A modular multi-functional environmental monitoring method

Abstract

The present application belongs to the technical field of environmental monitoring, and particularly relates to a modular multifunctional environmental monitoring method. The present application provides a modular multifunctional environmental monitoring method, which combines seasonal variation law of data and time series interpolation, guarantees the accuracy of data, and comprehensively considers various factors in the air environmental pollution process, i.e. time series trend, air pollutant influencing factors and physical and chemical processes, and thus realizes the monitoring of the environment.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring technology, and in particular relates to a modular, multifunctional environmental monitoring method. Background Technology

[0002] Summarizing and analyzing the spatiotemporal evolution characteristics of air quality helps in evaluating the effectiveness of policy implementation. Based on this, further exploration of the factors influencing urban air quality can provide more targeted policy recommendations and decision-making references for air pollution prevention and control. Existing research on regional air quality mostly focuses on single pollutants or analyzes the spatiotemporal characteristics of multiple pollutants in a single year, lacking comprehensive assessment studies based on multiple pollutants over multiple years. Due to the complexity of meteorological data, accurate pollutant concentration prediction still faces significant challenges. Currently, air environment monitoring tasks are typically modeled from two perspectives: based on atmospheric mechanisms and based on machine learning methods. The former analyzes the formation and transformation mechanisms of pollutants through physicochemical models, requiring extensive domain knowledge and exhibiting low short-term prediction accuracy. The latter achieves higher computational efficiency by constructing extrapolation relationships between historical observations and future predictions, but suffers from poor interpretability and long-term performance degradation. Therefore, this invention aims to combine the complementary characteristics of both approaches to explore new ideas for environmental monitoring methods. Summary of the Invention

[0003] This invention addresses the technical problems existing in current environmental monitoring by proposing a modular, multifunctional environmental monitoring method that is ingeniously designed, simple in method, and can effectively monitor environmental pollution.

[0004] To achieve the above objectives, the technical solution adopted by this invention is as follows: This invention provides a modular, multifunctional environmental monitoring method, comprising the following steps:

[0005] a. Data Acquisition Module: Utilizes distributed sensors to monitor meteorological and air pollution data, and connects the sensors to the Internet to achieve remote monitoring and data acquisition;

[0006] b. Communication module: transmits the collected meteorological data and air pollution data to the data processing center;

[0007] c. Data processing module: Preprocesses and calibrates the transmitted data, mainly including the removal of abnormal data, filling in missing values, and standardization;

[0008] d. Environmental assessment module: Based on real-time collected meteorological and air pollution data and historical meteorological and air pollution data, an air environment prediction model is constructed.

[0009] f. Abnormal alarm: The system presets air environment thresholds and uses an air environment prediction model to analyze the monitored data in real time. Once an abnormal situation is detected, the system will trigger an alarm.

[0010] Preferably, in step a, the meteorological data monitored by the distributed sensor includes: temperature, humidity, air pressure, wind direction, and wind speed; the air pollution data includes: CO, NO2, O3, PM10, and PM2.5. 2.5 SO2.

[0011] Preferably, in step c, the Mahalanobis distance algorithm is used to remove abnormal data from the meteorological and air pollution data; the ARIMA model is used to fill in the missing values ​​of the acquired meteorological and air pollution data; and the maximum-minimum normalization method is used to standardize the acquired data.

[0012] Preferably, in step c, when using the ARIMA model to fill in missing data, the data needs to be smoothed. Considering the seasonal variations in the collected meteorological and air pollution data, seasonal features are incorporated into the data smoothing process to make the smoothing results more accurate and reliable. The specific calculation steps are as follows:

[0013] (1) Calculate the seasonality index Se:

[0014]

[0015] Where i represents the month, i = 1, 2, ..., 12. Let be the monthly average concentration of a certain pollutant in the i-th month. This represents the annual average concentration of a certain pollutant.

[0016] (2) Seasonal adjustment:

[0017]

[0018] Where n is the size of the smoothing window, n∈[1,12], Q i This represents the seasonally adjusted data for the i-th month.

