation, and display the indoor environment quality score and evaluation to the user.

[0060] The beneficial effects of the present invention are:

[0061] The present invention is a method and system for evaluating indoor environmental quality based on a quantitative model. Compared with the prior art, the present invention has the following technical effects:

[0062] The present invention can improve the data preprocessing capability and enhance the model adaptability in indoor environmental quality evaluation through the steps of feature extraction and classification, data screening, obtaining the first quality coefficient, obtaining the second quality coefficient, calculating the harmful damage coefficient, and constructing a model, thereby improving the efficiency and accuracy of indoor environmental quality evaluation, optimizing the indoor environmental quality evaluation technology, greatly saving resources, and improving work efficiency, and can realize the evaluation of indoor environmental quality, providing a scientific basis for improving the indoor environment, which is of great significance to the indoor environmental quality evaluation, and can adapt to different indoor environmental quality evaluation systems based on quantitative models and terminal evaluation requirements of indoor environmental quality evaluation systems based on quantitative models of different users, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 The present invention is a flowchart of the steps of an indoor environment quality evaluation method based on a quantitative model. DETAILED DESCRIPTION

[0064] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0065] The present invention provides a method and system for evaluating indoor environmental quality based on a quantitative model, comprising the following steps:

[0066] like Figure 1 As shown, in this embodiment, the following steps are included:

[0067] Acquire indoor environment quality data and pre-process the indoor environment quality data; the indoor environment quality data includes indoor environment data and indoor building data;

[0068] Extracting features from the indoor building data to obtain indoor building features, constructing an indoor building scoring model, inputting the indoor building features into the indoor building scoring model to obtain a basic indoor environment score, and labeling the indoor environment basic score according to the indoor building data and the indoor environment data;

[0069] Classifying the indoor environmental data to obtain a harmful factor index, a sensory factor index, and an air factor index, and inputting the sensory factor index and the indoor building characteristics into a first quality function to obtain a first quality coefficient;

[0070] Inputting the air factor index into a second quality function to obtain a second quality coefficient, and determining a first quality score according to the second quality coefficient and the indoor environment basic score;

[0071] Screening the harmful factor index to obtain harmful characteristics, calculating a harmful damage coefficient according to the harmful characteristics, and determining a second quality score according to the harmful damage coefficient and the first quality score;

[0072] An indoor environment quality evaluation quantitative model is constructed according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and the indoor environment quality data to be evaluated is input into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation.

[0073] In this embodiment, the method for obtaining the basic indoor environment score includes:

[0074] The principal component analysis method is used to identify the important features in indoor building data, and the principal component with the largest variance is selected to obtain the indoor building features. The indoor building features are divided into training set, test set and evaluation set.

[0075] Constructing an indoor building scoring model, which includes an input layer, a base model layer, an integrated model layer and an output layer;

[0076] The training set enters the base model layer through the input layer, and the decision tree is used as the base model to predict the training set. The number of decision trees is 4. The prediction results enter the integrated model layer, and the majority voting strategy is used to merge the prediction results. The basic score of the indoor environment is output through the output layer;

[0077] The absolute error loss function is used to evaluate the difference between the predicted result and the true value, the gradient descent optimization algorithm is used to optimize the weights of the model parameters, and the test set is used to evaluate the indoor building scoring model;

[0078] Input the evaluation set into the indoor building scoring model to obtain the basic score of the indoor environment;

[0079] In the actual evaluation, taking the office of Company A and the residential area of ​​​​Community B as examples, the building data was randomly given and the feature extraction was obtained: Company A's office: window area 40m 2 、Wall area 500m 2 , greening rate 5%, floor quality score 8.5 points, wall quality score 7.8 points, residential area B: window area 60m 2 、Wall area 300m 2 , greening rate 10%, floor quality score 9 points, wall quality score 8.2;

[0080] The two sets of data were input into the indoor building scoring model to obtain the basic scores of the indoor environment: the basic score of the indoor environment of Company A's office was 36.5 points, and the basic score of the indoor environment of the residential area B was 43.4 points.

