A method for vapor intrusion mechanism identification and probabilistic risk assessment based on bayesian inference

Abstract

This invention discloses a method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference, comprising the following steps: S1, acquiring basic data; S2, establishing an indoor radon mass balance model; S3, constructing the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model; S4, constructing the likelihood function between the measured indoor air radon concentration and the predicted indoor air radon concentration, and performing Bayesian inference based on the prior probability distribution and the likelihood function to obtain the posterior probability distributions of the radon diffusion intrusion attenuation coefficient and the radon convection intrusion attenuation coefficient; S5, obtaining the target volatile organic compound (VOC) diffusion intrusion attenuation coefficient; S6, obtaining the target VOC convection intrusion attenuation coefficient; S7, obtaining the posterior probability distribution of the concentration contributing to the target VOC vapor intrusion; S8, calculating the posterior probability distribution of the health risk of the target VOC vapor intrusion; S9, identifying the vapor intrusion driving mechanism.

Description

Technical Field

[0001] This invention relates to the field of risk assessment technology for contaminated sites, specifically to a method for identifying vapor intrusion mechanisms and probabilistic risk assessment based on Bayesian inference. Background Technology

[0002] Volatile organic pollutants (VOCs) in a site can migrate from contaminated soil, groundwater, or soil vapor beneath the building's foundation into the indoor air, creating exposure concentrations. This process is commonly referred to as vapor intrusion. Vapor intrusion is a significant exposure pathway in human health risk assessments of contaminated sites, particularly in underground spaces, underground garages, basements, sites with bedrock fissures, and buildings with underground interconnected structures such as utility shafts, elevator shafts, construction joints, and garage ramps. Vapor intrusion can have a significant impact on indoor air quality and human health risks.

[0003] Existing vapor intrusion risk assessment methods primarily employ the Johnson-Ettinger model, or combine it with deterministic parameters, empirical attenuation factors, or Monte Carlo simulations to predict indoor air pollutant concentrations and assess health risks. These methods typically simplify vapor intrusion as a general attenuation process from soil vapor to indoor air, or assume that pollutants mainly enter indoor air through edge cracks at the junctions of building walls and floors, driven by molecular diffusion. However, existing methods have the following drawbacks: First, vapor intrusion processes in actual sites often involve both diffusion and convection intrusion. Diffusion intrusion is mainly controlled by factors such as pollutant concentration gradient, concrete slab thickness, effective diffusion coefficient of the building envelope, and indoor ventilation rate; while convection intrusion is mainly controlled by factors such as indoor-outdoor pressure difference, underground gas flow, fissure connectivity, construction joints, pipeline shafts, elevator shafts, garage ramps, and other priority pathways. Since the controlling factors of diffusion and convection intrusion differ, the corresponding risk management and engineering remediation measures also differ. Traditional methods treat vapor intrusion as a single process, making it difficult to distinguish between diffusion and convection contributions, and thus difficult to identify the main controlling mechanism of vapor intrusion.

[0004] Second, in bedrock fissure sites, underground garages, basements, or complex underground interconnected spaces, underground gases may migrate via convection through priority channels such as fissures, construction joints, pipeline shafts, elevator shafts, and garage ramps, significantly increasing the intensity of vapor intrusion. In such scenarios, using only a single diffusion model may underestimate the actual risk of vapor intrusion; while directly using a conservative attenuation factor makes it difficult to explain the differences in vapor intrusion intensity under different building structures, monitoring locations, and underground interconnected conditions, and also makes it difficult to provide a mechanistic basis for subsequent risk management measures.

[0005] Third, traditional Monte Carlo simulations essentially propagate prior parameter distributions forward, failing to fully utilize measured indoor air concentrations to update parameter distributions in reverse. When key parameters such as ventilation rate, pressure difference, underground gas flow rate, and fissure area lack actual measurements, forward simulations struggle to reliably identify vapor intrusion mechanisms, and the calculated results may lack practical physical meaning, leading to high uncertainty in management decisions.

[0006] Fourth, while directly using the measured indoor concentration of target VOCs can assess the overall indoor air risk of a site, it is difficult to distinguish the risk caused by vapor intrusion, and it is also difficult to answer mechanistic questions such as the respective contributions of diffusion and convection, the significance of priority pathways, and how the intrusion risk will change after future changes in ventilation or pressure conditions. Furthermore, when the building is not yet completed, it is also impossible to obtain the measured indoor concentration of target VOCs.

[0007] Fifth, diffusion intrusion models rely on the effective diffusion coefficient of pollutants in the concrete slab or building envelope. For most volatile organic compounds (VOCs), there is very little data in the literature directly reporting their effective diffusion coefficients in concrete of different strengths, ages, and moisture contents, making it difficult to construct a stable prior distribution of parameters. In contrast, engineering testing and literature data on concrete carbonation depth and carbonation coefficient are much richer. The concrete carbonation process essentially involves the migration of carbon dioxide into the concrete interior and its reaction with alkaline components. The equivalent effective diffusion coefficient of carbon dioxide in concrete can be derived using the concrete carbonation depth or carbonation coefficient. Then, based on the ratio of gas phase molecular diffusion coefficients, the effective diffusion coefficient of radon or target VOCs in concrete can be derived, thus providing a basis for constructing a priori parameters for diffusion intrusion attenuation coefficients.

[0008] Furthermore, radon is an inert radioactive gas naturally present in soil gas. The process by which radon enters indoor air from the soil beneath the building floor is also controlled by factors such as the building floor, envelope, ventilation, indoor-outdoor pressure difference, fissures, and underground gas connectivity, exhibiting some similarities in migration pathways to the VOCs vapor intrusion process. Compared to target VOCs, radon monitoring is relatively easier, and data on radon concentrations in soil gas and indoor air are more readily available. Therefore, radon can serve as a natural tracer characterizing the ability of underground gases to enter indoor air in buildings, used to infer the actual vapor intrusion capacity and mechanisms of buildings. Summary of the Invention

[0009] The purpose of this invention is to provide a method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference, in order to solve the problems mentioned in the background art, such as difficulty in distinguishing vapor intrusion mechanisms, lack of key parameters, difficulty in identifying indoor sources, and high uncertainty in risk assessment results.