[0019] Preferably, in step d, the air environment prediction model is a combined model that comprehensively considers time series trends, influencing factors, and physicochemical processes. The specific modeling process is as follows:

[0020] d1. Calculate the Air Quality Index (AQI);

[0021] d2. Use the ARIMA model to conduct time series analysis on historical AQI, construct a statistical model, and predict changes in air quality AQI1;

[0022] d3. An improved particle swarm optimization algorithm is used to optimize the key parameters of the SVR model, constructing a machine learning model to learn the complex relationship between AQI and various air pollution data from monitoring data, and predict changes in the air quality (AQI2). The main optimization steps are as follows:

[0023] d31. Set the particle population size N p Number of iterations k p Boundary values ​​of parameters C and γ;

[0024] d32. Initialize particle positions and speed

[0025] d33. Calculate the fitness value;

[0026] d34. Introducing a linear combination of individual optimality and swarm optimality in the velocity update term broadens the particle's search space, thereby increasing the likelihood of the algorithm finding the global optimum. This updates the velocity and position of each particle.

[0027]

[0028] Where, p i p represents the particle's position. i =(C i ,γ i ), pbest i is the optimal position experienced by the i-th particle; gbest is the optimal position of the particle swarm; w is the inertia weight; c1 and c2 are the learning step sizes; r1 and r2 are random numbers, r1 and r2 ∈ [0, 1];

[0029] d4. Based on monitoring meteorological data and air pollution data, establish a physical-chemical model based on atmospheric science and chemical principles to simulate the transport, transformation and deposition processes of pollutants in the atmosphere, and predict changes in the air environment (AQI3).

[0030] d5. The weighted average of the prediction results from the three prediction models is used to obtain the final prediction result, AQI. The calculation formula is as follows:

[0031] AQI = w1AQI1 + w2AQI2 + w3AQI3

[0032] Where w1, w2, and w3 are weighting coefficients, satisfying w1 + w2 + w3 = 1.

[0033] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0034] 1. This invention provides a modular, multifunctional environmental monitoring method. It acquires real-time meteorological and air pollution data through sensors, and performs missing and outlier processing and data standardization on the meteorological and air pollution data through remote communication and data analysis. By combining the seasonal variation patterns of the data with time series interpolation, the accuracy of the data can be guaranteed.

[0035] 2. When using the machine learning SVR model to predict AQI, the improved particle swarm optimization algorithm is used to optimize the key parameters of SVR. By introducing a linear combination of individual optimality and swarm optimality, the search space of particles can be broadened, increasing the possibility of the algorithm finding the global optimal solution.

[0036] 3. When modeling, consider combining statistical models, machine learning models, and physical-chemical models to leverage their respective advantages and improve the overall accuracy and robustness of predictions.

[0037] 4. Modular design allows for customized system configurations to meet specific needs, enhancing system flexibility and scalability. It integrates multiple sensors and functional modules to satisfy environmental monitoring requirements in various scenarios. Attached Figure Description

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This provides a modular, multifunctional environmental monitoring method for Example 1. Detailed Implementation

[0040] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0041] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0042] Example 1: This example provides a modular, multifunctional environmental monitoring method.

[0043] In this embodiment, existing environmental monitoring typically relies on two methods: atmospheric numerical weather prediction models and machine learning methods. Atmospheric numerical weather prediction models, based on mathematical tools, simulate a series of physicochemical processes in the atmosphere, including the transport, diffusion, removal, and aerosol chemistry of air pollutants. They can achieve spatially comprehensive simulation of large-scale air pollutant concentrations, but their overall prediction accuracy is relatively low. Machine learning, unlike atmospheric numerical weather prediction models, does not rely on the systematic knowledge of the modeling process but is trained on a large number of samples, effectively reducing model bias. However, machine learning methods are limited by data availability, and previous studies have mostly relied on measured data from ground-based air quality monitoring stations for prediction. As air pollutants, CO, NO2, O3, and PM2.5 are... 10 PM 2.5 The formation mechanism and process of SO2 are complex, exhibiting uncertainty and instability, and are nonlinear in both time and space. Machine learning algorithms can perform nonlinear fitting on air pollutant-related data. To improve environmental monitoring capabilities, this embodiment proposes a combined prediction model that integrates statistical models, machine learning models, and physicochemical models, leveraging their respective advantages to enhance prediction accuracy and robustness.