[0081] In this embodiment, the method for obtaining the first quality coefficient includes:

[0082] Factor analysis was used to extract the main factors in indoor environmental data, and cluster analysis was used to divide the data points into different groups according to the main factor scores to obtain harmful factor indicators, sensory factor indicators and air factor indicators;

[0083] Indoor architectural features include window area A c 、Wall area A q 、Green plant rate ρ 1 , Floor quality score Q 1 And wall quality score Q 2 , sensory factor indicators include light intensity I t , spectral deviation λ t and noise N t , substitute the indoor architectural characteristics and sensory factor index into the first quality function to obtain the first quality coefficient. The expression of the first quality function is:

[0084]

[0085] Where S 1 is the first quality coefficient, w 1 is the floor quality score weight coefficient, w 2 is the wall quality scoring weight coefficient, w 3 is the light intensity weight coefficient, w 4 is the spectral range weight coefficient, w 5 is the noise weight coefficient, I 0 is the appropriate light intensity, 0 For the appropriate spectral range, N 0 The first term represents the similarity between the indoor architectural feature vector x and the sensory factor index vector y, x = (x 1 , x 2 , x 3 ), y=(y 1 ,y 2 ,y 3 ), y 1 =ln(|I t -I 0 |+1),

[0086]

[0087] In the actual evaluation, the sensory factor indicators of the indoor environment of the office of Company A and the residential area of ​​​​the B community were randomly given: the office of Company A: light intensity 500lux, suitable light intensity 500lux, spectral deviation 10am, ideal spectrum 300-740nm, noise 45dB, ideal noise less than 50dB; the indoor environment of the residential area of ​​​​the B community: light intensity 300lux, suitable light intensity 400lux, spectral deviation 20nm, ideal spectrum 300-740nm, noise 35dB, ideal noise less than 70dB;

[0088] The sensory factor index and architectural features of the office of Company A and the residential room of Community B were substituted into the first quality function to calculate the similarity between the sensory factor index vector and the architectural feature vector, and the first quality coefficients of the office of Company A and the residential room of Community B were obtained to be 0.458 and 0.740 respectively.

[0089] In this embodiment, the method for determining the first quality score includes:

[0090] The air factor index is input into the second mass function to obtain the second mass coefficient. The expression of the second mass function is:

[0091]

[0092] Where S 2 is the second quality coefficient, k 1 is the oxygen concentration influence coefficient, k 2 is the carbon dioxide concentration influence coefficient, k 3 is the temperature influence coefficient, k 4 is the humidity influence coefficient, C o2 is the oxygen concentration, C co2 is the carbon dioxide concentration, c 1 The baseline concentration of oxygen in the atmosphere is 21%, c 2 The baseline concentration of carbon dioxide in the atmosphere is taken as 0.04%, r 1 is the oxygen response coefficient, r 2 is the response coefficient of carbon dioxide, C o ' 2 is the change in oxygen content, C c ' o2 is the change value of carbon dioxide content, ΔH is the humidity deviation, which indicates the deviation between the current air humidity and the ideal indoor humidity, and ΔT is the temperature deviation, which indicates the deviation between the current air temperature and the ideal indoor temperature;

[0093] The first quality score is determined according to the second quality coefficient and the basic score of the indoor environment, and the expression is:

[0094]

[0095] Where P 1 is the first quality score, P 0 is the basic score of indoor environment, k 5 is the basic scoring weight coefficient for indoor environment;

[0096] In the actual evaluation, the air factor indicators of the indoor environment of the office of Company A and the residential area of ​​​​Community B were randomly given: Company A's office: oxygen concentration 21%, carbon dioxide concentration 0.06%, temperature 22°C, humidity 48%; residential area of ​​​​Community B: oxygen concentration 20.9%, carbon dioxide concentration 0.08%, temperature 20°C, humidity 55%;

[0097] Substituting the air factor index into the second quality function, the second quality coefficients of the office of company A and the residential room of community B are 1.215 and 0.36 respectively. Based on the basic indoor environment scores of 36.5 points and 43.4 points for the office of company A and the residential room of community B, k 5 =6, and the first quality scores of the office of Company A and the residential room of Community B are determined to be 68.918 and 65.318 respectively.

[0098] In this embodiment, the method of screening the harmful factor index to obtain the harmful characteristics includes:

[0099] The harmful factor indicators are screened to obtain harmful characteristics. The harmful factor indicators are screened using the equal volume hypersphere screening method. The data meets the following requirements:

[0100]

[0101] Where θ is the element of the array, s is the dimension of the point, is the optimized equivalent radius bounding coefficient, λ is the hypercube [-α,α] S Side length, n sel is the number of valid representative points that satisfy the above formula;

[0102] The effective representative point set deviation is calculated by the EVHS method, and the expression is:

[0103]

[0104] in is the effective representative point set deviation, is a set of valid representative points, The radius is Hypersphere The number of valid representative points in n is the total number of representative points. The equivalent radius is Hypersphere The volume, V(B(r c ,s)) is the equivalent radius rc The hypersphere B(r c ,s) volume;

[0105] Searching for the smallest effective representative point set deviation obtains the effective representative point set with the best uniformity, and the expression is:

[0106]

[0107] in For the effective representative point set with the best uniformity, the data in the point set is classified to obtain harmful features.