[0010] To achieve the above objectives, this invention provides a method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference, comprising the following steps: S1. Acquire radon soil gas-indoor air paired monitoring data, outdoor air radon concentration, building structural parameters, and the concentration, physicochemical parameters, and toxicity parameters of the target volatile organic compounds in the soil gas under the foundation. The radon soil gas-indoor air paired monitoring data includes the radon concentration in the soil gas under the foundation at multiple monitoring points and the indoor air radon concentration corresponding to each monitoring point. S2. Based on the radon concentration in the soil under the base plate, the indoor air radon concentration, the outdoor air radon concentration, and the building structural parameters, establish an indoor radon mass balance model, decomposing the indoor air radon concentration into the outdoor air input contribution concentration, the building material release contribution concentration, the diffusion intrusion contribution concentration, and the convection intrusion contribution concentration. S3. Based on literature data, field monitoring data, engineering experience, or concrete carbonation parameters, construct the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model. The parameters to be inverted include radon diffusion intrusion attenuation coefficient, radon convection intrusion attenuation coefficient, indoor and outdoor and soil radon concentration observation error parameters, indoor ventilation rate, radon release rate of building materials, and radon effective diffusion coefficient. S4. Using the measured concentration of indoor radon as an observation with error, construct the likelihood function between the measured concentration of indoor radon and the predicted concentration of indoor radon, and perform Bayesian inference based on the prior probability distribution and the likelihood function to obtain the posterior probability distribution of the radon diffusion intrusion attenuation coefficient and the radon convection intrusion attenuation coefficient. S5. Based on the ratio of the molecular diffusion coefficients of the target volatile organic compound to radon in the air, the radon diffusion intrusion attenuation coefficient is calibrated to obtain the target volatile organic compound diffusion intrusion attenuation coefficient. S6. Based on the radon convection intrusion attenuation coefficient and the convection intrusion calibration coefficient of the target volatile organic compound, obtain the target volatile organic compound convection intrusion attenuation coefficient; S7. Based on the concentration of the target volatile organic compounds in the soil gas under the substrate, the diffusion intrusion attenuation coefficient of the target volatile organic compounds, and the convective intrusion attenuation coefficient of the target volatile organic compounds, calculate the diffusion intrusion contribution concentration and the convective intrusion contribution concentration of the target volatile organic compounds, and obtain the posterior probability distribution of the vapor intrusion contribution concentration of the target volatile organic compounds. S8. Based on the contribution concentration and toxicity parameters of the target volatile organic compound vapor intrusion, calculate the posterior probability distribution of the health risk of the target volatile organic compound vapor intrusion. S9. Identify the vapor intrusion driving mechanism based on the ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration of the target volatile organic compounds.

[0011] In a preferred embodiment, in step S1, the building structural parameters include one or more of the following: indoor space area, indoor floor height, indoor volume, concrete base slab thickness, concrete enclosure structure thickness, exposed concrete surface area, ventilation rate, and underground space functional type. The underground space functional type includes basement, underground garage, pipeline shaft, elevator shaft, equipment room, underground connecting passage, and garage ramp.

[0012] In a preferred embodiment, in step S2, the indoor radon mass balance model is: ; in, This refers to the indoor radon concentration. Contributes to the concentration of outdoor air input. Contributes to the concentration of radon release from building materials. Contributes concentration to diffusion and invasion. The convective intrusion contribution concentration is used; the building material release contribution concentration, diffusion intrusion contribution concentration, and convective intrusion contribution concentration are determined according to the following formulas: The concentrations contributed by building material release, diffusion intrusion, and convective intrusion are determined according to the following formulas: ; ; ; ; ; In the formula, The radon release rate from the surface of concrete materials; The exposed surface area of ​​the concrete; V represents the indoor ventilation rate; V represents the indoor volume. Soil radon concentration; This is the radon diffusion intrusion attenuation coefficient; The radon convection intrusion attenuation coefficient; denoted as the effective diffusion coefficient of radon in the concrete slab and building envelope; L is the thickness of the concrete slab or building envelope; H is the interior floor height. The volumetric flow rate of soil gas entering the room, when When measured values ​​are lacking, As a latent variable to be inverted within a non-negative range, its posterior distribution is obtained through Bayesian inference.

[0013] In a preferred embodiment, in step S3, the carbonation equivalent effective diffusion coefficient of carbon dioxide in concrete is inverted based on the carbonation depth of concrete, and the effective diffusion coefficient of radon in concrete is derived based on the ratio of the molecular diffusion coefficients of radon and carbon dioxide in air; the derived effective diffusion coefficient of radon in concrete is used to construct the prior probability distribution of the effective diffusion coefficient of radon or the radon diffusion intrusion attenuation coefficient.

[0014] In a preferred embodiment, in step S3, the equivalent effective diffusion coefficient of carbon dioxide in concrete is calculated using the following formula: ; ; ; In the formula, For unit volume of concrete Integration ability; The carbonation coefficient of concrete; For concrete surface concentration; The depth of concrete carbonation; Carbonization time; The amount of cement or binder material used per unit volume of concrete; In cement or cementitious materials Quality score; This is a correction factor for the degree of carbonizability; Based on the ratio of the molecular diffusion coefficients of radon to carbon dioxide in air, the effective diffusion coefficient of radon in concrete is derived: ; in, The molecular diffusion coefficient of radon in air. is the molecular diffusion coefficient of carbon dioxide in air.

[0015] In a preferred embodiment, constructing the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model includes: expressing the building's indoor ventilation rate, radon release rate from building materials, equivalent effective diffusion coefficient of carbon dioxide carbonization, effective radon diffusion coefficient, and convective intrusion attenuation coefficient as probability distributions, wherein: .

