A pollution monitoring system for groundwater
Through the groundwater-oriented pollution monitoring system, the pollutant concentration in sewage is monitored in real time and intelligent prediction is carried out, which solves the problems of untimely treatment of water quality exceeding standards and high operation and maintenance costs in the existing technology, real-time calculation and intelligent early warning of sewage treatment trends , reducing the risk of environmental pollution.
Patent Information
- Application Number
- CN202311746558.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-19
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2043-12-19
AI Technical Summary
The existing technology is difficult to achieve effective pretreatment when water quality exceeds the standard, the operation and maintenance costs are high and the experience judgment is inaccurate, which may lead to environmental pollution and misjudgment.
A pollution monitoring system for groundwater has been designed, including data acquisition module, data processing module, trend prediction module and alarm module, to monitor the concentration of pollutants in sewage in real time, conduct intelligent and real-time future trend prediction, and issue it based on the prediction results. Alarm prompt.
Real-time calculation and intelligent early warning of sewage treatment trends are achieved, which reduces environmental pollution risks, reduces operation and maintenance costs, and improves the accuracy of pretreatment.
Smart Images

Figure CN117689114B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of groundwater monitoring, and in particular relates to a groundwater pollution monitoring system. Background Art
[0002] Groundwater is one of the important water sources for agricultural irrigation and urban use. Under certain conditions, its overexploitation and pollution may cause swamping, salinization, landslides, ground subsidence and other phenomena that are not conducive to human beings. Even surface water has been polluted to varying degrees. Comprehensive and long-term monitoring of groundwater will help to take different protective measures for groundwater in different places.
[0003] The current treatment methods for water quality exceeding the standard have the following problems:
[0004] 1. When the water quality has exceeded the standard, taking remedial measures will cause some sewage that does not meet the standard to be discharged, thus causing environmental pollution;
[0005] 2. Pre-processing based on one's own work experience places high demands on operation and maintenance personnel, and requires personnel to keep a close eye on various operating data in real time. The operation and maintenance cost is high, and experience-based judgments are sometimes inaccurate and may lead to misjudgments.
[0006] Therefore, in view of the above problems, the present invention proposes a groundwater pollution monitoring system. Summary of the invention
[0007] In view of the shortcomings of the existing technology, the present invention proposes a groundwater pollution monitoring system, which monitors the relevant data of pollutant concentration in the sewage to be detected in real time, so as to calculate the emission of each pollutant in real time, and make intelligent and real-time future trend predictions on the development trends of various data and the trends of sewage treatment, truly realizing the real-time calculation and intelligent prediction and early warning of the development trends of sewage treatment.
[0008] To achieve the above object, the present invention provides the following solutions:
[0009] A groundwater pollution monitoring system includes: a data acquisition module, a data processing module, a trend prediction module and an alarm module;
[0010] The data acquisition module is used to collect groundwater data information in the monitoring area;
[0011] The data processing module is used to track and analyze the collected groundwater data information to obtain groundwater pollutant concentrations at different levels;
[0012] The trend prediction module is used to predict the future trend of deep pollution discharge behavior based on the groundwater pollutant concentrations at different levels and the pollution discharge data of the monitoring area within a preset period of time;
[0013] The alarm module is used to issue an alarm if it is determined that deep sewage discharge may occur.
[0014] Preferably, the data acquisition module includes a monitoring point distribution acquisition unit, a data signal receiver and a cloud platform;
[0015] The monitoring point distribution collection unit is used to collect data from monitoring points at different levels to obtain groundwater quality data within the monitoring area;
[0016] The data signal receiver is used to transmit the quality data to the cloud platform;
[0017] The cloud platform is used to preprocess the quality data and store the preprocessed quality data.
[0018] Preferably, the process of preprocessing the quality data includes: removing duplicates and outliers, supplementing missing values using a linear interpolation method, and unifying the time resolution of the quality data.
[0019] Preferably, the data processing module includes: a slicing unit, a construction unit, a recognition model obtaining unit and a pollutant concentration obtaining unit;
[0020] The slicing unit is used to slice the preprocessed quality data, and divide the sliced quality data into a training set and a prediction set;
[0021] The construction unit is used to construct a convolutional neural network;
[0022] The recognition model acquisition unit is used to train the convolutional neural network based on the training set, test the trained convolutional neural network based on the prediction set, and obtain a groundwater quality recognition model;
[0023] The pollutant concentration obtaining unit is used to obtain groundwater pollutant concentrations at different levels based on the groundwater quality identification model.
