Automobile cabin air quality multi-pollutant health risk assessment method

Abstract

The application provides a kind of automobile cabin air quality multi-pollutant health risk evaluation method, and relates to the technical field of automobile air quality evaluation.The method comprises obtaining the concentration data of multiple target pollutants in the air in multiple vehicles under a predetermined test condition;The concentration data is used as an input feature to construct an input feature set;The concentration data is converted into a health risk endpoint according to toxicological reference parameters;The input feature set and the health risk endpoint form a training sample set, a mapping relationship between the input feature set and the health risk endpoint is established, and a health risk evaluation model is obtained;The contribution of the concentration data of each target pollutant to the predicted value of the health risk endpoint and the global contribution are calculated;The health risk weight of each target pollutant is obtained.The beneficial effects of the application are that the weight distribution takes the health risk endpoint as the direct target, so that the maximum weight can reflect the concentration level of the pollutant and its inherent toxicological hazards.

Description

Technical Field

[0001] This application relates to the field of automotive air quality assessment technology, and more specifically, to a method for assessing the health risks of multiple pollutants in automotive cabin air quality. Background Technology

[0002] The car cabin is a relatively enclosed microenvironment, and its interior materials may retain and continuously release various volatile organic compounds (VOCs) during the manufacturing process. Drivers and passengers exposed to this environment for extended periods during daily commutes may experience non-carcinogenic health effects such as irritation, while some pollutants may also pose a potential carcinogenic risk.

[0003] Currently, the industry mostly uses single pollutant limits or total volatile organic compound (TVOC) concentration indicators to evaluate in-vehicle air quality. While these methods are simple to operate, they are difficult to reflect the differences in comprehensive health risks under mixed exposure to multiple pollutants, resulting in insufficient resolution for comparing the health performance of different vehicles.

[0004] To improve comprehensive evaluation capabilities, existing technologies also employ weighted scoring schemes, primarily including subjective and objective weighting methods. Subjective weighting methods rely on expert experience, resulting in poor stability. Objective weighting methods, such as entropy weighting and principal component analysis, determine weights based on the statistical characteristics of concentration data. However, they essentially focus on the dispersion and information content of the data, rather than the differences in the health effects of the pollutants themselves, which can easily lead to highly toxic pollutants having their weights underestimated due to small concentration fluctuations.

[0005] Furthermore, the VOC concentration inside the vehicle is significantly affected by the test conditions, with marked differences in pollutant release characteristics under different temperatures and ventilation conditions. Existing methods still have shortcomings in the unified processing and fusion of data from multiple operating conditions, making it difficult to generate comprehensive evaluation results that take into account the characteristics of each operating condition. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a method for assessing the health risks of multiple pollutants in automotive cabin air quality, which is particularly suitable for health risk assessment of automotive cabin air quality, vehicle health performance classification, and determination of pollutant treatment priorities.

[0007] The technical solution adopted in this invention is: to provide a method for assessing the health risks of multiple pollutants in automotive cabin air quality, comprising the following steps: S10. Obtain concentration data of at least two target pollutants in the air inside multiple vehicles under preset test conditions, wherein the target pollutants are volatile organic compounds; S20. Using the concentration data of each of the target pollutants as input features, construct an input feature set; S30. Convert the concentration data into health risk endpoints based on toxicological reference parameters, wherein the health risk endpoints include non-carcinogenic risk indicator HI and / or carcinogenic risk indicator TCR. S40. The input feature set and the health risk endpoint constitute a training sample set; S50. Based on the training sample set, establish a mapping relationship from the input feature set to the health risk endpoint for each health risk endpoint to obtain a health risk assessment model; S60. Based on the health risk assessment model, calculate the contribution of the concentration data of each target pollutant to the predicted value of the health risk endpoint; S70. Take the absolute value of the contribution of each target pollutant in the training sample set and calculate the mean to obtain the global contribution of the target pollutant. S80. Normalize the global contribution of each target pollutant to obtain the health risk weight of each target pollutant.

[0008] Furthermore, the target pollutants include formaldehyde, acetaldehyde, benzene, toluene, xylene, ethylbenzene, styrene, trichloroethylene, and tetrachloroethylene.