[0016] In a preferred embodiment, step S4 involves constructing a likelihood function between the measured indoor radon concentration and the predicted indoor radon concentration, including: For the At each of the specified locations, the model predicts the indoor radon concentration as follows: ; in: ; Indoor radon concentration measured as an observation with error: When the indoor radon concentration follows a normal distribution, the normal likelihood function is used: ; When the indoor radon concentration follows a log-normal distribution or exhibits a right-skewed characteristic, the log-normal likelihood function is used: ; in, The measured concentration of radon in indoor air. To predict indoor radon concentration, or These are observation error parameters, characterizing instrumental errors, short-term fluctuations, spatial heterogeneity, and model errors. Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ; in, The measured concentration of radon in indoor air. To predict indoor radon concentration, or These are observation error parameters, characterizing instrumental errors, short-term fluctuations, spatial heterogeneity, and model errors. Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ; in, The set of parameters to be inverted. For monitoring data sets, For the prior probability distribution, Let be the likelihood function. Let be the posterior probability distribution.

[0017] In a preferred embodiment, in step S5, the diffusion intrusion attenuation coefficient of the target volatile organic compound is determined according to the following formula: ; in, For the first The diffusion and intrusion attenuation coefficient of the target volatile organic compounds. For the first The molecular diffusion coefficient of the target volatile organic compound in the air. is the molecular diffusion coefficient of radon in air.

[0018] In step S6, the attenuation coefficient of the target volatile organic compound convective intrusion is determined according to the following formula: ; in, For the first The attenuation coefficient of convective intrusion of the target volatile organic compounds. For the first The convective intrusion calibration coefficient for a target volatile organic compound is taken as follows, under equilibrium steady state and when the target volatile organic compound does not undergo significant adsorption, degradation, or reaction loss in the convective path: .

[0019] In a preferred embodiment, in step S7, the target volatile organic compound vapor intrusion contribution concentration is calculated according to the following formula: ; in: ; ; In the formula, For the first The concentration contributed by the intrusion of volatile organic compound vapors from the target species. For the first The concentration contributed by the diffusion and intrusion of target volatile organic compounds. For the first The concentration of volatile organic compounds contributed by the convective intrusion of the target species. For the first The concentration of the target volatile organic compounds in the soil atmosphere beneath the substrate, and the concentration contributed by vapor intrusion does not include the outdoor air background concentration, release from indoor decoration materials, vehicle exhaust, cleaning agents, paint or other non-subsurface sources; In step S8, the posterior probability distribution of the health risk from the intrusion of the target volatile organic compound vapor is calculated, including: When the target volatile organic compound is a carcinogenic pollutant, the carcinogenic risk of vapor intrusion is calculated using the following formula: ; When the target volatile organic compound is a non-carcinogenic pollutant, the vapor intrusion non-carcinogenic hazard quotient is calculated according to the following formula: ; in, For the first The carcinogenic risk of certain volatile organic compounds caused by vapor intrusion. For non-carcinogenic hazards caused by vapor intrusion, As a risk factor per unit of air inhalation, This is the inhalation reference concentration; Calculates and outputs the mean, median, 95% posterior confidence interval, and probability of exceeding the risk threshold based on posterior samples.

[0020] In a preferred embodiment, in step S9, the ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration is determined according to the following formula: ; when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is diffusion-driven vapor intrusion; when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is driven by both diffusion and convection. when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is either convection-driven or preferential channel-driven.

[0021] Compared with the prior art, the beneficial effects of the present invention are: First, existing vapor intrusion assessment methods typically treat vapor intrusion as a single, overall attenuation process, making it difficult to determine the main controlling mechanism of pollutants entering indoor air. This invention utilizes paired monitoring data of radon in soil air and indoor air to establish an indoor radon mass balance model. This model decomposes indoor radon concentration into contributions from outdoor air input, building material release, diffusion intrusion, and convection intrusion. Bayesian inference is used to invert the attenuation coefficients of diffusion and convection intrusion that satisfy physical constraints. Based on this, the similarity between radon and typical volatile organic compound (VOC) intrusion mechanisms is leveraged to further identify the contributions of target VOC diffusion and convection intrusion to indoor air pollution, providing a quantitative basis for diagnosing vapor intrusion mechanisms.

[0022] Second, traditional Monte Carlo simulations primarily involve forward propagation of prior parameter distributions, making it difficult to use measured indoor air data to correct the parameter distributions in reverse, which can easily lead to parameter combinations that do not match the actual situation on site. This invention uses measured indoor radon concentration as an observation with error, constructs a likelihood function, and updates parameters such as ventilation rate, radon release rate from building materials, radon diffusion intrusion attenuation coefficient, radon convection intrusion attenuation coefficient, and observation error posteriorly based on Bayesian inference. This makes the resulting attenuation coefficients and risk results more consistent with the actual intrusion status of the target building.

[0023] Third, in underground garages, basements, bedrock fissure sites, and buildings with underground interconnected structures such as pipeline shafts, elevator shafts, construction joints, and garage ramps, convective intrusion and preferential channel intrusion may be significantly enhanced, while parameters such as soil gas flow rate, pressure difference, and fissure connectivity are usually difficult to measure in practice. This invention uses the radon convective intrusion attenuation coefficient as a comprehensive parameter characterizing the entry of underground gas into the room through preferential channels, and incorporates it as a latent variable into a Bayesian model for inversion when measured airflow data is lacking, thereby improving the applicability of vapor intrusion assessment in complex underground spaces.

[0024] Fourth, existing methods, if they directly use the indoor air concentration of target VOCs, are easily affected by non-underground sources such as decoration materials, cleaning agents, paint, vehicle exhaust, and outdoor air background, making it difficult to accurately identify the contribution of vapor intrusion. This invention uses radon as a natural tracer to first invert the actual intrusion capacity of buildings to underground gases, and then obtains the diffusion intrusion attenuation coefficient and convection intrusion attenuation coefficient of target VOCs based on the ratio of the molecular diffusion coefficient of target VOCs to radon and the convection calibration relationship. Thus, even when the indoor concentration of target VOCs is lacking or the indoor sources are complex, the indoor air contribution concentration and health risk caused by underground vapor intrusion can still be calculated.