[0108] In this embodiment, the method for determining the second quality score includes:

[0109] The harmful characteristics screened out include PM2.5, total volatile organic compounds TVOC and formaldehyde. The harmful loss coefficient is calculated based on the harmful characteristics and the health benchmark values ​​corresponding to the harmful characteristics. The expression is:

[0110]

[0111] Where S 3 is the harmful damage coefficient, p i is the damage factor corresponding to the i-th harmful feature, w n is the nth category weight corresponding to the i-th harmful feature, a i0 is the health benchmark value corresponding to the harmful characteristic, a i It is a harmful characteristic concentration;

[0112] The harmful loss coefficient is generally non-positive, and its absolute value increases exponentially with the increase of harmful characteristic concentration;

[0113] The second quality score is determined according to the harmful loss coefficient, the first quality score and the basic score of the indoor environment. The expression is:

[0114] P 2 =P 1 -ln|P 0 ·S 3 |

[0115] Where P 2 is the second quality score, the second quality score is negatively correlated with the harmful loss coefficient, and the second quality score is positively correlated with the second quality coefficient;

[0116] In the actual assessment, the harmful characteristic contents of the office of Company A and the residential area of ​​​​Community B were obtained through screening: Company A office: PM2.5: 45μg / m 3 、TVOC: 350μg / m 3 , Formaldehyde: 0.05mg / m 3Residential area B: PM2.5: 35 μg / m 3 、TVOC: 200μg / m 3 , Formaldehyde: 0.10mg / m 3 ; The corresponding health benchmark value for offices is: PM2.5≤35μg / m 3 、TVOC≤300μg / m 3 , Formaldehyde ≤ 0.1mg / m 3 ; Corresponding health benchmark value for residential areas: PM2.5≤70μg / m 3 、TVOC≤200μg / m 3 , Formaldehyde ≤ 0.08mg / m 3 ;

[0117] Substituting the harmful characteristics and corresponding health benchmark values ​​of Company A's office and Community B's residence into the calculation of the harmful loss coefficients is -0.37 and -0.1821. Based on the harmful loss coefficients, basic indoor environment scores and first quality scores of Company A's office and Community B's residence, the second quality scores are determined to be 66.323 and 63.344.

[0118] In this embodiment, the method for obtaining the indoor environment quality score includes:

[0119] The second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient form a comprehensive data set, and the comprehensive data set is divided into a training set and a test set;

[0120] Construct a quantitative model for indoor environmental quality evaluation, the specific structure of which includes input layer, base model layer, strategy layer and output layer;

[0121] The training set is input into the base model layer through the input layer. The base model layer is connected in parallel with the multi-layer perceptron model, the support vector machine model and the deep neural network model. The base models are trained respectively with the same training set. The performance of the base model is evaluated by cross-validation and the hyperparameters are adjusted to optimize the model. The base model outputs the prediction results to the strategy layer. The prediction results and the second quality score form a new data set to train the meta-model. The performance of the meta-model is evaluated by cross-validation and the hyperparameters are adjusted to optimize the model. The meta-model synthesizes the final quantitative score. The quantitative score enters the output layer. Two fully connected layers are set in the output layer. One fully connected layer generates the quantitative score, and the other fully connected layer is followed by a softmax activation function to extract feature labels, and the indoor environment quality score and evaluation are output.

[0122] The mean square error is used to calculate the difference between the predicted value and the true value, the adaptive moment estimation is used to adjust the model weight, and the test set is used to evaluate the model to output the quantitative model of indoor environmental quality evaluation;

[0123] Inputting the indoor environment quality data to be evaluated into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation;

[0124] In the actual evaluation, the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient are combined into a comprehensive data set. The model is trained by the training set, and the model is evaluated by the test set to obtain the indoor environmental quality evaluation quantitative model. The indoor environmental quality data of Company A's office and Community B's residence are input into the indoor environmental quality evaluation quantitative model to obtain the indoor environmental quality scores: The indoor environmental quality score of Company A's office is 83.315 points, and the overall indoor environmental quality is good. There are problems such as average indoor wall and floor quality, fewer green plants, dry air, and slightly exceeded PM2.5 and TVOC standards; the indoor environmental quality score of Community B's residence is 72.572 points, and the overall indoor environmental quality is average. There are problems such as small window area, poor ventilation, low room temperature, and slightly exceeded formaldehyde content.