[0009] Furthermore, the non-carcinogenic risk indicator HI is the sum of the hazard quotients HQ of each of the target pollutants, wherein the hazard quotient of the i-th target pollutant is... This is the ratio of the concentration of the target pollutant to its reference concentration RfC; The carcinogenic risk index TCR is the sum of the carcinogenic risks CR of each carcinogenic target pollutant, where the carcinogenic risk of the i-th carcinogenic target pollutant is... The concentration data of the target pollutant, converted from exposure parameters, is used to calculate its inhalation risk per unit (IUR).

[0010] Furthermore, the health risk assessment model includes linear regression model, regularized regression model, support vector regression model, and tree model ensemble regression model.

[0011] Furthermore, the tree model ensemble regression model includes a random forest regression model and a gradient boosting tree regression model.

[0012] Furthermore, the concentration data obtained in step S10 is obtained under at least two different test conditions, and the concentration data includes data subsets corresponding to each test condition; steps S40 to S70 are executed independently on the data subsets corresponding to each test condition to obtain at least two sets of condition-specific health risk weights corresponding to each test condition.

[0013] Furthermore, it also includes: weighting and summing the at least two sets of condition-specific health risk weights according to preset condition weight coefficients and then normalizing them to obtain a comprehensive health risk weight.

[0014] Furthermore, the test conditions include normal temperature conditions, high temperature conditions, and high temperature conditions with air conditioning on; the weighting coefficient of the test conditions is 4:3:3.

[0015] Furthermore, it also includes: determining the target pollutant with the largest contribution weight as the HI dominant pollutant and the TCR dominant pollutant based on the contribution weight of the pollutant corresponding to the non-carcinogenic risk indicator HI and the contribution weight of the pollutant corresponding to the carcinogenic risk indicator TCR, respectively; when the HI dominant pollutant and the TCR dominant pollutant are consistent, the vehicle sample is determined to be of the same dominant pollutant type; when the HI dominant pollutant and the TCR dominant pollutant are inconsistent, the vehicle sample is determined to be of the dual dominant pollutant type.

[0016] Furthermore, the contribution is the SHAP value.

[0017] The advantages and positive effects of this invention are as follows: by adopting the above technical solution, the weight allocation takes the health risk endpoint as the direct target, so that the final weight can simultaneously reflect the concentration level of pollutants and their inherent toxicological hazards; through machine learning models and interpretability algorithms, the objectivity and automation of weight allocation are realized, avoiding the human uncertainty of subjective weighting methods; the use of SHAP values ​​can ensure that the calculated pollutant contribution (i.e., weight) has better interpretability and stability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a health risk assessment method according to an embodiment of the present invention. Detailed Implementation

[0020] The present disclosure will now be described more fully with reference to the accompanying drawings, which illustrate exemplary embodiments of the present disclosure. The technical solutions of the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present disclosure, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present disclosure without creative effort are within the scope of protection of the present disclosure.

[0021] like Figure 1 As shown, the present invention provides a method for assessing the health risks of multiple pollutants in automotive cabin air quality, characterized by comprising the following steps: S10. Obtain concentration data of at least two target pollutants in the air inside multiple vehicles under preset test conditions. The target pollutants are volatile organic compounds. Specifically, in a standardized testing environment (such as a vehicle testing chamber), different driving scenarios are simulated (i.e., preset test conditions, such as normal temperature, high temperature exposure, and high temperature with air conditioning), and in-vehicle air samples are taken from multiple representative vehicles. Professional analytical equipment (such as gas chromatography-mass spectrometry) is used to accurately measure the concentrations of various volatile organic compounds (VOCs) in the air. Ultimately, a large database containing multiple vehicles, various operating conditions, and multiple pollutant concentrations is formed. The preset test conditions include at least one specific combination of environmental conditions, such as specific temperature, humidity, light intensity, and ventilation. The target pollutants are volatile organic compounds, and at least two types are included. Preferably, the target pollutants include at least two VOCs with reference concentration (RfC) parameters and / or inhalation unit risk (IUR) parameters to ensure that the non-carcinogenic risk indicator HI and / or carcinogenic risk indicator TCR and their corresponding pollutant contribution weights can be calculated.

[0022] S20. Construct an input feature set using the concentration data of each target pollutant as input features; Specifically, the test data of each vehicle obtained in step S10 is taken as a sample, and the concentration value of each target pollutant is taken as a feature of this sample. All the features of all samples are gathered together to form a structured data matrix, namely the "input feature set".