[0025] Fifth, for most VOCs, there is limited measured data on their effective diffusion coefficients under different concrete conditions, making it difficult to construct stable prior diffusion parameters using traditional methods. This invention utilizes the carbonation depth or carbonation coefficient of concrete to invert the carbonation-equivalent effective diffusion coefficient of carbon dioxide in concrete, and derives its effective diffusion coefficient in concrete based on the ratio of the molecular diffusion coefficients of radon or target VOCs to carbon dioxide, providing a parameter basis for the diffusion intrusion attenuation coefficient.

[0026] Sixth, this invention identifies the vapor intrusion driving mechanism by the ratio of convective intrusion contribution to diffusion intrusion contribution, so that the risk assessment results not only include the posterior probability distribution of concentration and health risk, but also can directly indicate whether subsequent treatment should focus on the bottom plate barrier sealing, or the control of priority channels such as cracks, pipeline wells, elevator shafts, and garage ramps. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below. All other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present invention. Example 1

[0029] like Figure 1As shown, the vapor intrusion mechanism identification and probabilistic risk assessment method based on Bayesian inference of the present invention includes the following steps: Step S1: Acquisition of basic data: Acquire radon soil gas-indoor air paired monitoring data, outdoor air radon concentration, building structural parameters, and the concentration, physicochemical parameters, toxicity parameters, and exposure parameters of target volatile organic compounds in the soil gas under the foundation. Among them, the radon soil gas-indoor air paired monitoring data includes the radon concentration in the soil gas under the foundation at multiple monitoring points and the corresponding indoor air radon concentration at each monitoring point.

[0030] Building structural parameters include one or more of the following: indoor space area, indoor floor height, indoor volume, concrete base slab thickness, concrete enclosure structure thickness, exposed concrete surface area, ventilation rate, and underground space function type. Underground space function types include basement, underground garage, pipeline shaft, elevator shaft, equipment room, underground connecting passage, and garage ramp.

[0031] Step S2: Establish an indoor radon mass balance model: Based on the radon concentration in the soil under the foundation, the indoor air radon concentration, the outdoor air radon concentration, and the building structural parameters, establish an indoor radon mass balance model, decomposing the indoor air radon concentration into the outdoor air input contribution concentration, the building material release contribution concentration, the diffusion intrusion contribution concentration, and the convection intrusion contribution concentration.

[0032] The indoor radon mass balance model is as follows: ; in, This refers to the indoor radon concentration. Contributes to the concentration of outdoor air input. Contributes to the concentration of radon release from building materials. Contributes concentration to diffusion and invasion. The convective intrusion contribution concentration is used; the building material release contribution concentration, diffusion intrusion contribution concentration, and convective intrusion contribution concentration are determined according to the following formulas: The concentrations contributed by building material release, diffusion intrusion, and convective intrusion are determined according to the following formulas: ; ; ; ; ; In the formula, The radon release rate from the surface of concrete materials; The exposed surface area of ​​the concrete; V represents the indoor ventilation rate; V represents the indoor volume. Soil radon concentration; This is the radon diffusion intrusion attenuation coefficient; The radon convection intrusion attenuation coefficient; denoted as the effective diffusion coefficient of radon in the concrete slab and building envelope; L is the thickness of the concrete slab or building envelope; H is the interior floor height. The volumetric flow rate of soil gas entering the room, when When measured values ​​are lacking, As a latent variable to be inverted within a non-negative range, its posterior distribution is obtained through Bayesian inference.

[0033] Step S3: Based on literature data, field monitoring data, engineering experience or concrete carbonation parameters, construct the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model. The parameters to be inverted include radon diffusion intrusion attenuation coefficient, radon convection intrusion attenuation coefficient, indoor and outdoor and soil radon concentration observation error parameters, indoor ventilation rate, radon release rate of building materials, and radon effective diffusion coefficient.

[0034] When measured data on the effective diffusion coefficient of radon or the target volatile organic compound in concrete are lacking, the carbonation equivalent effective diffusion coefficient of carbon dioxide in concrete is inverted based on the carbonation depth or carbonation coefficient of concrete. Furthermore, the effective diffusion coefficient of radon or the target volatile organic compound in concrete is derived based on the ratio of the molecular diffusion coefficients of radon or the target volatile organic compound to carbon dioxide in air. The derived effective diffusion coefficient of radon in concrete is used to construct the prior probability distribution of the effective diffusion coefficient of radon or the radon diffusion intrusion attenuation coefficient. Similarly, the derived effective diffusion coefficient of the target volatile organic compound in concrete is used to construct the prior probability distribution of the target volatile organic compound diffusion intrusion attenuation coefficient.

[0035] In step S3, the equivalent effective diffusion coefficient of carbon dioxide in concrete is calculated using the following formula: ; ; ; In the formula, For unit volume of concrete Combining capabilities, the unit is ; The carbonation coefficient of concrete, in units of ; For concrete surface Concentration, taking values ; This refers to the carbonation depth of concrete, expressed in meters (m). Carbonization time, in seconds; This refers to the amount of cement or binder material used per unit volume of concrete, and is taken as a value. ; In cement or cementitious materials Mass fraction, with a value of 0.64; This is a correction factor for the degree of carbonizability, with a value of 0.75.

[0036] Based on the ratio of the molecular diffusion coefficients of radon to carbon dioxide in air, the effective diffusion coefficient of radon in concrete is derived: ; in, The molecular diffusion coefficient of radon in air. is the molecular diffusion coefficient of carbon dioxide in air.

[0037] The prior probability distribution in step S3 includes one or more of the following: ; ; ; ; .

[0038] The prior probability distributions of indoor ventilation rate, radon release rate from building materials, and effective radon diffusion coefficient are determined by literature data, field monitoring data, engineering experience, or a combination thereof. Preferably, the convective intrusion attenuation coefficient is non-negative and less than 1, used to characterize the comprehensive ability of soil gas to enter the room through fissures, construction joints, or other preferred channels under pressure differential.