[0125] In the second aspect, an indoor environment quality evaluation system based on a quantitative model includes:

[0126] Data acquisition module: used to obtain indoor environment quality data and pre-process the indoor environment quality data;

[0127] Data processing module: used for extracting features and classifying the indoor environment quality data, for obtaining the first quality coefficient, the second quality coefficient and the harmful damage coefficient, for determining the first quality score and the second quality score, and for transmitting data to the basic scoring module and the quality scoring module;

[0128] Basic scoring module: used to construct an indoor building scoring model according to indoor building characteristics to obtain a basic indoor environment score, and transmit the indoor environment basic score to the data module;

[0129] Evaluation model module: used for constructing an indoor environment quality evaluation quantitative model according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and inputting the indoor environment quality data to be evaluated into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation;

[0130] Intelligent supervision module: used to store, view and manage the indoor environment quality data, the indoor environment quality score and evaluation, and display the indoor environment quality score and evaluation to the user.

[0131] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for evaluating indoor environmental quality based on a quantitative model, characterized in that: The following steps are involved: S1. Acquire indoor environment quality data and pre-process the indoor environment quality data; the indoor environment quality data includes indoor environment data and indoor building data; S2, extracting features from the indoor building data to obtain indoor building features, constructing an indoor building scoring model, inputting the indoor building features into the indoor building scoring model to obtain a basic indoor environment score, and labeling the indoor environment basic score according to the indoor building data and the indoor environment data; S3, classifying the indoor environment data to obtain a harmful factor index, a sensory factor index, and an air factor index, and inputting the sensory factor index and the indoor building characteristics into a first quality function to obtain a first quality coefficient; S4, inputting the air factor index into a second quality function to obtain a second quality coefficient, and determining a first quality score according to the second quality coefficient and the indoor environment basic score; S5. Filter the harmful factor index to obtain harmful characteristics, calculate a harmful damage coefficient according to the harmful characteristics, and determine a second quality score according to the harmful damage coefficient and the first quality score; S6. Constructing an indoor environment quality evaluation quantitative model according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and inputting the indoor environment quality data to be evaluated into the indoor environment quality evaluation quantitative model to obtain an indoor environment quality score and evaluation; Obtaining the indoor environmental quality score and evaluation includes: The second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient form a comprehensive data set, and the comprehensive data set is divided into a training set and a test set; Construct a quantitative model for indoor environmental quality evaluation, the specific structure of which includes input layer, base model layer, strategy layer and output layer; The training set is input into the base model layer through the input layer, the base model layer connects the multi-layer perceptron model, the support vector machine model and the deep neural network model in parallel to perform prediction and output the prediction result to the strategy layer, the strategy layer uses the stacking method to synthesize the final quantitative score and input it into the output layer, the two fully connected layers of the input layer generate the quantitative score and extract the label, and output the indoor environment quality score and evaluation; The difference between the predicted value and the true value is calculated by using the mean square error, the model weight is adjusted by using the adaptive moment estimation, and the model is evaluated by using the test set to output the quantitative model for indoor environmental quality evaluation; The indoor environment quality data to be evaluated is input into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation.

2. The indoor environment quality evaluation method based on a quantitative model according to claim 1 is characterized in that: The method for obtaining the basic indoor environment score includes: The principal component analysis method is used to extract and select indoor architectural data to obtain indoor architectural features, and the indoor architectural features are divided into training set, test set and evaluation set; Constructing an indoor building scoring model, which includes an input layer, a base model layer, an integrated model layer and an output layer; The training set enters the base model layer through the input layer. The base model layer outputs multiple prediction results to the integrated model layer. The base model layer merges the prediction results and outputs the basic indoor environment score through the output layer. The absolute error loss function is used to evaluate the difference between the prediction result and the true value. The gradient descent optimization algorithm is used to optimize the model parameters. The test set is used to evaluate the indoor building scoring model. The evaluation set is input into the indoor building scoring model to obtain the basic score of the indoor environment.