[0023] S30. Based on toxicological reference parameters, the concentration data are converted into health risk endpoints, which include non-carcinogenic risk indicators HI and / or carcinogenic risk indicators TCR. Specifically, using internationally recognized toxicological parameters (such as reference concentration (RfC) and inhalation unit risk (IUR), the concentration value of each pollutant is converted into its potential hazard value to the human body using a standard risk assessment formula. For pollutants that are not carcinogenic but cause irritation or other discomfort, their hazard quotients (HQ) are calculated and summed to obtain the total non-carcinogenic risk index (HI). For pollutants with carcinogenic risk, their carcinogenic risk (CR) is calculated and summed to obtain the total carcinogenic risk index (TCR).

[0024] S40. The training sample set consists of the input feature set and health risk endpoints; S50. Based on the training sample set, establish a mapping relationship from the input feature set to the health risk endpoint for each health risk endpoint to obtain a health risk assessment model. Specifically, the training sample set is input into a machine learning model (such as ridge regression, random forest, etc.) to allow the model to learn and master the complex mathematical relationship between "pollutant concentration combinations" and "health risk values". After training, inputting VOC concentration data into the model can predict the corresponding HI and TCR values.

[0025] S60. Based on the health risk assessment model, calculate the contribution of the concentration data of each target pollutant to the predicted value of the health risk endpoint. Specifically, the concentration data of each target pollutant acquired inside the vehicle is used to construct an input feature vector, which is then fed into a pre-trained health risk assessment model for forward inference, outputting predicted values ​​for health risk endpoints. Subsequently, the processor invokes a pre-defined model interpretation algorithm to quantify and analyze the marginal influence of each dimension (i.e., the concentration data of each target pollutant) in the input feature vector on the final predicted value of the health risk endpoint, thereby calculating the contribution of each target pollutant to the predicted value of the health risk endpoint. The contribution of the predicted value of the health risk endpoint characterizes the degree and direction of influence (i.e., positively increasing risk or negatively decreasing risk) of the concentration data of each target pollutant on the final predicted value of the health risk endpoint in a specific test sample.

[0026] S70. Take the absolute value of the contribution of each target pollutant in the training sample set and calculate the mean to obtain the global contribution of the target pollutant. Specifically, the contribution of the same pollutant (such as formaldehyde) is taken from all vehicle samples. First, its absolute value is calculated (focusing only on the magnitude of the impact, without distinguishing between positive and negative directions). Then, the average of these absolute values ​​is calculated to obtain the global contribution. The global contribution represents the average intensity of the target pollutant's impact on health risk across the entire study sample population.

[0027] S80. Normalize the global contribution of each target pollutant to obtain the health risk weight of each target pollutant.

[0028] Using the above method, the weight allocation takes the health risk endpoint as the direct target, so that the final weight can simultaneously reflect the concentration level of pollutants and their inherent toxicological hazards, and the evaluation results are more health-guiding. Through machine learning models and interpretable algorithms, the objectivity and automation of weight allocation are realized, avoiding the human uncertainty of subjective weighting methods.

[0029] In one embodiment, the target pollutants include formaldehyde, acetaldehyde, benzene, toluene, xylene, ethylbenzene, styrene, trichloroethylene, and tetrachloroethylene.

[0030] In one embodiment, the non-carcinogenic risk indicator HI is the sum of the hazard quotients HQ of each target pollutant, wherein the hazard quotient of the i-th target pollutant is... The ratio of the concentration of the target pollutant to its reference concentration (RfC); the carcinogenic risk index (TCR) is the sum of the carcinogenic risks (CR) of each carcinogenic target pollutant, where the carcinogenic risk of the i-th carcinogenic target pollutant is... The risk threshold value (IUR) is calculated from the concentration of the target pollutant and its inhalation risk per unit (IUR). The use of internationally accepted toxicological risk assessment formulas ensures the scientific rigor and authority of the health risk endpoint calculations, thus guaranteeing the accuracy of subsequent model training.

[0031] In one embodiment, the carcinogenic risk index TCR is the sum of the carcinogenic risks CR of each carcinogenic target pollutant, wherein the carcinogenic risk of the i-th carcinogenic target pollutant is... It is calculated from the average daily exposure dose (ADD) and its inhalation unit risk (IUR).

[0032] In one embodiment, the health risk assessment model includes a linear regression model, a regularized regression model, a support vector regression model, and a tree model ensemble regression model, wherein the tree model ensemble regression model includes a random forest regression model and a gradient boosting tree regression model.