[0024] Preferably, the convolutional neural network includes 4 convolutional layers, 4 batch normalization layers, 4 Relu layers, 2 maximum pooling layers, 1 fully connected layer and a softmax layer, and adam is set as the solver.
[0025] Preferably, the trend prediction module includes: an impact relationship quantification unit, a prediction model construction unit, a training unit and a prediction unit;
[0026] The influence relationship quantification unit is used to combine the groundwater pollutant concentrations at different levels in the monitoring area and the pollution discharge data of the monitoring area within a preset period of time, and divide the acquired data into a data set and a test set, and use big data analysis and processing methods to perform multi-dimensional characterization of the factors and parameters affecting the change of groundwater pollutant concentrations in the monitoring area, and determine the quantitative influence relationship of different influencing factors on the change of groundwater pollutant concentrations in the monitoring area;
[0027] The prediction model building unit is used to establish a NARX neural network prediction model for deep-seated sewage discharge behavior based on quantitative influencing factors of groundwater pollutant concentration changes in the monitoring area;
[0028] The training unit is used to train the deep sewage discharge behavior NARX neural network prediction model based on the data set to obtain a trained deep sewage discharge behavior NARX neural network prediction model;
[0029] The prediction unit is used to input the test set into the trained deep sewage discharge behavior NARX neural network prediction model to complete the future trend prediction of deep sewage discharge behavior.
[0030] Preferably, the test set is input into the trained NARX neural network prediction model for deep sewage discharge behavior, and the process of completing the prediction of the future trend of deep sewage discharge behavior includes:
[0031] Obtaining time series sample data of groundwater pollutant concentration and pollution discharge data at different levels, and calculating the autocorrelation coefficient and partial correlation coefficient of the sample data;
[0032] Determining the order of the time series model based on the autocorrelation coefficient and the partial correlation coefficient;
[0033] Using the maximum likelihood estimation method, the parameters of the time series model are accurately estimated to obtain unknown parameter values;
[0034] Performing an error test on the time series model, if the test result meets the preset conditions, obtaining the final time series model, if not, re-determining the order and re-estimating the parameters;
[0035] Based on the final time series model, future trend prediction of deep sewage discharge behavior is carried out.
[0036] Preferably, the alarm module divides the indicators of deep sewage discharge behavior into the following levels: threat to human health as the primary level, threat to ecological safety as the secondary level, and potential pollution as the third level, to obtain three-level groundwater pollution monitoring indicators.
[0037] Compared with the prior art, the present invention has the following beneficial effects:
[0038] The present invention collects groundwater data information in the monitoring area; tracks and analyzes the collected groundwater data information to obtain groundwater pollutant concentrations at different levels; predicts future trends of deep-layer pollution discharge behavior based on groundwater pollutant concentrations at different levels and pollution discharge data in the monitoring area within a preset period of time; and issues an alarm if it is determined that deep-layer pollution discharge behavior may exist. The present invention monitors the relevant data of pollutant concentrations in the sewage to be detected in real time to calculate the emission of each pollutant in real time, and makes intelligent and real-time future trend predictions on the development trends of various data and the trends of sewage treatment, truly realizing real-time calculation and intelligent prediction and early warning of the development trend of sewage treatment. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0040] Figure 1 The figure is an overall framework diagram of a groundwater pollution monitoring system in an embodiment of the present invention. DETAILED DESCRIPTION
[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0043] Embodiment 1
[0044] like Figure 1 As shown, the present invention discloses a groundwater pollution monitoring system, comprising: a data acquisition module, a data processing module, a trend prediction module and an alarm module;
[0045] The data acquisition module is used to collect groundwater data information in the monitoring area;
[0046] The data processing module is used to track and analyze the collected groundwater data information to obtain the groundwater pollutant concentrations at different levels;
[0047] The trend prediction module is used to predict the future trend of deep-seated pollution discharge behavior based on the groundwater pollutant concentrations at different levels and the pollution discharge data of the monitoring area within a preset period of time;
[0048] The alarm module is used to issue an alarm if it is determined that there may be deep sewage discharge.
[0049] In this embodiment, the data acquisition module includes a monitoring point distribution acquisition unit, a data signal receiver and a cloud platform;
[0050] The monitoring point distribution collection unit is used to collect data from monitoring points at different levels to obtain the quality data of groundwater in the monitoring area;
[0051] The data signal receiver is used to transmit the quality data to the cloud platform;
[0052] The cloud platform is used to preprocess the quality data and store the preprocessed quality data.