[0044] Due to the influence of internal and external factors such as the structural characteristics of sensors and signal transmission paths, the meteorological and air pollution data collected by sensors often contain anomalous data and are accompanied by some missing data. Data processing is required before modeling. First, the ARIMA model is used to interpolate and fill in the missing values ​​of the acquired meteorological and air pollution data. Before filling, the data needs to be smoothed. This embodiment considers the seasonal variations in the collected meteorological and air pollution data. Seasonal features are incorporated into the data smoothing process, which helps to remove seasonal fluctuations in the data, making it easier to identify the true trends and periodic changes, accurately capture the core features of the trend, and reduce the impact of noise and random fluctuations in the data, making the data more accurate and reliable. The specific calculation steps are as follows:

[0045] (1) Calculate the seasonality index Se:

[0046]

[0047] Where i represents the month, i = 1, 2, ..., 12. Let be the monthly average concentration of a certain pollutant in the i-th month. This represents the annual average concentration of a certain pollutant.

[0048] (2) Seasonal adjustment:

[0049]

[0050] Where n is the size of the smoothing window, n∈[1,12], Q i This represents the seasonally adjusted data for the i-th month.

[0051] The collected data, after seasonal smoothing adjustment, were interpolated and filled using the ARIMA model for meteorological and air pollution data. Then, the Mahalanobis distance algorithm was used to remove outliers from the acquired meteorological and air pollution data. Finally, the data were standardized using the max-min normalization method.

[0052] The formation mechanisms and processes of air pollutants such as CO, NO2, O3, PM10, PM2.5, and SO2 are complex, exhibiting uncertainty and instability, and are nonlinear in both time and space. A single predictive model is unlikely to fully capture all these characteristics, resulting in unsatisfactory prediction accuracy. This embodiment proposes a combined predictive model, such as... Figure 1 As shown, statistical models, machine learning models, and physicochemical models are combined to leverage their respective advantages and improve the accuracy and robustness of environmental measurements. The specific modeling method is as follows:

[0053] (1) Calculate the Air Quality Index (AQI);

[0054] (2) Use the ARIMA model to conduct time series analysis on historical AQI, construct a statistical model, and predict changes in air quality AQI1.

[0055] (3) Due to the uncertainty and instability of air pollution data, and its nonlinearity in time and space, machine learning algorithms can perform nonlinear fitting on air pollutant-related data. In this embodiment, the SVR machine learning algorithm is selected to determine the nonlinear relationship between air pollutants and AQI. However, the SVR algorithm is prone to local optima. Therefore, in this embodiment, the particle swarm optimization algorithm is used to optimize the key parameters of the SVR model. The traditional particle swarm optimization algorithm consists of a group of particles moving in the search space. It is sensitive to the selection of parameters and is prone to getting trapped in local optima. It has limited global search capability for complex problems and has a slow convergence speed. Therefore, this embodiment introduces a linear combination of individual optima and swarm optima to optimize the particle swarm algorithm, making the particle search space wider. The main optimization steps are as follows:

[0056] a. Set the particle population size N p The number of iterations k, the boundary values ​​of parameters C and γ;

[0057] b. Initialize particle positions and speed

[0058] c. Calculate the fitness value;

[0059] d. Introduce a linear combination of individual optimality and group optimality in the velocity update term. To replace pbest in the traditional particle swarm optimization algorithm i k and gbest k This allows for a wider search space for particles, thereby increasing the likelihood that the algorithm will find the global optimum, and updating the velocity and position of each particle:

[0060]

[0061] Where, p i p represents the particle's position. i =(C i ,γ i ), pbest i k This represents the optimal position experienced by the i-th particle after the k-th iteration; gbest k The optimal position of the particle swarm after the k-th iteration is given by: w is the inertia weight; c1 and c2 are the learning step sizes; r1 and r2 are random numbers, r1 and r2 ∈ [0, 1]; v i k Indicates the particle's current velocity. The attractive force of the particle's current position on the positive direction of the global best position (gbest) and the positive direction of its own best position (pbest); The particle's current position is attracted to the average of its own best position in the positive direction (pbest) and the global best position in the negative direction (-gbest). A linear combination strategy of individual optimality and swarm optimality is introduced to update the particle's velocity and position, making it easier for the particle to escape local minima, improving the global search capability, and achieving higher convergence accuracy.

[0062] (4) Based on the monitoring meteorological data and air pollution data, establish a physical-chemical model based on atmospheric science and chemical principles to simulate the transmission, transformation and deposition of pollutants in the atmosphere and predict changes in the air environment (AQI3).