[0033] In one embodiment, when the health risk assessment model is a linear model (such as a ridge regression model), the model contains weight coefficients corresponding to each input feature. By multiplying the concentration data of a target pollutant by its corresponding weight coefficient, the contribution of the target pollutant to the predicted value of the health risk endpoint can be calculated.

[0034] In one embodiment, when the health risk assessment model is a nonlinear model (such as a random forest model), a model interpretation method based on the Shapley value idea is used to perform feature attribution analysis on the model prediction results, and the change in the predicted value of the health risk endpoint is allocated to each input feature.

[0035] In one embodiment, the concentration data obtained in step S10 is obtained under at least two different test conditions, and the concentration data includes data subsets corresponding to each test condition; steps S40 to S70 are executed independently on the data subsets corresponding to each test condition to obtain at least two sets of condition-specific health risk weights corresponding to each test condition.

[0036] In one embodiment, the method further includes: weighting and summing at least two sets of work condition-specific health risk weights according to preset work condition weight coefficients and then normalizing them to obtain a comprehensive health risk weight.

[0037] In one embodiment, pollutant treatment priorities are generated based on comprehensive health risk weights.

[0038] In one embodiment, the test conditions include normal temperature conditions, high temperature conditions, and high temperature conditions with air conditioning on; the weighting coefficient for each condition is 4:3:3.

[0039] In one embodiment, the method further includes: determining the target pollutant with the largest contribution weight as the dominant pollutant for HI and the dominant pollutant for TCR based on the contribution weight of the pollutant corresponding to the non-carcinogenic risk indicator HI (the contribution weight of the pollutant is obtained by normalizing the contribution of the pollutant) and the contribution weight of the pollutant corresponding to the carcinogenic risk indicator TCR, respectively; when the dominant pollutant for HI and the dominant pollutant for TCR are the same, the vehicle sample is determined to be of the same dominant pollutant type; when the dominant pollutant for HI and the dominant pollutant for TCR are different, the vehicle sample is determined to be of the dual dominant pollutant type.

[0040] In one embodiment, a health risk score for vehicle samples is generated and output based on the health risk weights and concentration data of the target pollutants. The health risk weights of the target pollutants are derived from the global contribution of the pollutants calculated based on SHAP values ​​and normalized; therefore, the contribution weights for each target pollutant range from 0 to 1. The health risk score is obtained by weighting the contribution weights of the target pollutants at the corresponding health risk endpoints with the standardized concentration data of the target pollutants. Specifically, the HI health risk weight is used to calculate the non-carcinogenic health risk score, and the TCR health risk weight is used to calculate the carcinogenic health risk score. The health risk score is a dimensionless relative evaluation index used to compare the health risk levels among vehicle samples.

[0041] In one embodiment, the contribution is a feature-attribution-based contribution, which is either a SHAP value or a feature contribution calculated based on the Shapley value concept. Using SHAP values ​​ensures that the calculated pollutant contribution (i.e., weight) has better interpretability and stability.

[0042] This embodiment provides a preferred implementation method and describes in detail the weight allocation process based on a single test condition.

[0043] A sample of 515 vehicles was selected, and in-vehicle air samples were collected under three different testing conditions. The concentrations of nine target pollutants (VOCs) were measured. These nine VOCs included formaldehyde, acetaldehyde, benzene, toluene, xylene, ethylbenzene, styrene, trichloroethylene, and tetrachloroethylene. The sample sizes for the three testing conditions were as follows: 246 samples under normal temperature conditions, 153 samples under high temperature conditions, and 116 samples under high temperature conditions with air conditioning on. All collected data were divided into three independent subsets according to these three conditions to ensure that subsequent analysis reflected the risk characteristics under different conditions.