[0039] The prior probability distribution of the parameters to be inverted in the indoor radon mass balance model is as follows: the building indoor ventilation rate, radon release rate of building materials, equivalent effective diffusion coefficient of carbon dioxide carbonization, effective radon diffusion coefficient, and convective intrusion attenuation coefficient are all expressed as probability distributions, where: ; ; .

[0040] Step S4: Construct the likelihood function and perform Bayesian inference: Using the measured indoor radon concentration as the error-bearing observation, construct the likelihood function between the measured indoor radon concentration and the predicted indoor radon concentration, and perform Bayesian inference based on the prior probability distribution and the likelihood function to obtain the posterior probability distributions of the radon diffusion intrusion attenuation coefficient and the radon convection intrusion attenuation coefficient.

[0041] Construct a likelihood function between the measured concentration and the predicted concentration of indoor radon, including: For the At each of the specified locations, the model predicts the indoor radon concentration as follows: ; in: ; Indoor radon concentration measured as an observation with error: When the indoor radon concentration follows a normal distribution, the normal likelihood function is used: ; When the indoor radon concentration follows a log-normal distribution or exhibits a right-skewed characteristic, the log-normal likelihood function is used: ; in, The measured concentration of radon in indoor air. To predict indoor radon concentration, or These are observation error parameters, characterizing instrumental errors, short-term fluctuations, spatial heterogeneity, and model errors. Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ; in, The measured concentration of radon in indoor air. To predict indoor radon concentration, or These are observation error parameters, characterizing instrumental errors, short-term fluctuations, spatial heterogeneity, and model errors. Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ; in, The set of parameters to be inverted. For monitoring data sets, For the prior probability distribution, Let be the likelihood function. Let be the posterior probability distribution.

[0042] Step S5: Based on the ratio of the molecular diffusion coefficients of the target volatile organic compounds to radon in the air, calibrate the radon diffusion intrusion attenuation coefficient to obtain the target volatile organic compound diffusion intrusion attenuation coefficient.

[0043] Specifically, under the same building envelope, diffusion path, and steady-state diffusion conditions, the target volatile organic compound diffusion intrusion attenuation coefficient is determined according to the following formula: ; in, For the first The diffusion and intrusion attenuation coefficient of the target volatile organic compounds. For the first The molecular diffusion coefficient of the target volatile organic compound in the air. is the molecular diffusion coefficient of radon in air.

[0044] Step S6: Based on the radon convection intrusion attenuation coefficient and the target volatile organic compound (VOC) convection intrusion calibration coefficient, calibrate the radon convection intrusion attenuation coefficient to obtain the target VOC convection intrusion attenuation coefficient. The target VOC convection intrusion calibration coefficient is used to characterize the influence of adsorption, degradation, or reaction loss of the target VOC in the convection path.

[0045] Specifically, the attenuation coefficient of the convective intrusion of the target volatile organic compounds is determined according to the following formula: ; in, For the first The attenuation coefficient of convective intrusion of the target volatile organic compounds. For the first The convective intrusion calibration coefficient for a target volatile organic compound is taken as follows, under equilibrium steady state and when the target volatile organic compound does not undergo significant adsorption, degradation, or reaction loss in the convective path: .

[0046] Step S7: Based on the concentration of the target volatile organic compounds (VOCs) in the soil gas beneath the substrate, the diffusion intrusion attenuation coefficient of the target VOCs, and the convective intrusion attenuation coefficient of the target VOCs, calculate the diffusion intrusion contribution concentration and the convective intrusion contribution concentration of the target VOCs, respectively, and obtain the posterior probability distribution of the vapor intrusion contribution concentration of the target VOCs. The vapor intrusion contribution concentration of the target VOCs is calculated according to the following formula: ; in: ; ; In the formula, For the first The concentration contributed by the intrusion of volatile organic compound vapors from the target species. For the first The concentration contributed by the diffusion and intrusion of target volatile organic compounds. For the first The concentration of volatile organic compounds contributed by the convective intrusion of the target species. For the first The concentration of the target volatile organic compounds in the soil atmosphere beneath the substrate, and the concentration contributed by vapor intrusion does not include the outdoor air background concentration, release from indoor decoration materials, vehicle exhaust, cleaning agents, paint or other non-subsurface sources.

[0047] Step S8: Based on the contribution concentration and toxicity parameters of the target volatile organic compound vapor intrusion, calculate the posterior probability distribution of the health risk of the target volatile organic compound vapor intrusion.

[0048] When the target volatile organic compound is a carcinogenic pollutant, the carcinogenic risk of vapor intrusion is calculated using the following formula: ; When the target volatile organic compound is a non-carcinogenic pollutant, the vapor intrusion non-carcinogenic hazard quotient is calculated according to the following formula: ; in, For the first The carcinogenic risk of certain volatile organic compounds caused by vapor intrusion. For non-carcinogenic hazards caused by vapor intrusion, As a risk factor per unit of air inhalation, This is the inhalation reference concentration; Calculated based on posterior samples:

[0049] It outputs the mean, median, 95% posterior confidence interval, and probability of exceeding the risk threshold for vapor intrusion risk.

[0050] Step S9: Identify the vapor intrusion driving mechanism based on the ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration of the target volatile organic compound. The ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration is determined according to the following formula: ; when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is diffusion-driven vapor intrusion; when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is driven by both diffusion and convection. when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is either convection-driven or preferential channel-driven.

[0051] Example 2: This embodiment takes the underground space of a high-rise building that has been built but not yet put into use in a bedrock fracture site as the object, and describes the vapor intrusion mechanism identification and probabilistic risk assessment method based on Bayesian inference described in this invention, and compares it with the results of traditional Monte Carlo probabilistic model.