[0063] (5) The prediction results of the three prediction models are weighted and averaged to obtain the final prediction result AQI. The calculation formula is as follows:

[0064] AQI = w1AQI1 + w2AQI2 + w3AQI3

[0065] Where w1, w2, and w3 are weighting coefficients, satisfying w1 + w2 + w3 = 1.

[0066] Finally, based on the preset air environment thresholds, the system uses the aforementioned air environment prediction model to analyze the monitored data in real time. Once an anomaly is detected, the system will trigger an alarm.

[0067] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments for application in other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A modular, multifunctional environmental monitoring method, characterized in that, Includes the following steps: a. Data Acquisition: Utilize distributed sensors to acquire meteorological and air pollution data, and connect the sensors to the Internet to achieve remote monitoring and data acquisition; b. Data communication: Transmitting the collected meteorological data and air pollution data to the data processing center; c. Data processing: Preprocessing and calibrating the transmitted data, mainly including the removal of abnormal data, filling in missing values, and standardization; d. Environmental assessment: Based on real-time meteorological and air pollution data and historical meteorological and air pollution data, construct an air environment prediction model. f. Abnormal alarm: The system presets air environment thresholds and uses an air environment prediction model to analyze the monitored data in real time. Once an abnormal situation is detected, the system will trigger an alarm. In step c, when using the ARIMA model to fill in missing data, the data needs to be smoothed. Considering the seasonal variations in the collected meteorological and air pollution data, seasonal features are incorporated into the smoothing process to make the smoothing results more accurate and reliable. The specific calculation steps are as follows: (1) Calculate the seasonality index Se: Where i represents the month, i = 1, 2, ..., 12. Let be the monthly average concentration of a certain pollutant in the i-th month. This represents the annual average concentration of a certain pollutant. (2) Seasonal adjustment: Where n is the size of the smoothing window, n∈[1,12], Q i This represents the seasonally adjusted data for the i-th month. In step d, the air environment prediction model is a combined model that comprehensively considers time series trends, influencing factors, and physicochemical processes. The specific modeling process is as follows: d1. Calculate the Air Quality Index (AQI); d2. Use the ARIMA model to perform time series analysis on historical AQI values ​​and construct a statistical model. Predicting changes in air quality (AQI1); d3. An improved particle swarm optimization algorithm is used to optimize the key parameters of the SVR model, constructing a machine learning model to learn the complex relationship between AQI and various air pollution data from monitoring data, and predict changes in the air quality (AQI2). The main optimization steps are as follows: d31. Set the particle population size N p Number of iterations k p Boundary values ​​of parameters C and γ; d32. Initialize particle positions and speed d33. Calculate the fitness value; d34. Introducing a linear combination of individual optimality and swarm optimality in the velocity update term broadens the particle's search space, thereby increasing the likelihood of the algorithm finding the global optimum. This updates the velocity and position of each particle. Where, p i p represents the particle's position. i =(C i ,γ i ), pbest i is the optimal position experienced by the i-th particle; gbest is the optimal position of the particle swarm; w is the inertia weight; c1 and c2 are the learning step sizes; r1 and r2 are random numbers, r1 and r2 ∈ [0, 1]; d4. Based on meteorological and air pollution data, establish a physical-chemical model based on atmospheric science and chemical principles to simulate the transport, transformation, and deposition processes of pollutants in the atmosphere and predict changes in the air environment (AQI3). d5. Calculate the weighted average of the prediction results from the three prediction models to obtain the final prediction result, AQI. The calculation formula is as follows: AQI = w1AQI1 + w2AQI2 + w3AQI3 Where w1, w2, and w3 are weighting coefficients, satisfying w1 + w2 + w3 = 1.

2. The modular, multifunctional environmental monitoring method according to claim 1, characterized in that, In step a), the meteorological data monitored by the distributed sensors include: temperature, humidity, air pressure, wind direction, and wind speed; the air pollution data includes: CO, NO2, O3, and PM2.

5. 10 PM 2.5 SO2.

3. The modular, multifunctional environmental monitoring method according to claim 1, characterized in that, In step c, the Mahalanobis distance algorithm is used to remove outlier data from the acquired meteorological and air pollution data; the ARIMA model is used to fill in missing values ​​in the acquired meteorological and air pollution data; and the max-min normalization method is used to standardize the acquired data.

Images