An Internet of Things-based Indoor Air Monitoring and Early Warning System and Method

Abstract

The application provides an indoor air monitoring and early warning system and method based on the Internet of Things, and relates to the technical field of indoor environment monitoring. The method comprises the following steps: acquiring air quality detection data corresponding to a plurality of data collection points respectively, extracting a plurality of instability characteristic parameters of each data collection point, and calculating a local state instability score of each data collection point with respect to each monitored gas; fusing a plurality of state instability scores to generate a global state instability score of a target monitoring area; calculating a dynamic diffusion coefficient of the target monitoring area with respect to gas component diffusion, determining a local abnormal area of the target monitoring area and constructing an abnormal state diffusion model, determining a pollution risk score of a plurality of core areas according to the abnormal state diffusion model, and generating air quality early warning information about the plurality of core areas according to the pollution risk score. The application realizes accurate monitoring of dynamic changes of air quality in a complex indoor environment and dynamic early warning.

Description

Technical Field

[0001] This invention relates to the field of indoor environmental monitoring technology, and in particular to an indoor air monitoring and early warning system and method based on the Internet of Things. Background Technology

[0002] With increasing attention being paid to indoor environmental health issues, IoT-based indoor air quality monitoring technology has been widely applied. Current monitoring systems typically deploy multiple sensors to collect data on the concentration of pollutants in indoor air, such as PM2.5, CO2, and VOCs, and combine this data with analysis methods to provide air quality assessment and early warning services.

[0003] Some methods rely on single-point pollutant concentration data, comparing it with preset thresholds to generate early warning information. This approach fails to extract information on the unstable state of pollutants within a region based on the potential anomalies of multi-point data. In practical applications, alarms are typically triggered only after pollution has already occurred, lacking the ability to capture early signs of instability. In complex environments with significant air quality fluctuations, this static analysis method cannot effectively capture the dynamic characteristics of pollutant concentration changes over time and space within the monitoring area, easily overlooking potential risks. Furthermore, indoor air quality is affected by multiple factors, including pollutant diffusion, ventilation conditions, and environmental layout. Point-to-point analysis fails to consider the global fusion of multi-point data, making it difficult to accurately reflect the overall state of the entire monitoring area. Summary of the Invention

[0004] To address the aforementioned technical issues, this invention proposes an Internet of Things-based indoor air monitoring and early warning system and method, which can effectively enhance the perception and prediction capabilities of pollution risks within the monitoring area, provide accurate and personalized air quality early warning services, and is suitable for complex indoor environmental monitoring scenarios.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] The first aspect of this invention provides an indoor air quality monitoring and early warning method based on the Internet of Things, comprising:

[0007] The system acquires air quality monitoring data from multiple data collection points, extracts multiple instability characteristic parameters for each data collection point for various monitored gases based on the air quality monitoring data, including volatility parameters and correlation parameters, and calculates the local state instability score for each data collection point for each monitored gas based on the multiple instability characteristic parameters.

[0008] Acquire location data for each data collection point, and fuse the state instability scores for each data collection point for each monitored gas based on the location data and air quality detection data to generate a global state instability score for the target monitoring area;

[0009] The dynamic diffusion coefficient of gas component diffusion in the target monitoring area is calculated based on the environmental condition data of the target monitoring area. Local anomaly areas in the target monitoring area are determined based on the global state instability score of the target monitoring area and the local state instability score of each data collection point for each monitored gas.

[0010] An abnormal state diffusion model for the target monitoring area is constructed based on air quality monitoring data and dynamic diffusion coefficients from data collection points within the local anomaly area. Based on the abnormal state diffusion model, pollution risk scores for multiple core areas are determined, and air quality early warning information for multiple core areas is generated based on the pollution risk scores.

[0011] Preferably, the state instability scores for each data acquisition point for each monitored gas are fused based on the location data and air quality detection data of the data acquisition points to generate a global state instability score for the target monitoring area, including:

[0012] Synchronization deviation analysis is performed on each data acquisition point based on air quality monitoring data. The time synchronization parameter between any two data acquisition points is calculated. The synchronization deviation parameter of each data acquisition point is calculated based on multiple time synchronization parameters. The state instability score of the data acquisition point for each monitored gas is corrected based on the synchronization deviation parameter, and the state instability correction score of the data acquisition point for each monitored gas is obtained.

[0013] Perform gas co-analysis for each data acquisition point, calculate the dynamic correlation parameters between any two monitoring gases at each data acquisition point, and calculate the co-change parameters of each data acquisition point with respect to each monitoring gas based on multiple dynamic correlation parameters;

[0014] Based on the location data of each data collection point and the corresponding multiple coordinated change parameters, multiple state instability correction scores are fused to generate a global state instability score for the target monitoring area, where: In the formula, I ns C is the global state instability score for the target monitoring area. i,j Let R be the cooperative variation parameter of the i-th data acquisition point with respect to the j-th monitored gas. i,j The score for state instability correction at the i-th data acquisition point with respect to the j-th monitored gas is given. Let be the location weight of the i-th data acquisition point, m be the total number of data acquisition points, and n be the number of categories of monitored gases.