[0053] In this embodiment, the process of preprocessing the quality data includes: removing duplicate items and outliers, supplementing missing values using a linear interpolation method, and unifying the time resolution of the quality data.
[0054] In this embodiment, the data processing module includes: a slicing unit, a construction unit, a recognition model obtaining unit and a pollutant concentration obtaining unit;
[0055] The slicing unit is used to slice the preprocessed quality data, and divide the sliced quality data into a training set and a prediction set;
[0056] The building block is used to construct a convolutional neural network;
[0057] The recognition model acquisition unit is used to train the convolutional neural network based on the training set, test the trained convolutional neural network based on the prediction set, and obtain a groundwater quality recognition model;
[0058] The pollutant concentration acquisition unit is used to obtain groundwater pollutant concentrations at different levels based on a groundwater quality identification model.
[0059] Specifically, the minimum sample slice length that can retain the original water quality characteristics is determined according to the concentration of pollutants at each level. Each divided segment must not only meet the requirements of short slicing time and dynamic prediction accuracy, but also ensure that the original water quality characteristics are covered.
[0060] Specifically: The construction process of the groundwater quality identification framework based on convolutional neural network includes:
[0061] A typical CNN consists of a convolutional layer, a pooling layer, a fully connected layer, and an output layer.
[0062] In the convolution layer, the output features of the previous layer are convolved with the convolution kernel of this layer, and new features are formed through the activation function as the input of the next layer. The convolution operation can be expressed by mathematical expression as follows:
[0063]
[0064] In formula (1), is the j-th output of layer l; M j is the jth convolution region of layer l-1; is the jth convolution kernel associated with the i-th feature in the l-th layer; is the bias of the jth convolution kernel in the lth layer;
[0065] The main purpose of pooling is to downsample and thus reduce the dimension of features. The expression of the maximum pooling layer is as follows:
[0066]
[0067] In formula (2), is the value of the tth neuron in the ith feature of the lth layer; W is the width of the pooling area; is the value of the jth neuron in the i-th feature of the l+1-th layer.
[0068] The fully connected layer connects each node to all nodes in the previous layer to integrate the features extracted previously. Its expression is:
[0069] ε i =f(w i p i +b i ) (3)
[0070] In formula (3), p i is the input of the fully connected layer; w i and b i are weight and threshold respectively; ε i is the output of the fully connected layer.
[0071] The output layer usually uses a softmax classifier to output classification labels. The softmax function can be expressed as:
[0072]
[0073] In formula (4), f(z i ) is the probability of output for each category, and M is the number of categories in the multi-classification problem.
[0074] The code used to build the model is written in MATLAB. The first layer defines the size and type of the input data. After the data is preprocessed, each data record contains 16 channels and 2560 data points (the data sampling frequency is 512Hz, and the duration of each sample is 5s). The middle layer of the network defines the core architecture of the network. The CNN structure designed in this paper includes 4 convolutional layers, 4 batch normalization layers, 4 Relu layers, 2 maximum pooling layers, 1 fully connected layer and a softmax layer. Adam is set as the solver, and miniBatchSize is set to 128. (When using minibatch, all samples in a batch can be placed in a matrix, and the linear algebra library can be used to accelerate the calculation of gradients. This is an optimization method in engineering implementation).
[0075] In this embodiment, the trend prediction module includes: an influence relationship quantification unit, a prediction model construction unit, a training unit, and a prediction unit;
[0076] The influence relationship quantitative unit is used to combine the groundwater pollutant concentrations at different levels in the monitoring area and the pollution discharge data of the monitoring area within a preset period, and divide the acquired data into a data set and a test set, and use big data analysis and processing methods to perform multi-dimensional characterization of the factors and parameters affecting the change of groundwater pollutant concentrations in the monitoring area, and determine the quantitative influence relationship of different influencing factors on the change of groundwater pollutant concentrations in the monitoring area;
[0077] The prediction model building unit is used to establish a NARX neural network prediction model for deep-seated sewage discharge behavior based on the quantitative influencing factors of groundwater pollutant concentration changes in the monitoring area;
[0078] The training unit is used to train the deep sewage discharge behavior NARX neural network prediction model based on the data set to obtain the trained deep sewage discharge behavior NARX neural network prediction model;
[0079] The prediction unit is used to input the test set into the trained deep sewage discharge behavior NARX neural network prediction model to complete the prediction of the future trend of deep sewage discharge behavior.