[0044] Based on the concentration data of each target pollutant and the preset occupant exposure parameters, the average daily exposure dose (ADD) of each target pollutant is calculated. The average daily exposure dose (ADD) is calculated by multiplying the concentration data of the target pollutant by the occupant's breathing rate, daily exposure time, exposure frequency, and exposure years, and then dividing by the product of the occupant's average weight and average exposure time. The reference concentration (RfC) and inhalation risk per unit (IUR) for each target pollutant were obtained from a pre-defined toxicological parameter library; the toxicological parameters were sourced from the Integrated Risk Information System (IRIS) database of the U.S. Environmental Protection Agency. For non-carcinogenic risk assessment, the average daily exposure dose (ADD) of each target pollutant is compared with the corresponding reference concentration (RfC) to obtain the hazard quotient (HQ) of each target pollutant. The non-carcinogenic risk indicator HI is obtained by summing the hazard quotients HQ of each target pollutant; For carcinogenic risk assessment, the average daily exposure dose (ADD), the corresponding inhalation unit risk (IUR), and the concentration unit conversion factor (CF) of each target pollutant are multiplied together to obtain the carcinogenic risk value (CR) of each target pollutant. Among them, the concentration unit conversion factor (CF) is used to realize the unit conversion between the average daily exposure dose (ADD) and the inhalation unit risk (IUR), and its value is 1000 to realize the unit conversion between mg / m³ and μg / m³. The total carcinogenic risk index (TCR) is obtained by summing the carcinogenic risk values ​​(CR) of each target pollutant. The non-carcinogenic risk indicator HI and the total carcinogenic risk indicator TCR are used as predicted values ​​for the health risk endpoint.

[0045] Using the concentration data of nine pollutants from all vehicle samples as the input feature set X, and the calculated HI and TCR values ​​as the output labels Y, a health risk assessment model was trained. The health risk assessment model employed a ridge regression model, and the trained model demonstrated extremely high fitting accuracy for both HI and TCR predictions under three operating conditions. (All values ​​are greater than 0.999), which proves that the model can stably and accurately establish the mapping relationship from pollutant concentration to health risk endpoints.

[0046] Based on the trained model, the SHAP algorithm was used to calculate the contribution of each pollutant concentration to the predicted HI and TCR values ​​in each subset of the data for each operating condition. By averaging the absolute values ​​of the contributions of all samples and normalizing the average, three sets of "operating condition-specific health risk weights" were obtained. According to the C-GCAP 2024 evaluation procedure, a 4:3:3 operating condition weight coefficient was used to weight and sum the three sets of operating condition-specific weights, and then normalized again to obtain a final set of "comprehensive health risk weights". The results show that at the non-carcinogenic risk HI endpoint, the comprehensive weights indicate that acetaldehyde (54.74%) and formaldehyde (31.34%) are the main risk contributors. At the carcinogenic risk TCR endpoint, the comprehensive weights show that formaldehyde (61.99%) is the primary risk contributor, followed by acetaldehyde (29.27%).

[0047] Based on the overall weighting, the priority of pollutant treatment can be directly output. In this embodiment, for HI risk, acetaldehyde and formaldehyde should be controlled first; for TCR risk, formaldehyde should be controlled first. By comparing the dominant pollutants in the HI and TCR weighting lists, the risk type of the vehicle can be diagnosed. In this embodiment, acetaldehyde is the dominant HI pollutant and formaldehyde is the dominant TCR pollutant, which can be diagnosed as "separation type of non-carcinogenic and carcinogenic dominant pollutants", providing precise guidance for differentiated pollution control.

[0048] To better utilize the multi-pollutant health risk assessment method for automotive cabin air quality, this application also provides a multi-pollutant health risk assessment device for automotive cabin air quality, comprising: The data acquisition module is used to receive and manage the raw concentration data collected from the test vehicle, and can classify and store the data according to the test conditions. The feature building module is used to extract the concentration values ​​of each target pollutant from the raw data and organize them into the set of input features required by the model. The risk endpoint calculation module has built-in toxicology calculation formulas and reference parameters, which are used to automatically calculate the non-carcinogenic risk indicator HI and the carcinogenic risk indicator TCR based on the input concentration data. The sample set construction module is used to pair the input feature set output by the feature construction module with the health risk endpoint values ​​output by the risk endpoint calculation module to construct a training sample set for model training. The model building module contains a variety of machine learning algorithm libraries, which are used to train health risk assessment models based on training sample sets and save the trained models. The contribution calculation module integrates model interpretability algorithms (such as SHAP) to load the trained model and calculate the contribution of each input feature to the model output. The global contribution calculation module is used to perform statistical analysis (take absolute value and calculate mean) on the contribution of all samples and calculate the global contribution of each pollutant. The weight generation module is used to normalize the global contribution and finally generate health risk weights; this module may also include a weight fusion unit for performing weighted fusion of weights for multiple operating conditions. The results output module is used to output the final health risk weights, pollutant treatment priorities, risk type diagnosis results, etc. in the form of visual reports or data interfaces.