[0015] Preferably, an abnormal state diffusion model of the target monitoring area is constructed based on air quality monitoring data and dynamic diffusion coefficients from data collection points within the local anomaly area, including:

[0016] The concentration parameters of each monitored gas in the local anomaly area are determined based on the air quality detection data of the data collection points. The initial intensity parameters of each monitored gas are obtained by correcting the concentration parameters based on the local state instability score corresponding to the monitored gas. A two-dimensional convection-diffusion equation for the local anomaly area is generated based on the dynamic diffusion coefficient and the initial intensity parameters of the data collection points. An anomaly state diffusion model for the target monitoring area is constructed based on the two-dimensional convection-diffusion equation for each local anomaly area in the target monitoring area.

[0017] For local anomaly areas in the target monitoring area, after detecting that the global state instability score of the target monitoring area is greater than the preset global instability threshold, multiple data collection points in the target monitoring area with local state instability scores greater than the preset global instability threshold are identified and recorded as anomaly nodes. Multiple anomaly nodes are merged based on the location data of the data collection points to determine multiple local anomaly areas.

[0018] Preferably, the dynamic diffusion coefficient of the gas component diffusion in the target monitoring area is calculated based on the environmental condition data of the target monitoring area, including:

[0019] The environmental condition data of the target monitoring area is fed into the diffusion coefficient analysis model, and the dynamic diffusion coefficient corresponding to the environmental condition data of the target monitoring area is generated through the diffusion coefficient analysis model.

[0020] For the diffusion coefficient analysis model, a training dataset is constructed by acquiring multiple sets of historical data from the target monitoring area. Each set of historical data contains a set of correlated environmental condition data and gas composition detection data. The target diffusion function is generated by discretizing the two-dimensional convection-diffusion equation in time and space. The target diffusion function is fitted to multiple sets of historical data to generate the target diffusion coefficient corresponding to each set of historical data. The environmental condition data in each set of historical data is correlated with the corresponding target diffusion coefficient to construct the training dataset. The diffusion coefficient analysis model is trained using the training dataset to obtain the trained diffusion coefficient analysis model.

[0021] Preferably, for the synchronous deviation analysis of data acquisition points and the coordinated analysis of monitored gases, the method further includes:

[0022] Based on the air quality monitoring data corresponding to the data collection points, the local synchronization parameters between the two data collection points for any one monitored gas are calculated using the following formula: In the formula, q k1，k2 (t) represents the local time synchronization parameter of data acquisition points k1 and k2 at time t with respect to a certain monitored gas. A k1 (t), A k2(t) represent the concentration values ​​of a certain monitored gas at data acquisition points k1 and k2 at time t, respectively. k1,k2 This represents the spatial distance between data collection point k1 and data collection point k2;

[0023] After calculating the local synchronization parameters for each monitored gas between any two data acquisition points using the above formula based on air quality monitoring data, the average of multiple local synchronization parameters is taken as the time synchronization parameter between the two data acquisition points. The average of multiple time synchronization parameters corresponding to each data acquisition point is calculated to obtain the synchronization deviation parameter of the data acquisition point.

[0024] Based on the air quality monitoring data corresponding to the data collection points, the time-series concentration vectors of each monitored gas at the data collection points are extracted. Correlation analysis is performed on any two monitored gases based on the time-series concentration vectors to calculate the dynamic correlation parameters between the two monitored gases. The average of the multiple dynamic correlation parameters corresponding to each monitored gas is calculated to obtain the co-variation parameters of each monitored gas.

[0025] Preferably, pollution risk scores for multiple core areas are determined based on an anomalous state diffusion model, including:

[0026] The reference coordinate data corresponding to each core area is determined, and the reference coordinate data corresponding to each core area is analyzed through an abnormal state diffusion model. This includes inputting the reference coordinate data corresponding to each core area into multiple two-dimensional convection-diffusion equations, processing the two-dimensional convection-diffusion equations through the finite difference method, obtaining the concentration prediction data corresponding to the reference coordinate data at a preset time, and determining the core area based on the concentration prediction data to obtain a pollution risk score.

[0027] Preferably, multiple instability characteristic parameters corresponding to various monitored gases for each data collection point are extracted based on air quality monitoring data, including:

[0028] The mean and standard deviation of the time-series concentration vector for each monitored gas are calculated. The ratio of the standard deviation to the mean of the monitored gas is used as the fluctuation parameter of the monitored gas. Autocorrelation analysis is performed on the time-series concentration vector of each monitored gas to calculate the correlation parameter of the monitored gas.

[0029] A second aspect of the present invention provides an Internet of Things (IoT)-based indoor air quality monitoring and early warning system for implementing the aforementioned IoT-based indoor air quality monitoring and early warning method, comprising:

[0030] The local state analysis module is used to acquire air quality detection data corresponding to multiple data collection points, extract multiple instability characteristic parameters for each data collection point for various monitored gases based on the air quality detection data, including volatility parameters and correlation parameters, and calculate the local state instability score for each data collection point for each monitored gas based on the multiple instability characteristic parameters.