[0080] Specifically, two functions, f(x) = tanh(x) and g(x) = Softplus(x), are used as activation functions of the neural network model to achieve forward propagation of the signal in the model. At the same time, the output error is calculated, and the weight of each synapse is updated by differentiating f(x) or g(x) to achieve backward propagation of the error.
[0081] in:
[0082]
[0083] g(x)=Softplus(x)=log(1+e x )
[0084]
[0085] Taking the prediction model of groundwater pollutant concentration changes in a certain area as an example, the weights of the neural network model are trained through the activation function, and the initial influencing factors must be set first:
[0086]
[0087] Where m1 1n (1×7), m1 2n (1×7) are the weights from the input layer to the hidden layer H(1) and the weights from the input layer to the hidden layer H(2), m2 n (1×2) is the weight from the hidden layer to the output layer.
[0088] Multiply each factor of the input layer by its weight:
[0089]
[0090] In the formula, x 1n are the factors of the input layer,
[0091] Taking the tanh(x) activation function as an example, calculate the output of the hidden layer H(1) of the neural network in this model:
[0092]
[0093] Where, X n It is the result of multiplying each factor of the input layer by the weight from the input layer to the hidden layer H(1).
[0094] Comparing the output of the hidden layer H(1) with the true result, the error is:
[0095]
[0096] In the formula, ΔR represents the error, y net1 is the received hidden layer input signal, that is, the weighted sum of the factors in the input layer. out1 It is the output value of the hidden layer after being processed by the activation function tanh(x).
[0097] After the neural network model is established, a large number of training samples and multiple training sessions are used to select a function that is more suitable for predicting changes in regional groundwater pollutant concentrations.
[0098] The neural network model was built with tanh(x) as the activation function. After training, the model was built using Softplus(x) to compare the accuracy of the two models.
[0099] Specifically, the process of inputting the test set into the trained NARX neural network prediction model of deep sewage discharge behavior to complete the prediction of the future trend of deep sewage discharge behavior includes:
[0100] Obtain time series sample data of groundwater pollutant concentration and pollution discharge data at different levels, and calculate the autocorrelation coefficient and partial correlation coefficient of the sample data;
[0101] Determining the order of the time series model based on the autocorrelation coefficient and the partial correlation coefficient;
[0102] The maximum likelihood estimation method is used to accurately estimate the parameters of the time series model and obtain the unknown parameter values;
[0103] Perform error test on the time series model. If the test result meets the preset conditions, the final time series model is obtained. If not, the order is determined again and the parameters are accurately estimated.
[0104] Based on the final time series model, future trend predictions of deep sewage discharge behavior are conducted.
[0105] In this embodiment, the alarm module divides the indicators of deep sewage discharge behavior into the following levels: threat to human health as the primary level, threat to ecological security as the secondary level, and potential pollution as the third level, to obtain three-level groundwater pollution monitoring indicators.
[0106] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.
