An automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds

Abstract

The present invention discloses an automatic operation monitoring system for a two-phase extraction and remediation well contaminated by volatile organic compounds. The system covers soil gas and groundwater extraction pipelines, multiple sensors, intelligent control and adaptive learning algorithm modules, etc. The sensor accurately collects data such as liquid level, pressure, tail gas concentration, water quality, etc., and transmits and processes them through the intelligent control module. The adaptive learning algorithm realizes intelligent adjustment of operating parameters through data preprocessing, feature screening, neural network modeling training and parameter optimization. The system controls submersible pumps, electric valves and other equipment in real time based on sensor data to maintain liquid level stability, adjust negative pressure, and optimize water quality monitoring and tail gas extraction. At the same time, it can remotely access the cloud platform, realize remote monitoring and control through a computer or mobile phone APP, and can automatically generate data records to assist in evaluation. It has the advantages of improving remediation efficiency, saving manpower, and intelligently optimizing operating parameters, and is suitable for the field of soil and groundwater remediation.

Description

Technical Field

[0001] The patent of this invention relates to the field of soil and groundwater remediation, and in particular to an automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds. Background Art

[0002] Two-phase extraction technology is an in-situ remediation technology that repairs contaminated soil and groundwater below the ground by simultaneously extracting soil gas in the vadose zone and contaminated groundwater (including non-aqueous phase fluid) in the aquifer. The remediation of the vadose zone soil is mainly achieved by extracting the soil gas in the vadose zone from the extraction well, extracting the volatile pollutants in the vadose zone soil to the ground for treatment, and achieving the remediation and treatment of underground soil pollution. At the same time, soil gas extraction strengthens the biological ventilation of the soil in the vadose zone and enhances the biological decomposition of organic pollutants. The remediation and enhancement of groundwater is due to the fact that vacuum extraction in the extraction well strengthens the enrichment of groundwater and non-aqueous phase pollutants in the well, which is conducive to the extraction and treatment of contaminated groundwater and non-aqueous phase pollutants.

[0003] At present, the control of two-phase extraction remediation technology in actual operation in China is relatively extensive. The remediation well group is pumped and exhausted simultaneously, with large opening and closing. There is no accurate measurement of groundwater pollution in each well, and it is impossible to timely judge the trend of the pollution degree of each remediation well with the remediation cycle, and it is impossible to timely close some low-concentration remediation wells that have met the standards. The limited exhaust vacuum equipment extraction load and sewage treatment extraction volume are timely adjusted from the low-concentration pollution area to the high-concentration pollution remediation well area. In addition, the high and low concentration remediation wells that are close to each other are operated at the same time, which may cause pollutants in the high-concentration area to transfer to the low-concentration pollution area or the qualified area, increase the remediation cycle, and affect the remediation effect. Summary of the invention

[0004] The present invention aims to address the technical deficiencies described in the background and propose an automatic operation monitoring system for two-phase extraction remediation wells contaminated by volatile organic compounds, so as to realize automatic operation monitoring of the entire remediation cycle of two-phase extraction wells contaminated by volatile organic compounds. According to the requirements of on-site management personnel, the exhaust gas extraction load of the two-phase extraction equipment and the extraction volume of contaminated groundwater can be adjusted in time from the low-concentration remediation well to the high-concentration remediation well area, so as to avoid the migration of high-concentration pollution to the low-concentration pollution area and improve the remediation efficiency.

[0005] The present invention is achieved by adopting the following technical solutions:

[0006] A volatile organic compound contaminated two-phase extraction repair well automatic operation monitoring system, characterized in that it includes a soil gas extraction pipeline, a liquid level sensor, a pressure sensor, an instrument signal circuit, an exhaust gas concentration detector, a groundwater extraction pipeline, a ground hardening layer, a repair well solid pipe section, a repair well flower pipe section, a repair well filler, groundwater in the repair well, a groundwater extraction pump control circuit, an online water quality detector, a submersible pump, an electric valve, an intelligent control module and an adaptive learning algorithm module;

[0007] The liquid level sensor is used to detect the liquid level of groundwater in the two-phase extraction well and input the liquid level signal to the PLC. The pressure sensor is used to detect the vacuum degree in the two-phase extraction well and input the pressure signal to the PLC. The tail gas concentration detector is used to detect the pollutant concentration of the extracted soil gas and transmit the tail gas concentration detection signal to the PLC control system. The online water quality detector is used to detect the organic pollution of the polluted water in the well in real time and send the detection data signal to the PLC in real time. The submersible pump is controlled to start and stop according to the program by the PLC control system, and the electric valve is controlled to open and close by the PLC control system. The intelligent control module is connected to the PLC control system, receives the data of each sensor, and analyzes and processes it according to the preset rules and algorithms to realize the whole system. Intelligent control, the adaptive learning algorithm module is integrated in the intelligent control module, including data preprocessing, feature engineering, neural network model construction and training, parameter optimization, model verification and adjustment functions; the PLC control system realizes the functions of automatic start of high liquid level extraction pump and automatic stop of low liquid level extraction pump according to the data detected by the liquid level sensor, so as to control the liquid level in the two-phase extraction well within the set fluctuation range, and at the same time judges the operating effect of the two-phase extraction equipment to extract groundwater and soil gas according to the data detected by the pressure sensor, online water quality detector and exhaust gas concentration detector, and adjusts the operating status of the two-phase extraction well group. The system can also access the cloud platform through wired and wireless networks, 4G, etc., and realize remote monitoring and control of the two-phase extraction system through a computer or mobile phone APP.