[0031] The global state instability analysis module is used to acquire the location data of each data acquisition point, and to fuse the state instability scores of each data acquisition point for each monitored gas based on the location data and air quality detection data to generate a global state instability score for the target monitoring area.

[0032] The anomaly area analysis module is used to calculate the dynamic diffusion coefficient of gas component diffusion in the target monitoring area based on the environmental condition data of the target monitoring area, and to determine the local anomaly areas of the target monitoring area based on the global state instability score of the target monitoring area and the local state instability score of each data acquisition point for each monitored gas.

[0033] The air monitoring and early warning module is used to construct an abnormal state diffusion model of the target monitoring area based on air quality detection data and dynamic diffusion coefficients from data collection points in the local abnormal area, determine the pollution risk score of multiple core areas based on the abnormal state diffusion model, and generate air quality early warning information for multiple core areas based on the pollution risk score.

[0034] The present invention has the following beneficial effects:

[0035] This invention analyzes air quality monitoring data collected by an IoT system, extracts instability features and calculates local state scores. It combines the temporal synchronization of data collection points with the synergistic relationship of monitored gases to generate a global state instability score for the target monitoring area. Based on environmental condition data, it dynamically generates a diffusion coefficient and constructs an abnormal state diffusion model by combining the instability features of abnormal areas. The model predicts future pollution risk scores for core areas, generating personalized air quality early warning information. This invention can accurately monitor dynamic changes in air quality in complex indoor environments, improve the accuracy of anomaly location and early warning, effectively reduce monitoring errors and false alarms, and achieve precise monitoring and early warning of air quality within the region. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating an indoor air monitoring and early warning method based on the Internet of Things, as provided in one embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram of an indoor air monitoring and early warning system based on the Internet of Things, provided as one embodiment of the present invention. Detailed Implementation

[0038] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention.

[0039] Please see Figure 1 The diagram illustrates a flowchart of an indoor air monitoring and early warning method based on the Internet of Things (IoT) according to one embodiment of the present invention. The method specifically includes the following steps:

[0040] Step S10: Obtain air quality detection data corresponding to multiple data collection points, extract multiple instability characteristic parameters for each data collection point for each type of monitored gas based on the air quality detection data, and calculate the local state instability score for each data collection point for each type of monitored gas based on the multiple instability characteristic parameters.

[0041] In this embodiment, the air quality monitoring data corresponding to the data collection points includes concentration change data of various monitored gases such as PM2.5, CO2, VOCs, and formaldehyde over a period of time, collected through an IoT system. The data collection points are equipped with various sensor devices, enabling real-time monitoring and collection of gas data from different areas, and data integration and analysis through the IoT system. Depending on different indoor environments, such as industrial parks, factory buildings, and office buildings, a corresponding monitoring cycle can be reasonably set according to actual needs, for example, performing periodic data analysis every 15 minutes. For the data from each data collection point within a certain time period, multiple instability characteristic parameters for various monitored gases are extracted to describe the critical instability characteristics of different data collection points within the target monitoring area, i.e., the potential abnormal characteristics before the occurrence of abnormal air quality events. Taking volatility parameters and correlation parameters as examples, the mean and standard deviation of the time-series concentration vector for each monitored gas are calculated. The ratio of the standard deviation to the mean of the monitored gas is used as the volatility parameter of the monitored gas. Autocorrelation analysis is performed on the time-series concentration vector of each monitored gas, and the correlation parameter of the monitored gas is calculated using the autocorrelation coefficient formula. Then, multiple instability characteristic parameters of the monitored gas are used to comprehensively calculate the local instability score for each data acquisition point for each monitored gas. The instability score can be defined as a weighted combination of instability characteristic parameters, reflecting the degree of air quality anomaly in the local area corresponding to the data acquisition point during the current time period.

[0042] Step S20: Obtain the location data of each data collection point, and fuse the state instability scores of each data collection point for each monitored gas based on the location data of the data collection points and the air quality detection data to generate a global state instability score for the target monitoring area.

[0043] In this embodiment, after determining the air quality performance of different local areas within the target monitoring area, multiple local instability scores corresponding to multiple data collection points are further fused. During the fusion process, the influence of the spatial distribution of different data collection points is considered, as well as the potential data lag or synchronization differences in sensor data at different locations. That is, the temporal-spatial synchronization of the signal is comprehensively considered to reflect the diffusion dynamics of pollutants. If the synchronization suddenly decreases, it may indicate anomalies in a local area, such as the sudden activation of pollution sources. This achieves efficient fusion of the characteristics of different data collection points, which can be better applied to regional air pollution monitoring and early warning in complex indoor environments. Finally, a global state instability score is generated to characterize the global state of the target monitoring area, reflecting the overall degree of air quality anomalies at various locations within the area.