[0049] Based on embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0050] An electronic device includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the multi-pollutant health risk assessment method for automotive cabin air quality provided in this disclosure.

[0051] Electronic devices are intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0052] A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the health risk assessment method for multiple pollutants in automotive cabin air quality provided in this disclosure.

[0053] The various embodiments of this disclosure can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0054] A computer program product includes a computer program / instructions, which are executed by a processor. This disclosure provides a method for assessing the health risks of multiple pollutants in automotive cabin air quality.

[0055] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0056] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0057] The embodiments of the present invention have been described in detail above, but the content described is only a preferred embodiment of the present invention and should not be considered as limiting the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the patent coverage of the present invention.

Claims

1. A method for assessing the health risks of multiple pollutants in automotive cabin air quality, characterized in that, Includes the following steps: S10. Obtain concentration data of at least two target pollutants in the air inside multiple vehicles under preset test conditions, wherein the target pollutants are volatile organic compounds; S20. Using the concentration data of each of the target pollutants as input features, construct an input feature set; S30. Convert the concentration data into health risk endpoints based on toxicological reference parameters, wherein the health risk endpoints include non-carcinogenic risk indicator HI and / or carcinogenic risk indicator TCR. S40. The input feature set and the health risk endpoint constitute a training sample set; S50. Based on the training sample set, establish a mapping relationship from the input feature set to the health risk endpoint for each health risk endpoint to obtain a health risk assessment model; S60. Based on the health risk assessment model, calculate the contribution of the concentration data of each target pollutant to the predicted value of the health risk endpoint; S70. Take the absolute value of the contribution of each target pollutant in the training sample set and calculate the mean to obtain the global contribution of the target pollutant. S80. Normalize the global contribution of each target pollutant to obtain the health risk weight of each target pollutant.

2. The method according to claim 1, characterized in that: The target pollutants include formaldehyde, acetaldehyde, benzene, toluene, xylene, ethylbenzene, styrene, trichloroethylene, and tetrachloroethylene.

3. The method according to claim 1, characterized in that: The non-carcinogenic risk indicator HI is the sum of the hazard quotients HQ of each of the target pollutants, where the hazard quotient of the i-th target pollutant is... This is the ratio of the concentration of the target pollutant to its reference concentration RfC; The carcinogenic risk index TCR is the sum of the carcinogenic risks CR of each of the target carcinogenic pollutants, where the carcinogenic risk of the i-th target carcinogenic pollutant is... It is calculated from the concentration data of the target pollutant and its inhalation risk per unit (IUR).

4. The method according to claim 1, characterized in that: The health risk assessment models include linear regression models, regularized regression models, support vector regression models, and tree-based integrated regression models.

5. The method according to claim 4, characterized in that: The tree model ensemble regression model includes a random forest regression model and a gradient boosting tree regression model.

6. The method according to claim 1, characterized in that: The concentration data obtained in step S10 is obtained under at least two different test conditions, and the concentration data includes data subsets corresponding to each test condition; steps S40 to S70 are executed independently on the data subsets corresponding to each test condition to obtain at least two sets of condition-specific health risk weights corresponding to each test condition.

7. The method according to claim 6, characterized in that, Also includes: The weights of the at least two sets of work-specific health risks are weighted and summed according to preset work-condition weight coefficients and then normalized to obtain the comprehensive health risk weights.

8. The method according to claim 7, characterized in that: The test conditions include normal temperature conditions, high temperature conditions, and high temperature conditions with air conditioning on; the weighting coefficient of the conditions is 4:3:

3.

9. The method according to claim 1, characterized in that, Also includes: Based on the contribution weight of the pollutant corresponding to the non-carcinogenic risk indicator HI and the contribution weight of the pollutant corresponding to the carcinogenic risk indicator TCR, the target pollutant with the largest contribution weight is determined as the HI dominant pollutant and the TCR dominant pollutant, respectively. When the HI dominant pollutant and the TCR dominant pollutant are the same, the vehicle sample is determined to be of the same dominant pollutant type. When the HI dominant pollutant and the TCR dominant pollutant are different, the vehicle sample is determined to be of the dual dominant pollutant type.

10. The method according to claim 1, characterized in that: The contribution is referred to as the SHAP value.

Images