[0008] Furthermore, the data preprocessing in the adaptive learning algorithm module includes data cleaning, normalization and data segmentation. The normalization uses the formula Scale the features to the interval [0,1].

[0009] In the formula, X min and X max Represent the minimum and maximum values ​​of the feature respectively;

[0010] The data was split into 70% training set, 15% validation set, and 15% test set.

[0011] Furthermore, the feature engineering in the adaptive learning algorithm module includes calculating the correlation between the features and the target variable, using the Pearson correlation coefficient to calculate, adopting recursive feature elimination (RFE) to select features, using the logistic regression model as the base model to train the relationship between the data set and the target variable, calculating and sorting the feature weights, and removing the features with the smallest weights until the preset conditions are met.

[0012] Furthermore, the construction of the neural network model in the adaptive learning algorithm module includes determining the number of input layer neurons according to the number of features, designing the hidden layer, selecting the ReLU function as the activation function, determining the number of output layer neurons according to the number of optimization parameters, selecting the mean square error (MSE) as the loss function, selecting Adam as the optimizer, and setting the learning rate and batch size.

[0013] Furthermore, the parameter optimization in the adaptive learning algorithm module includes inputting real-time monitoring data into the trained neural network model to obtain predicted optimization parameters, adjusting the operating parameters of the dual-phase extraction well, such as the extraction rate, according to the adjustment formula, and feeding back the adjusted parameters to the PLC control system.

[0014] Furthermore, the model verification and adjustment in the adaptive learning algorithm module includes using a test set to evaluate model performance, calculating indicators such as accuracy and recall, adjusting model hyperparameters such as learning rate, batch size, and the number of hidden layer neurons based on the results, and adjusting the model structure if necessary.

[0015] Furthermore, the PLC control system determines the actual extraction negative pressure in each extraction well according to the extraction negative pressure signal collected by the pressure sensor. When the negative pressure of a single well is too low or too high, maintenance and adjustment are performed to keep the two-phase extraction well in an efficient operation range.

[0016] Furthermore, the PLC control system reduces the operating time of the well or controls the opening and closing of the groundwater extraction pump and the electric valve of the exhaust gas extraction pipeline according to the set COD concentration when the water quality COD index of one or some wells in the repair well group is reduced to the repair target value or even lower, and the extracted exhaust gas concentration is also low, based on the groundwater COD data detected by the online water quality detector.

[0017] Furthermore, the automatic operation monitoring system of the two-phase extraction and repair well of volatile organic compound pollution includes an intelligent control module and an adaptive learning algorithm module; the intelligent control module is connected to the PLC control system, receives the data of each sensor, analyzes and processes according to preset rules and algorithms, and realizes intelligent control of the entire system, including automatically adjusting the working status of the submersible pump and the electric valve, judging the system operation failure and giving maintenance suggestions, etc.; the adaptive learning algorithm module is integrated in the intelligent control module, including data preprocessing, feature engineering, neural network model construction and training, parameter optimization, model verification and adjustment functions. Through the learning and analysis of a large amount of historical data and real-time monitoring data, the prediction and optimization adjustment of the operation parameters of the repair well can be realized, and the repair efficiency and accuracy can be improved.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] (1) The present invention proposes an automatic operation monitoring system for a two-phase extraction remediation well contaminated by volatile organic compounds, which realizes automatic and accurate operation monitoring of the entire remediation cycle: through real-time monitoring and intelligent control, it can dynamically adjust the operation parameters of the remediation well according to the pollution status, significantly improve the remediation efficiency, and shorten the remediation cycle.

[0020] (2) The present invention proposes an automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds, which automatically generates data record tables and fluctuation curves, including records of changes in data such as liquid level, pressure, pollutant concentration, COD, etc. over time, providing detailed and accurate data materials for subsequent project effect evaluation, which helps to optimize remediation strategies and technical solutions.

[0021] (3) The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds proposed in the present invention supports remote monitoring and control, realizes unmanned operation on site, reduces manpower input, and avoids operators from being exposed to the contaminated environment for a long time, thereby reducing operational risks.

[0022] (4) The present invention proposes an automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds. The adaptive learning algorithm module can continuously optimize the model and prediction capabilities based on historical and real-time data, realize intelligent optimization and adjustment of operating parameters, improve resource utilization efficiency, reduce energy consumption, and adapt to changes in different pollution conditions and remediation scenarios.