[0044] Step S30: Calculate the dynamic diffusion coefficient of the gas component diffusion in the target monitoring area based on the environmental condition data of the target monitoring area. Determine the local abnormal areas of the target monitoring area based on the global state instability score of the target monitoring area and the local state instability score of each data acquisition point for each monitored gas.

[0045] In this embodiment, considering the diffusion phenomenon of pollution sources in complex environments, and taking into account the influence of key environmental factors such as temperature, humidity, and ventilation coefficient on gas diffusion, the dynamic diffusion coefficient corresponding to the current environment is calculated based on the environmental condition data of the target monitoring area. Regarding the global performance of the target monitoring area, after detecting that the global state instability score of the target monitoring area exceeds a preset global instability threshold, precise detection and classification of local abnormal areas are initiated. The preset global instability threshold can be reasonably set based on information such as the characteristics of the monitoring area, air quality standards, or risk levels. For multiple data collection points in the target monitoring area, the local state instability score of each data collection point for each monitored gas is analyzed one by one. Data collection points with any local state instability score exceeding the preset global instability threshold are identified and recorded as abnormal nodes. For some data collection points that are relatively close together, considering the diversity of data collection point locations in different scenarios, such as office areas and green channel areas, the distribution of data collection points is relatively close, and abnormal scores may occur in all of them when a pollution source is present. Therefore, the location data of the data collection points are fused to avoid mistaking a single pollution source for multiple sources. During the fusion process, data collection points with a distance less than a pre-set distance threshold can be fused, and the area corresponding to the fused data collection points is recorded as a local abnormal area. In this way, one or more local abnormal areas that may exist in the target monitoring area can be identified.

[0046] Step S40: Construct an abnormal state diffusion model for the target monitoring area based on air quality detection data and dynamic diffusion coefficients from data collection points within the local abnormal area; determine pollution risk scores for multiple core areas based on the abnormal state diffusion model; and generate air quality early warning information for multiple core areas based on the pollution risk scores.

[0047] In this embodiment, for each defined local anomaly area, an anomaly diffusion model is constructed based on air quality monitoring data from multiple data collection points within the area and the overall dynamic diffusion coefficient of the target monitoring area. During this process, the location of the pollution source and its corresponding pollution intensity are initially determined based on the air quality monitoring data from the data collection points within the local anomaly area. The diffusion phenomenon of the corresponding monitored gas is described using the dynamic diffusion coefficient, thus constructing the anomaly diffusion model. The anomaly diffusion model is used to predict and analyze the pollution risk scores of multiple core areas over a future period. These core areas can be pre-defined areas requiring focused monitoring, or areas with high population density determined based on real-time personnel monitoring. Finally, air quality warning information for these core areas is generated based on the pollution risk scores. For example, if it is determined that light pollution may occur in the future, increased ventilation is recommended; if it is determined to be moderate pollution, air purification equipment is recommended to be turned on; if heavy pollution may occur, personnel are recommended to evacuate in advance or strengthen area sealing measures, etc.

[0048] The aforementioned indoor air monitoring and early warning method starts from air quality detection data at data collection points, gradually extracts local instability characteristics, integrates them to generate a global state score, analyzes local abnormal areas using dynamic diffusion coefficients, and achieves pollution risk assessment and refined early warning information generation for high-risk areas based on an abnormal diffusion model. This solves the shortcomings of some simple threshold-based monitoring and early warning technologies in terms of dynamic monitoring, anomaly location, and early warning accuracy, providing a comprehensive and accurate solution for air quality management in complex indoor environments.

[0049] In one implementation, for step S20 above, the state instability scores for each data acquisition point for each monitored gas are fused based on the location data and air quality detection data of the data acquisition points to generate a global state instability score for the target monitoring area, including:

[0050] Synchronization deviation analysis is performed on each data acquisition point based on air quality monitoring data. The time synchronization parameter between any two data acquisition points is calculated. The synchronization deviation parameter of each data acquisition point is calculated based on multiple time synchronization parameters. The state instability score of the data acquisition point for each monitored gas is corrected based on the synchronization deviation parameter, and the state instability correction score of the data acquisition point for each monitored gas is obtained.

[0051] In this embodiment, the synchronization deviation analysis mainly addresses the time synchronization problem between different data acquisition points. By comparing the gas concentration data of any two data acquisition points at the same time, the local time synchronization parameters between the two points are calculated.

[0052] Based on the air quality monitoring data corresponding to the data collection points, the local synchronization parameters between the two data collection points for any one monitored gas are calculated using the following formula: In the formula, q k1，k2 (t) represents the local time synchronization parameter of data acquisition points k1 and k2 at time t with respect to a certain monitored gas. A k1 (t), A k2 (t) represent the concentration values ​​of a certain monitored gas at data acquisition points k1 and k2 at time t, respectively. k1,k2 This represents the spatial distance between data collection point k1 and data collection point k2.