Claims
1. A groundwater pollution monitoring system, characterized in that: include: Data acquisition module, data processing module, trend prediction module and alarm module; The data acquisition module is used to collect groundwater data information in the monitoring area; The data processing module is used to track and analyze the collected groundwater data information to obtain groundwater pollutant concentrations at different levels; The trend prediction module is used to predict the future trend of deep pollution discharge behavior based on the groundwater pollutant concentrations at different levels and the pollution discharge data of the monitoring area within a preset period of time; The alarm module is used to issue an alarm if it is determined that there may be deep sewage discharge behavior; The trend prediction module includes: an influence relationship quantification unit, a prediction model construction unit, a training unit and a prediction unit; The influence relationship quantification unit is used to combine the groundwater pollutant concentrations at different levels in the monitoring area and the pollution discharge data of the monitoring area within a preset period of time, and divide the acquired data into a data set and a test set, and use big data analysis and processing methods to perform multi-dimensional characterization of the factors and parameters affecting the change of groundwater pollutant concentrations in the monitoring area, and determine the quantitative influence relationship of different influencing factors on the change of groundwater pollutant concentrations in the monitoring area; The prediction model building unit is used to establish a NARX neural network prediction model for deep-seated sewage discharge behavior based on quantitative influencing factors of groundwater pollutant concentration changes in the monitoring area; The training unit is used to train the deep sewage discharge behavior NARX neural network prediction model based on the data set to obtain a trained deep sewage discharge behavior NARX neural network prediction model; The prediction unit is used to input the test set into the trained deep sewage discharge behavior NARX neural network prediction model to complete the future trend prediction of deep sewage discharge behavior; The two functions f(x) = tanh(x) and g(x) = softplus(x) are used as activation functions of the neural network model to realize the forward propagation of the signal in the model; at the same time, the output error is calculated, and the weight of each synapse is updated by differentiating f(x) or g(x) to realize the backward propagation of the error; in: To train the weights of the neural network model through the activation function, the initial influencing factors must be set first: Where m1 1n (1×7), m1 2n (1×7) are the weights from the input layer to the hidden layer H(1) and the weights from the input layer to the hidden layer H(2), m2 n (1×2) is the weight from the hidden layer to the output layer; Multiplying the factors of the input layer with the weights from the input layer to the hidden layer H(1) yields: In the formula, x 1n are the factors of the input layer, Tanh(x) is the activation function, and the output of the hidden layer H(1) of the neural network in this model is calculated as: Where, X n It is the result of multiplying each factor of the input layer by the weight from the input layer to the hidden layer H(1); Comparing the output of the hidden layer H(1) with the true result, the error is: In the formula, ΔR represents the error, y net1 is the received hidden layer input signal, that is, the weighted sum of the factors in the input layer; Y out1 It is the output value of the hidden layer after being processed by the activation function tanh(x); After the neural network model is established, a large number of training samples and multiple trainings are used to select a function that is more suitable for predicting changes in regional groundwater pollutant concentrations; The neural network model was built with tanh(x) as the activation function. After training, the model was built using Softplus(x) to compare the accuracy of the two models. The test set is input into the trained deep sewage discharge behavior NARX neural network prediction model to complete the process of predicting the future trend of deep sewage discharge behavior, including: Obtaining time series sample data of groundwater pollutant concentration and pollution discharge data at different levels, and calculating the autocorrelation coefficient and partial correlation coefficient of the sample data; Determining the order of the time series sample data based on the autocorrelation coefficient and the partial correlation coefficient; Using the maximum likelihood estimation method, the time series sample data is accurately estimated to obtain unknown parameter values; Performing an error check on the time series sample data, and if the check result meets the preset conditions, obtaining the final time series sample data; if not, re-determining the order and re-estimating the parameters; Based on the final time series sample data, future trend prediction of deep sewage discharge behavior is carried out.
2. The groundwater pollution monitoring system according to claim 1, characterized in that: The data acquisition module includes a monitoring point distribution acquisition unit, a data signal receiver and a cloud platform; The monitoring point distribution collection unit is used to collect data from monitoring points at different levels to obtain groundwater quality data within the monitoring area; The data signal receiver is used to transmit the quality data to the cloud platform; The cloud platform is used to preprocess the quality data and store the preprocessed quality data.
3. The groundwater pollution monitoring system according to claim 2, characterized in that: The process of preprocessing the quality data includes: removing duplicates and outliers, supplementing missing values using a linear interpolation method, and unifying the time resolution of the quality data.
4. The groundwater pollution monitoring system according to claim 2, characterized in that: The data processing module includes: a slicing unit, a construction unit, a recognition model obtaining unit and a pollutant concentration obtaining unit; The slicing unit is used to slice the preprocessed quality data, and divide the sliced quality data into a training set and a prediction set; The construction unit is used to construct a convolutional neural network; The recognition model acquisition unit is used to train the convolutional neural network based on the training set, test the trained convolutional neural network based on the prediction set, and obtain a groundwater quality recognition model; The pollutant concentration obtaining unit is used to obtain groundwater pollutant concentrations at different levels based on the groundwater quality identification model.
5. The groundwater pollution monitoring system according to claim 4, characterized in that: The convolutional neural network includes 4 convolutional layers, 4 batch normalization layers, 4 ReLU layers, 2 maximum pooling layers, 1 fully connected layer and a softmax layer, and Adam is set as the solver.
6. The groundwater pollution monitoring system according to claim 1, characterized in that: The alarm module divides the indicators of deep sewage discharge behavior into the following levels: threat to human health as the primary level, threat to ecological security as the secondary level, and potential pollution as the third level, to obtain three-level groundwater pollution monitoring indicators.