[0052] 1. Basic data acquisition: The study area is located in the foothills, covering an area of ​​approximately 6676 m² with a perimeter of approximately 436 m. It comprises four 27-31 story high-rise buildings with pile foundations, completed but not yet handed over, and the interiors are unfinished. The study area has two basement levels, each with a clear height of approximately 3.5 m. The underground spaces of the buildings are interconnected. The second basement level has a C30 concrete slab in direct contact with the foundation, with a thickness of 0.2-1.2 m and a P6 impermeability rating. It is surrounded by a 0.3 m thick C30 reinforced concrete retaining wall. The first basement level has one garage entrance and one exit, each 8 m wide, which will remain open during monitoring. The first and second basement levels are connected by a garage ramp. The strata in the study area are mainly interbedded sandstone and mudstone. Due to disturbance during construction leveling, the second basement level slab is subdivided into plain fill and moderately weathered bedrock. A total of 38 soil gas sampling points were set up under the foundation plate in the study area. The RAD7 radon (Rn) meter was used to measure the Rn concentration in the soil gas at a depth of approximately 0.5 m below the foundation plate, and the corresponding indoor air Rn concentration was simultaneously measured at a depth of approximately 0.5 m above each monitoring well. Additionally, 11 points were set up around the study area to measure the outdoor air Rn concentration. Furthermore, the benzene concentration in the soil gas monitoring wells at these 38 points was also measured according to specifications. The main parameters used in this embodiment are shown in Table 1. Table 1 Basic Data

[0053] 2. Establish an indoor radon mass balance model: In this embodiment, the indoor radon concentration is composed of outdoor air input, release from building materials, diffusion intrusion, and convective intrusion: ; The concentration contributed by building materials is: ; ; ; ; ; The indoor Rn concentration caused by vapor intrusion is the sum of the contributions from diffusion intrusion and convection intrusion: .

[0054] In this comparative analysis with the traditional Monte Carlo method, to avoid comparison bias caused by using different models, both the traditional Monte Carlo simulation and the Bayesian posterior simulation of this invention use the same indoor intrusion concentration model as described above. .

[0055] 3. Construct the prior probability distribution: The statistical distributions of ventilation rate, Rn leaching rate on concrete surface, Rn concentration in outdoor air, Rn concentration in indoor air, Rn concentration in soil gas under the foundation slab, and effective diffusion of Rn in concrete, as shown in Table 1, were used as prior inputs. For the convective intrusion attenuation coefficient, since there was a lack of measured values ​​for underground gas flow rate and pressure difference between indoor and foundation slab locations, it was included as a non-negative latent variable to be inverted in the Bayesian model (less than 1).

[0056] In traditional Monte Carlo simulations, all parameters are directly sampled randomly from the prior distributions listed in Table 1 and calculated forward. In this invention, the distributions listed in Table 1 are first used as prior distributions and then updated to posterior distributions by the likelihood constraints of the measured indoor Rn concentrations.

[0057] 4. Construct the likelihood function and perform Bayesian inference: Since the indoor air Rn concentration is positive and right-skewed in this embodiment, the log-normal likelihood function is used, and the measured indoor Rn concentration is used as the observation constraint for the model's predicted concentration. ; ; Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ; In the formula, Let D be the set of parameters to be inverted, representing the indoor air Rn, outdoor air Rn, building structural parameters, and related monitoring data. Through this step, if a certain set of prior parameters leads to a predicted indoor... If the concentration deviates significantly from the actual on-site concentration, its posterior weight decreases; conversely, parameter combinations that are consistent with the on-site mass balance relationship obtain higher posterior weights.

[0058] The posterior distribution functions and means of the radon diffusion intrusion attenuation coefficient and convective intrusion attenuation coefficient inferred using Bayesian inference are shown in Table 2.

[0059] Table 2. Posterior distribution results of radon diffusion intrusion attenuation coefficient and convective intrusion attenuation coefficient.

[0060] 5. Calculation of benzene diffusion and convection intrusion attenuation coefficients: Under steady-state equilibrium, the convective migration of benzene in a concrete slab can be considered as follows: ， Therefore, the average convective intrusion attenuation coefficient of benzene is: The posterior distribution is approximately: .

[0061] Based on the diffusion coefficients of benzene and radon in air, the mean benzene diffusion intrusion attenuation coefficient is calculated as follows: The posterior distribution is approximately: .

[0062] 6. Calculation of the concentration contribution of benzene vapor intrusion: In this embodiment, the concentration of benzene in the soil gas below the substrate follows the following order: Its arithmetic mean is: ; The average concentration contributed by benzene convection intrusion is: ; The average concentration contributed by benzene diffusion intrusion is: ; Therefore, the average concentration contributed by benzene vapor intrusion is: .

[0063] 7. Benzene vapor intrusion risk probability assessment: The average risk of cancer from benzene vapor intrusion is: .

[0064] Typical statistics for probabilistic risk assessment results are shown in Table 3. If we conservatively consider the 95th percentile risk value as the recipient's health risk, it exceeds an acceptable level. Nearly one order of magnitude.

[0065] Table 3. Posterior probability assessment results of benzene vapor intrusion risk

[0066] 8. Mechanism Identification This invention further decomposes the underground intrusion concentration into diffusion intrusion contribution and convective intrusion contribution. Based on Bayesian posterior simulation, in this embodiment: ; ; Construct the mean of the ratio of convective intrusion contribution to diffuse intrusion contribution: .

[0067] According to the mechanism identification rules of the present invention, when When, it is determined to be diffusion-driven; when When, it is determined to be driven by both diffusion and convection; when When this occurs, it is determined to be a convection-driven or priority channel-driven vapor intrusion. In this embodiment, benzene... The value is approximately 60.2, which is significantly greater than 10, therefore it is determined to be a convection-driven or preferential channel-driven vapor intrusion.

[0068] 9. Comparison of the unified model of traditional Monte Carlo simulation and the posterior simulation of this invention Based on the following intrusion concentration model: ; Traditional Monte Carlo simulations directly sample from the prior distribution for calculation. This invention, however, first updates the prior parameters using Bayesian methods with the measured indoor Rn concentration, and then calculates using the posterior distribution. The calculation results are shown in Table 4.