[0053] Based on air quality monitoring data, multiple local synchronization parameters for each monitored gas between any two data acquisition points can be calculated using the formula described above. Considering that air quality monitoring data may be acquired in real-time at high frequencies, for the time period corresponding to the air quality monitoring data, a sliding window can be used to iterate through the air quality monitoring data at 1-minute intervals to obtain multiple sets of characteristic data. Specifically, the average gas concentration within each sliding window is used as the gas concentration value at the time corresponding to the center point of that sliding window. This determines multiple local synchronization parameters for each monitored gas in the air quality monitoring data. The average of these multiple local synchronization parameters is then used as the time synchronization parameter between the two data acquisition points. Finally, the average of the multiple time synchronization parameters corresponding to each data acquisition point is calculated to obtain the synchronization deviation parameter for the data acquisition point.

[0054] The process of correcting the state instability score of each monitored gas at the data acquisition point based on the synchronization deviation parameter can be carried out as follows: State instability correction score = State instability score × (1 - synchronization deviation parameter). This yields the state instability correction score of each monitored gas at the data acquisition point, effectively compensating for scoring errors caused by differences in time distribution or noise between acquisition points, and improving the accuracy of the global state instability score.

[0055] For each data acquisition point, a collaborative analysis of the monitored gases is performed to calculate the dynamic correlation parameters between any two monitored gases at each data acquisition point. Based on multiple dynamic correlation parameters, the collaborative change parameters of each data acquisition point with respect to each monitored gas are calculated.

[0056] In this embodiment, the time-series concentration vectors for each monitored gas at the data collection points are extracted based on the air quality monitoring data corresponding to the data collection points. Correlation analysis is then performed on any two monitored gases based on these time-series concentration vectors to calculate the dynamic correlation parameters between them. For example, the Pearson correlation coefficient is used to measure the dynamic correlation characteristics between the two time-series concentration vectors. Finally, the average of the multiple dynamic correlation parameters for each monitored gas is calculated to obtain the synergistic change parameters for each monitored gas. Through dynamic synergistic change analysis, the complex correlation relationships among multiple monitored gases can be comprehensively evaluated, effectively compensating for the shortcomings of single-gas scoring in scenarios with the synergistic effects of multiple pollutants. This can be used to further optimize the rationality of local state instability scoring.

[0057] To generate a global state instability score for the target monitoring area, this method combines the location information of data acquisition points with air quality monitoring data. Compensation is achieved through synchronization deviation analysis and collaborative analysis of monitored gases at each data acquisition point, further improving the accuracy and robustness of the global state instability score. This method fully utilizes the local state information of the acquisition points while considering the spatial and temporal correlation characteristics of the monitored gases, ensuring higher adaptability and reliability in the global score calculation. In this process, multiple state instability correction scores are fused based on the location data of each data acquisition point and corresponding multiple collaborative change parameters to generate a global state instability score for the target monitoring area, where: In the formula, I ns C is the global state instability score for the target monitoring area. i,j Let R be the cooperative variation parameter of the i-th data acquisition point with respect to the j-th monitored gas. i,j The score for state instability correction at the i-th data acquisition point with respect to the j-th monitored gas is given. Let be the position weight of the i-th data acquisition point. Specifically, it can be reasonably set according to the relative distance between the data acquisition points. In this process, the data acquisition point with the highest state instability correction score can be used as the initial position for reference. The farther the data acquisition point is from the initial position, the smaller the corresponding position weight. m is the total number of data acquisition points, and n is the number of categories of monitored gases.

[0058] Synchronization deviation analysis resolved the issue of insufficient time synchronization between data collection points, improving the spatial consistency of the state score. Co-variation analysis compensated for the deficiency of single-gas scoring in potentially ignoring complex relationships among multiple pollutants, comprehensively enhancing the robustness of the score and enabling the calculated global state instability score to more realistically reflect the actual dynamic changes in air quality within the target monitoring area.

[0059] In one implementation process, for step S40 above, an abnormal state diffusion model of the target monitoring area is constructed based on the air quality detection data and dynamic diffusion coefficient of the data collection points within the local anomaly area, including:

[0060] The concentration parameters of each monitored gas in the local anomaly area are determined based on the air quality monitoring data from the data collection points. The initial intensity parameters of each monitored gas are obtained by correcting the concentration parameters based on the local instability score corresponding to the monitored gas.

[0061] In this embodiment, for the local anomaly area identified in the aforementioned steps, since it may contain multiple data collection points, the concentration parameters for each monitored gas at each data collection point are first determined. For a single data collection point, the concentration parameters are corrected using the local instability score corresponding to the monitored gas to estimate the intensity of the pollution source, thus calculating the initial intensity parameters for each monitored gas. For cases with multiple data collection points, after calculating the initial intensity parameters for each monitoring gas at each collection point, the maximum value is taken as a representative value to obtain the initial intensity parameters for the local anomaly area.

[0062] Two-dimensional convection-diffusion equations for local anomalies are generated based on the dynamic diffusion coefficient and the initial intensity parameters of the data acquisition points. An anomaly state diffusion model for the target monitoring area is then constructed based on the two-dimensional convection-diffusion equations for each local anomaly in the target monitoring area.