3. The indoor environment quality evaluation method based on a quantitative model according to claim 1 is characterized in that: The method for obtaining the first quality coefficient comprises: Extract and classify the indoor environmental data to obtain harmful factor index, sensory factor index and air factor index; Indoor architectural features include window area A c 、Wall area A q , green plant rate ρ1, floor quality score Q1 and wall quality score Q2, sensory factor indicators include light intensity I t , spectral deviation λ t and noise N t , substitute the indoor architectural characteristics and sensory factor index into the first quality function to obtain the first quality coefficient. The expression of the first quality function is: Among them, S1 is the first quality coefficient, w1 is the floor quality score weight coefficient, w2 is the wall quality score weight coefficient, w3 is the light intensity weight coefficient, w4 is the spectral range weight coefficient, w5 is the noise weight coefficient, I0 is the appropriate light intensity, λ0 is the appropriate spectral range size, N0 is the appropriate noise, and the first item represents the similarity between the indoor building feature vector x and the sensory factor index y vector, x = (x1, x2, x3), 4. The indoor environment quality evaluation method based on a quantitative model according to claim 1 is characterized in that: The method of determining the first quality score comprises: The air factor index is input into the second mass function to obtain the second mass coefficient. The expression of the second mass function is: Where S2 is the second mass coefficient, k1 is the oxygen concentration influence coefficient, k2 is the carbon dioxide concentration influence coefficient, k3 is the temperature influence coefficient, k4 is the humidity influence coefficient, C o2 is the oxygen concentration, C co2 is the concentration of carbon dioxide, c1 is the reference concentration of oxygen in the atmosphere, c2 is the reference concentration of carbon dioxide in the atmosphere, r1 is the response coefficient of oxygen, r2 is the response coefficient of carbon dioxide, C′ o2 is the change in oxygen content, C′ co2 is the change value of carbon dioxide content, ΔH is the humidity deviation, which indicates the deviation between the current air humidity and the ideal indoor humidity, and ΔT is the temperature deviation, which indicates the deviation between the current air temperature and the ideal indoor temperature; The first quality score is determined according to the second quality coefficient and the basic score of the indoor environment, and the expression is: Among them, P1 is the first quality score, P0 is the basic score of the indoor environment, and k5 is the weight coefficient of the basic score of the indoor environment.

5. The indoor environment quality evaluation method based on a quantitative model according to claim 1 is characterized in that: The method of screening the harmful factor index to obtain harmful characteristics includes: The harmful factor indicators are screened to obtain harmful characteristics. The harmful factor indicators are screened using the equal volume hypersphere screening method. The data meets the following requirements: Where θ is the element of the array, s is the dimension of the point, is the optimized equivalent radius bounding coefficient, λ is the hypercube [-λ,λ] S Side length, n sel is the number of valid representative points that satisfy the above formula; The effective representative point set deviation is calculated by the EVHS method, and the expression is: in is the effective representative point set deviation, is a set of valid representative points, The radius is Hypersphere The number of valid representative points in n is the total number of representative points. The equivalent radius is Hypersphere The volume, V(B(r c ,s)) is the equivalent radius r c The hypersphere B(r c ,s) volume; Searching for the smallest effective representative point set deviation obtains the effective representative point set with the best uniformity, and the expression is: in For the effective representative point set with the best uniformity, the data in the point set is classified to obtain harmful features.

6. The indoor environment quality evaluation method based on a quantitative model according to claim 1 is characterized in that: The method of determining the second quality score comprises: The harmful characteristics screened out include PM2.5, total volatile organic compounds TVOC and formaldehyde. The harmful loss coefficient is calculated based on the harmful characteristics and the health benchmark values ​​corresponding to the harmful characteristics. The expression is: Where S3 is the harmful damage coefficient, p i is the damage factor corresponding to the i-th harmful feature, w n is the weight of the nth category corresponding to the i-th harmful feature, a i0 is the health benchmark value corresponding to the harmful characteristic, a i It is the harmful characteristic concentration; The second quality score is determined according to the harmful loss coefficient, the first quality score and the basic score of the indoor environment. The expression is: P2=P1-ln|P0·S3| Among them, P2 is the second quality score, the second quality score is negatively correlated with the harmful loss coefficient, and the second quality score is positively correlated with the second quality coefficient.

7. An indoor environment quality evaluation system based on a quantitative model, used to execute the method according to any one of claims 1 to 6, comprising: Data acquisition module: used to obtain indoor environment quality data and pre-process the indoor environment quality data; Data processing module: used for extracting features and classifying the indoor environment quality data, for obtaining the first quality coefficient, the second quality coefficient and the harmful damage coefficient, for determining the first quality score and the second quality score, and for transmitting data to the basic scoring module and the quality scoring module; Basic scoring module: used to construct an indoor building scoring model according to indoor building characteristics to obtain a basic indoor environment score, and transmit the indoor environment basic score to the data module; Evaluation model module: used for constructing an indoor environment quality evaluation quantitative model according to the second quality score, the first quality coefficient, the second quality coefficient and the harmful damage coefficient, and inputting the indoor environment quality data to be evaluated into the indoor environment quality evaluation quantitative model to obtain the indoor environment quality score and evaluation; Intelligent supervision module: used to store, view and manage the indoor environment quality data, the indoor environment quality score and evaluation, and display the indoor environment quality score and evaluation to the user.

Images