[0069] Table 4 Comparison of traditional Monte Carlo probability assessment with the calculation results of this invention

[0070] As can be seen, under the same indoor intrusion concentration model, the mean indoor Rn intrusion concentration obtained by traditional Monte Carlo simulation is about 27.5 Bq / m³, and the 95th percentile is about 69.6 Bq / m³; the mean indoor Rn intrusion concentration obtained by the Bayesian posterior simulation of this invention is about 42.1 Bq / m³, and the 95th percentile is about 48.0 Bq / m³.

[0071] The 95th percentile decreased by: (69.6-48.0) / 69.6×100%=31.0%; This indicates that the present invention can reduce the impact of extreme prior combinations that do not conform to the on-site quality balance on high-end prediction results, making the probabilistic simulation results more consistent with the measured constraints of the case site.

[0072] The simulation results of this invention are more consistent with the actual site conditions because in traditional Monte Carlo simulations, as long as the parameter combination is within the prior distribution range, it can be sampled and participated in the forward calculation. For example, when a combination of high concrete surface Rn exudation rate and low ventilation rate is sampled:

[0073] The contribution from building material release alone is: ; This value is close to or even exceeds most of the measured indoor Rn values ​​at this site, and it does not yet include contributions from outdoor air input, diffusion intrusion, and convective intrusion. Although this type of parameter combination falls within the a priori range mathematically, it does not match the indoor measured concentration constraints at this case site.

[0074] This invention constructs a likelihood function using the measured indoor Rn concentration. The aforementioned parameter combination receives a lower posterior weight due to the predicted concentration deviating from the measured concentration. Therefore, the 95th percentile of the Bayesian posterior simulation decreases from 69.6 Bq / m³ to 48.0 Bq / m³, indicating that the high-end quantile is no longer dominated by extreme prior combinations that do not conform to the physical meaning of the field.

[0075] This embodiment demonstrates that, under the premise of using the same indoor intrusion concentration model, the main difference between traditional Monte Carlo simulation and the Bayesian posterior simulation of this invention lies in whether the parameter distribution is constrained by the measured indoor Rn concentration. Traditional Monte Carlo simulation directly uses the prior distribution, obtaining an indoor Rn intrusion concentration mean of approximately 27.5 Bq / m³ and a 95th percentile of approximately 69.6 Bq / m³; this invention uses the measured indoor Rn concentration as a likelihood constraint, obtaining a posterior simulation mean of approximately 42.1 Bq / m³ and a 95th percentile of approximately 48.0 Bq / m³.

[0076] The 95th percentile of this invention is reduced by approximately 31.0% compared to traditional Monte Carlo methods, indicating that this invention can effectively reduce the impact of extreme prior parameter combinations that do not conform to the field mass balance relationship on high-end prediction results. Furthermore, this invention obtains... The value was approximately 60.2, clearly indicating that the site was a convection or priority channel driven vapor intrusion scenario, providing a more direct mechanism basis than traditional Monte Carlo methods for subsequent VOCs risk screening and priority channel control.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference, characterized in that: Includes the following steps: S1. Acquire radon soil gas-indoor air paired monitoring data, outdoor air radon concentration, building structural parameters, and the concentration, physicochemical parameters, and toxicity parameters of the target volatile organic compounds in the soil gas under the foundation. The radon soil gas-indoor air paired monitoring data includes the radon concentration in the soil gas under the foundation at multiple monitoring points and the indoor air radon concentration corresponding to each monitoring point. S2. Based on the radon concentration in the soil under the base plate, the indoor air radon concentration, the outdoor air radon concentration, and the building structural parameters, establish an indoor radon mass balance model, decomposing the indoor air radon concentration into the outdoor air input contribution concentration, the building material release contribution concentration, the diffusion intrusion contribution concentration, and the convection intrusion contribution concentration. S3. Based on literature data, field monitoring data, engineering experience, or concrete carbonation parameters, construct the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model. The parameters to be inverted include radon diffusion intrusion attenuation coefficient, radon convection intrusion attenuation coefficient, indoor and outdoor and soil radon concentration observation error parameters, indoor ventilation rate, radon release rate of building materials, and radon effective diffusion coefficient. S4. Using the measured concentration of indoor radon as an observation with error, construct the likelihood function between the measured concentration of indoor radon and the predicted concentration of indoor radon, and perform Bayesian inference based on the prior probability distribution and the likelihood function to obtain the posterior probability distribution of the radon diffusion intrusion attenuation coefficient and the radon convection intrusion attenuation coefficient. S5. Based on the ratio of the molecular diffusion coefficients of the target volatile organic compound to radon in the air, the radon diffusion intrusion attenuation coefficient is calibrated to obtain the target volatile organic compound diffusion intrusion attenuation coefficient. S6. Based on the radon convection intrusion attenuation coefficient and the convection intrusion calibration coefficient of the target volatile organic compound, obtain the target volatile organic compound convection intrusion attenuation coefficient; S7. Based on the concentration of the target volatile organic compounds in the soil gas under the substrate, the diffusion intrusion attenuation coefficient of the target volatile organic compounds, and the convective intrusion attenuation coefficient of the target volatile organic compounds, calculate the diffusion intrusion contribution concentration and the convective intrusion contribution concentration of the target volatile organic compounds, and obtain the posterior probability distribution of the vapor intrusion contribution concentration of the target volatile organic compounds. S8. Based on the concentration and toxicity parameters of the target volatile organic compound vapor intrusion, calculate the posterior probability distribution of the health risk of the target volatile organic compound vapor intrusion. S9. Identify the vapor intrusion driving mechanism based on the ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration of the target volatile organic compounds.

2. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference as described in claim 1, characterized in that: In step S1, the building structural parameters include one or more of the following: indoor space area, indoor floor height, indoor volume, concrete base slab thickness, concrete enclosure structure thickness, exposed concrete surface area, ventilation rate, and underground space function type. The underground space function type includes basement, underground garage, pipeline shaft, elevator shaft, equipment room, underground connecting passage, and garage ramp.

3. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 1, characterized in that: In step S2, the indoor radon mass balance model is as follows: ； in, This refers to the indoor radon concentration. Contributes to the concentration of outdoor air input. Contributes to the concentration of radon release from building materials. Contributes concentration to diffusion and invasion. The convective intrusion contribution concentration is used to determine the contribution concentrations of building material release, diffusion intrusion, and convective intrusion, respectively, according to the following formulas: ； ； ； ； ； In the formula, represents the radon release rate from the surface of the concrete material; The exposed surface area of ​​the concrete; V represents the indoor ventilation rate; V represents the indoor volume. Soil radon concentration; This is the radon diffusion intrusion attenuation coefficient; Radon convection intrusion attenuation coefficient; denoted as the effective diffusion coefficient of radon in the concrete slab and building envelope; L is the thickness of the concrete slab or building envelope; H is the interior floor height. The volumetric flow rate of soil gas entering the room, when When measured values ​​are lacking, As a latent variable to be inverted within a non-negative range, its posterior distribution is obtained through Bayesian inference.

4. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 1, characterized in that: In step S3, the carbonation equivalent effective diffusion coefficient of carbon dioxide in concrete is inverted based on the carbonation depth of concrete, and the effective diffusion coefficient of radon in concrete is derived based on the ratio of the molecular diffusion coefficients of radon and carbon dioxide in air; the derived effective diffusion coefficient of radon in concrete is used to construct the prior probability distribution of the effective diffusion coefficient of radon or the radon diffusion intrusion attenuation coefficient.

5. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 4, characterized in that: In step S3, the equivalent effective diffusion coefficient of carbon dioxide in concrete is calculated using the following formula: ； ； ； in, For unit volume of concrete Integration ability; The carbonation coefficient of concrete; For concrete surface concentration; The depth of concrete carbonation; Carbonization time; The amount of cement or binder material used per unit volume of concrete; In cement or cementitious materials Quality score; This is a correction factor for the degree of carbonizability; Based on the ratio of the molecular diffusion coefficients of radon to carbon dioxide in air, the effective diffusion coefficient of radon in concrete is derived: ； in, The molecular diffusion coefficient of radon in air. is the molecular diffusion coefficient of carbon dioxide in air.

6. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 5, characterized in that: Constructing the prior probability distribution of the parameters to be inverted in the indoor radon mass balance model includes: expressing the building's indoor ventilation rate, radon release rate from building materials, equivalent effective diffusion coefficient of carbon dioxide carbonization, effective radon diffusion coefficient, and convective intrusion attenuation coefficient as probability distributions, wherein: 。 7. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 6, characterized in that: In step S4, the likelihood function between the measured indoor radon concentration and the predicted indoor radon concentration is constructed, including: For the At each of the specified locations, the model predicts the indoor radon concentration as follows: ； in: ; Indoor radon concentration measured as an observation with error: When the indoor radon concentration follows a normal distribution, the normal likelihood function is used: ； When the indoor radon concentration follows a log-normal distribution or exhibits a right-skewed characteristic, the log-normal likelihood function is used: ； in, The measured concentration of radon in indoor air. To predict indoor radon concentration, or These are observation error parameters, characterizing instrumental errors, short-term fluctuations, spatial heterogeneity, and model errors. Bayesian inference is performed according to the following formula to obtain the posterior probability distribution of the parameters to be inverted: ； in, The set of parameters to be inverted. For monitoring data sets, For the prior probability distribution, Let be the likelihood function. Let be the posterior probability distribution.

8. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 7, characterized in that: In step S5, the diffusion intrusion attenuation coefficient of the target volatile organic compounds is determined according to the following formula: ； in, For the first The diffusion and intrusion attenuation coefficient of the target volatile organic compounds. For the first The molecular diffusion coefficient of the target volatile organic compound in the air. The molecular diffusion coefficient of radon in air; In step S6, the attenuation coefficient of the target volatile organic compound convective intrusion is determined according to the following formula: ； in, For the first The attenuation coefficient of convective intrusion of the target volatile organic compounds. For the first The convective intrusion calibration coefficient for a target volatile organic compound is taken as follows, under equilibrium steady state and when the target volatile organic compound does not undergo significant adsorption, degradation, or reaction loss in the convective path: .

9. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 8, characterized in that: In step S7, the concentration of volatile organic compound vapor intrusion contribution is calculated according to the following formula: ； in: ; ； In the formula, For the first The concentration contributed by the intrusion of volatile organic compound vapors from the target species. For the first The concentration contributed by the diffusion and intrusion of target volatile organic compounds. For the first The concentration of volatile organic compounds contributed by the convective intrusion of the target species. For the first The concentration of the target volatile organic compounds in the soil atmosphere beneath the substrate, and the concentration contributed by vapor intrusion does not include the outdoor air background concentration, release from indoor decoration materials, vehicle exhaust, cleaning agents, paint or other non-subsurface sources; In step S8, the posterior probability distribution of the health risk from the intrusion of the target volatile organic compound vapor is calculated, including: When the target volatile organic compound is a carcinogenic pollutant, the carcinogenic risk of vapor intrusion is calculated using the following formula: ； When the target volatile organic compound is a non-carcinogenic pollutant, the vapor intrusion non-carcinogenic hazard quotient is calculated according to the following formula: ； in, For the first The carcinogenic risk of certain volatile organic compounds caused by vapor intrusion. For non-carcinogenic hazards caused by vapor intrusion, As a risk factor per unit of air inhalation, This is the inhalation reference concentration; Calculates and outputs the mean, median, 95% posterior confidence interval, and probability of exceeding the risk threshold based on posterior samples.

10. The method for identifying vapor intrusion mechanisms and assessing probabilistic risks based on Bayesian inference according to claim 9, characterized in that: In step S9, the ratio of the convective intrusion contribution concentration to the diffusion intrusion contribution concentration is determined according to the following formula: ； when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is diffusion-driven vapor intrusion; when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is driven by both diffusion and convection. when At that time, the judgment of the first The vapor intrusion mechanism of the target volatile organic compounds is either convection-driven or preferential channel-driven.

Images