[0063] In this embodiment, the construction process of the two-dimensional convection-diffusion equation involves a dynamic diffusion coefficient describing the diffusion state, as well as the initial location and intensity of the pollution source. The core location of the local anomaly region can be used as the initial pollution source location. Based on the initial intensity parameters of the local anomaly region and the dynamic diffusion coefficient of the target monitoring area, a two-dimensional convection-diffusion equation for each monitored gas is constructed. The two-dimensional convection-diffusion equation is a well-known technique and will not be elaborated further here. It is worth noting that the two-dimensional convection-diffusion equation considers boundary conditions compared to the one-dimensional convection-diffusion equation. In practical implementation, indoor space boundaries, such as walls, can be set as impermeable boundaries to reflect the effect of the walls. Finally, for each local anomaly region, a two-dimensional convection-diffusion equation corresponding to each monitored gas is constructed, thereby generating an anomaly state diffusion model for the target monitoring area.

[0064] Based on air quality monitoring data and dynamic diffusion coefficients from data collection points in local anomaly areas within the target monitoring region, an anomaly diffusion model for the target monitoring region is constructed. This process aims to accurately describe the diffusion dynamics of each monitored gas within the target monitoring region by modeling the diffusion patterns of anomalies, thus providing support for subsequent pollution risk assessment and air quality early warning.

[0065] In constructing the two-dimensional convection-diffusion equation, a key aspect is determining the dynamic diffusion coefficient that describes the diffusion state. Specifically, the dynamic diffusion coefficient used to describe the diffusion state of gas components within the target monitoring area is obtained through the following analysis:

[0066] The environmental condition data of the target monitoring area is fed into the diffusion coefficient analysis model, and the dynamic diffusion coefficient corresponding to the environmental condition data of the target monitoring area is generated through the diffusion coefficient analysis model.

[0067] For the diffusion coefficient analysis model, a training dataset is constructed by acquiring multiple sets of historical data of the target monitoring area. Each set of historical data contains a set of correlated environmental condition data and gas composition detection data. The environmental condition data includes data on the changes in temperature, humidity, ventilation speed, etc. of the target monitoring area over a period of time. The gas composition detection data includes data on the changes in the concentration of various gases in different local areas of the target monitoring area over time during this period.

[0068] In constructing the training dataset by analyzing multiple sets of historical data, the two-dimensional convection-diffusion equation is discretized both temporally and spatially to generate the target diffusion function. Specifically, temporal discretization utilizes the finite difference method to transform the time derivative into a difference form. Spatial discretization involves dividing the target monitoring area into a two-dimensional grid and calculating the gas concentration variation at each grid point based on the difference, thereby constructing a target diffusion function that describes the spatial and temporal diffusion behavior of the gas components. The discretization of the two-dimensional convection-diffusion equation is a technique well-known to those skilled in the art, and its specific application is not limited here.

[0069] Then, the target diffusion function is fitted using multiple sets of historical data. The data fitting process can be achieved through multinomial fitting, random forest model fitting, etc. Those skilled in the art can choose an appropriate fitting scheme according to actual needs. The main process of data fitting is to analyze the distribution law of the diffusion coefficient in time and space dimensions based on the actual change law of the volume concentration, and finally generate the target diffusion coefficient corresponding to each set of historical data. Then, the environmental condition data in each set of historical data is associated with the corresponding target diffusion coefficient to construct a training dataset. The diffusion coefficient analysis model is trained through the training dataset. In the training process, the environmental condition data in the training dataset is used as the input feature, and the fitted target diffusion coefficient is used as the output target to complete the training process. For the diffusion coefficient analysis model, a multilayer perceptron is used as an example in this embodiment. After iteratively training the model through the training dataset, the final trained diffusion coefficient analysis model can be used to predict the diffusion coefficient corresponding to the target monitoring area under the current environment based on the input environmental condition data. Thus, after generating the dynamic diffusion coefficient of the gas component diffusion in the target monitoring area based on the environmental condition data of the target monitoring area, an abnormal state diffusion model of the target monitoring area is constructed.

[0070] In one implementation process, for step 40 above, pollution risk scores for multiple core areas are determined based on the anomalous state diffusion model, including:

[0071] The reference coordinate data corresponding to each core area is determined. Specifically, it can be the center position of the core area. For the abnormal state diffusion model, after substituting the reference coordinate data corresponding to the core area into any two-dimensional convection-diffusion equation, the finite difference method (FDM) can be used to discretize the convection-diffusion equation. The gas concentration data under the reference coordinates of the core area is calculated step by step through iterative methods such as explicit or implicit methods to obtain the concentration prediction data of the monitoring gas corresponding to the two-dimensional convection-diffusion equation at different times. Then, the core area is determined based on the concentration prediction data to obtain the pollution risk score. Solving the two-dimensional convection-diffusion equation under specific boundary conditions is a technical means well known to those skilled in the art. This embodiment does not specifically limit it.

[0072] The pollution risk score can be reasonably set according to pre-defined scoring rules. For example, the concentration ranges of different gases under light, moderate, and heavy pollution levels can be predetermined, as well as the pollution risk scores corresponding to different pollution levels. This allows for the prediction of pollution in different core areas at a future time, facilitating early warning of potential localized air pollution. The indoor air monitoring and early warning method provided in this invention can be applied to various indoor scenarios, such as industrial parks, office buildings, and laboratories. For environments where localized air pollution may occur due to production processes or experimental operations, the aforementioned monitoring and early warning method enables real-time monitoring and analysis to provide accurate and personalized air quality early warning services.

[0073] Please see Figure 2 The diagram illustrates a structural schematic of an IoT-based indoor air monitoring and early warning system according to one embodiment of the present invention. This system can be specifically used to implement the aforementioned IoT-based indoor air monitoring and early warning method, and includes the following structure:

[0074] The local state analysis module is used to acquire air quality detection data corresponding to multiple data collection points, extract multiple instability characteristic parameters for each data collection point for various monitored gases based on the air quality detection data, including volatility parameters and correlation parameters, and calculate the local state instability score for each data collection point for each monitored gas based on the multiple instability characteristic parameters.

[0075] The global state instability analysis module is used to acquire the location data of each data acquisition point, and to fuse the state instability scores of each data acquisition point for each monitored gas based on the location data and air quality detection data to generate a global state instability score for the target monitoring area.

[0076] The anomaly area analysis module is used to calculate the dynamic diffusion coefficient of gas component diffusion in the target monitoring area based on the environmental condition data of the target monitoring area, and to determine the local anomaly areas of the target monitoring area based on the global state instability score of the target monitoring area and the local state instability score of each data acquisition point for each monitored gas.

[0077] The air monitoring and early warning module is used to construct an abnormal state diffusion model of the target monitoring area based on air quality detection data and dynamic diffusion coefficients from data collection points in the local abnormal area, determine the pollution risk score of multiple core areas based on the abnormal state diffusion model, and generate air quality early warning information for multiple core areas based on the pollution risk score.

[0078] The above are merely specific embodiments of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art. Parts not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. An indoor air monitoring and early warning method based on Internet of Things, characterized in that, The method comprises the following steps: obtaining air quality detection data corresponding to a plurality of data collection points respectively, extracting a plurality of instability characteristic parameters corresponding to each data collection point for each monitoring gas respectively from the air quality detection data, the instability characteristic parameters including fluctuation parameters and correlation parameters, and calculating a local state instability score of each data collection point for each monitoring gas according to the plurality of instability characteristic parameters; obtaining position data of each data collection point, fusing the state instability scores of each data collection point for each monitoring gas according to the position data and the air quality detection data of the data collection points, generating a global state instability score of the target monitoring area, including performing synchronous deviation analysis on each data collection point according to the air quality detection data, calculating time synchronization parameters between any two data collection points, calculating a synchronous deviation parameter of each data collection point according to a plurality of time synchronization parameters, correcting the state instability score of each data collection point for each monitoring gas according to the synchronous deviation parameter, and obtaining a state instability correction score of each data collection point for each monitoring gas; performing monitoring gas coordination analysis on each data collection point, calculating dynamic correlation parameters between any two monitoring gases at each data collection point, and calculating a coordination change parameter of each data collection point for each monitoring gas according to a plurality of dynamic correlation parameters; The global state instability score of the target monitoring area is generated by fusing a plurality of state instability correction scores based on the position data of each data collection point and the corresponding plurality of synergistic change parameters, wherein: In the formula, is the global state instability score of the target monitoring area, is the state instability correction score of the first data collection point about the first monitoring gas, is the state instability correction score of the first data collection point about the first monitoring gas, is the position weight of the first data collection point, is the total number of data collection points, is the number of categories of monitoring gases; calculating a dynamic diffusion coefficient of the target monitoring area for gas component diffusion according to environmental condition data of the target monitoring area, including inputting the environmental condition data of the target monitoring area into a diffusion coefficient analysis model, and generating a dynamic diffusion coefficient corresponding to the environmental condition data of the target monitoring area through the diffusion coefficient analysis model; for the diffusion coefficient analysis model, a training data set is constructed by obtaining a plurality of groups of historical data of the target monitoring area, wherein each group of historical data contains a group of associated environmental condition data and gas component detection data, a target diffusion function is generated by time discretization and space discretization of a two-dimensional convection diffusion equation, the target diffusion function is fitted by a plurality of groups of historical data respectively, a target diffusion coefficient corresponding to each group of historical data is generated, the environmental condition data in each group of historical data is associated with the corresponding target diffusion coefficient, and a training data set is constructed, the diffusion coefficient analysis model is trained through the training data set, and a trained diffusion coefficient analysis model is obtained; determining a local abnormal area of the target monitoring area according to the global state instability score of the target monitoring area and the local state instability score of each data collection point for each monitoring gas. The abnormal state diffusion model of the target monitoring area is constructed according to the air quality detection data and the dynamic diffusion coefficient of the data collection points in the local abnormal area, including determining the concentration parameters of each monitoring gas in the local abnormal area according to the air quality detection data of the data collection points, correcting the concentration parameters based on the local state instability score of the corresponding monitoring gas to obtain the initial intensity parameters of each monitoring gas, generating the two-dimensional convection diffusion equation of the local abnormal area according to the dynamic diffusion coefficient and the initial intensity parameters of the data collection points, and constructing the abnormal state diffusion model of the target monitoring area according to the two-dimensional convection diffusion equation of each local abnormal area in the target monitoring area. For the local abnormal area of the target monitoring area, after detecting that the global state instability score of the target monitoring area is greater than the preset global instability threshold, a plurality of data collection points with a local state instability score greater than the preset global instability threshold in the target monitoring area are determined and recorded as abnormal nodes, and the plurality of abnormal nodes are fused according to the position data of the data collection points to determine a plurality of local abnormal areas. The pollution risk scores of the plurality of core areas are determined according to the abnormal state diffusion model, and the air quality early warning information about the plurality of core areas is generated according to the pollution risk scores. For the synchronization deviation analysis and monitoring gas collaborative analysis of the data collection points, further comprising: Based on the air quality monitoring data corresponding to the data collection points, the local synchronization parameters between the two data collection points for any one monitored gas are calculated using the following formula: In the formula, Indicates data collection point and data collection points exist Time-local synchronization parameters for a certain monitored gas , These represent the data collection points. and data collection points exist The concentration value of a certain monitored gas at any given time. Indicates data collection point and data collection points Spatial distance between them; After calculating the local synchronization parameters of each monitoring gas between any two data collection points based on the air quality detection data using the above formula, taking the mean value of the plurality of local synchronization parameters as the time synchronization parameter between the two data collection points, and calculating the mean value of the plurality of time synchronization parameters corresponding to each data collection point to obtain the synchronization deviation parameter of the data collection point. According to the air quality detection data corresponding to the data collection point, the time sequence concentration vector of the data collection point about each monitoring gas is extracted, the correlation analysis of any two monitoring gases is performed according to the time sequence concentration vector to calculate the dynamic correlation parameter between the two monitoring gases, and the collaborative change parameter of each monitoring gas is obtained by calculating the mean value of the plurality of dynamic correlation parameters corresponding to each monitoring gas. 2.The indoor air monitoring and early warning method based on the Internet of Things according to claim 1, characterized in that, The pollution risk scores of the plurality of core areas are determined according to the abnormal state diffusion model, including: determining the reference coordinate data corresponding to each core area, analyzing the reference coordinate data corresponding to each core area through the abnormal state diffusion model, including inputting the reference coordinate data corresponding to each core area into the plurality of two-dimensional convection diffusion equations respectively, processing the two-dimensional convection diffusion equations by finite difference method to obtain the concentration prediction data corresponding to the reference coordinate data at the preset time, and determining the pollution risk score of the core area according to the concentration prediction data. 3.The indoor air monitoring and early warning method based on the Internet of Things according to claim 2, characterized in that, According to the air quality detection data, the plurality of instability feature parameters corresponding to each data collection point about the plurality of monitoring gases are extracted, including: taking the mean value and standard deviation of the time sequence concentration vector of each monitoring gas, taking the ratio of the standard deviation to the mean value of the monitoring gas as the volatility parameter of the monitoring gas, and performing autocorrelation analysis on the time sequence concentration vector of each monitoring gas to calculate the correlation parameter of the monitoring gas.

4. An indoor air monitoring and early warning system based on Internet of Things, characterized in that, The system is used for realizing the indoor air monitoring and early warning method based on Internet of Things in any one of claims 1-3, comprising: a local state analysis module, configured to acquire air quality detection data corresponding to each data acquisition point, extract a plurality of instability characteristic parameters corresponding to each data acquisition point for each monitoring gas, including volatility parameters and correlation parameters, and calculate a local state instability score of each data acquisition point for each monitoring gas according to the plurality of instability characteristic parameters; a global state instability analysis module, configured to acquire position data of each data acquisition point, fuse the state instability score of each data acquisition point for each monitoring gas according to the position data and the air quality detection data, and generate a global state instability score of the target monitoring area; an abnormal area analysis module, configured to calculate a dynamic diffusion coefficient of the target monitoring area for gas component diffusion according to environmental condition data of the target monitoring area, determine a local abnormal area of the target monitoring area according to the global state instability score of the target monitoring area and the local state instability score of each data acquisition point for each monitoring gas; an air monitoring and early warning module, configured to construct an abnormal state diffusion model of the target monitoring area according to the air quality detection data and the dynamic diffusion coefficient of the data acquisition points in the local abnormal area, determine a pollution risk score of a plurality of core areas according to the abnormal state diffusion model, and generate air quality early warning information of the plurality of core areas according to the pollution risk score.

Images