[0023] (5) The present invention proposes an automatic operation monitoring system for a two-phase extraction and remediation well contaminated by volatile organic compounds. The intelligent algorithm module can perform real-time analysis on the system operation data, promptly detect abnormal conditions such as negative pressure abnormalities, sudden changes in water quality, etc., and perform fault diagnosis and early warning, so as to facilitate timely maintenance measures and ensure stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a schematic diagram of an automatic operation monitoring system for a two-phase extraction and remediation well of volatile organic compound pollution proposed by the present invention;

[0025] Figure 2 This is a technical roadmap for the automatic operation monitoring system of a two-phase extraction and remediation well contaminated by volatile organic compounds proposed by the present invention.

[0026] Among them: (1) - soil gas extraction pipeline, (2) - liquid level sensor, (3) - pressure sensor, (4) - instrument signal circuit, (5) - tail gas concentration detector, (6) - groundwater extraction pipeline, (7) - ground hardening layer, (8) - repair well solid pipe section, (9) - repair well flower pipe section, (10) - repair well filler, (11) - repair well groundwater, (12) - groundwater extraction pump control circuit, (13) - online water quality detector, (14) - submersible pump, (15) - electric valve. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe some details of 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.

[0028] Example:

[0029] An automatic operation monitoring system for a two-phase extraction and remediation well of volatile organic compound pollution, such as Figure 1 As shown, the details are as follows:

[0030] At the well repair site, complete the installation of hardware equipment according to the engineering design requirements. The soil gas extraction pipeline and the groundwater extraction pipeline must be laid accurately to ensure that the connection is tight and there is no risk of leakage. (2) The liquid level sensor is installed vertically in the repair well to ensure that its probe is completely immersed in the groundwater and away from the water inlet pipe and other components that may interfere with the liquid level measurement. It is connected to the PLC control system through the (4) instrument signal line, which can accurately detect the groundwater level and transmit the liquid level signal to the PLC control system in real time. (3) The pressure sensor is installed at a suitable position to detect the vacuum degree in the well, and its signal is also transmitted to the PLC control system through the instrument signal line. (5) The exhaust gas concentration detector and (13) the online water quality detector are installed at the corresponding positions to detect the exhaust gas pollutant concentration and groundwater COD data. The submersible pump (14) is installed at a suitable position at the bottom of the well and connected to the PLC control system through the groundwater extraction pump control line (12) so that the PLC control system can control its start and stop according to the program. The electric valve (15) is installed on the tail gas extraction pipeline and the groundwater extraction pipeline to ensure that the valve can be opened and closed flexibly and has good sealing performance. It is controlled by a PLC control system. The ground hardening layer (7) is laid as required to protect the repair well and surrounding facilities and ensure the stability of the system operation environment. The solid pipe section (8) and the flower pipe section (9) of the repair well constitute the main structure of the repair well. The flower pipe section is used for water and air intake. The repair well is filled with fillers (10) to promote the flow exchange of groundwater and soil gas.

[0031] Furthermore, after completing the installation of the parts, carefully connect the instrument signal line to ensure that the line connection is stable, the signal transmission is stable and reliable, and there is no interference and interruption, so as to realize the real-time transmission of data between the sensor and the PLC control system. At the same time, configure the intelligent control module, and set reasonable initial parameters according to the specific situation and repair target of the repair well, such as the liquid level fluctuation range (for example, set the high liquid level threshold to 5 meters and the low liquid level threshold to 1 meter, and the specific value can be adjusted according to the well depth and actual needs), pressure threshold (such as low negative pressure threshold is -0.05MPa, high negative pressure threshold is -0.08MPa), pollutant concentration limit (such as tail gas concentration high threshold is 500ppm, low threshold is 50ppm), COD target value (such as 50mg / L), etc. Perform the corresponding initialization configuration of the system adaptive learning algorithm module, set the relevant parameters of data preprocessing and the initial structure parameters of the neural network model.

[0032] Furthermore, after the equipment is put into operation, various sensors continuously collect operating data of the two-phase extraction and repair well. The liquid level sensor collects liquid level data at a certain time interval (such as every 5 minutes), the pressure sensor monitors the extraction negative pressure in real time, and the exhaust gas concentration detector and the online water quality detector also detect the exhaust gas pollutant concentration and groundwater COD data at a set frequency (such as every 10 minutes). After receiving these data, the intelligent control and algorithm module first performs data cleaning operations. For missing values ​​in the collected data, a reasonable interpolation method is used to supplement them. For obviously abnormal data points, a judgment is made based on the operating status and historical data of the repair well. If it is determined to be erroneous data, it is corrected or deleted.

[0033] Furthermore, after the data cleaning is completed, normalization is performed. For the liquid level data, if the lowest liquid level is L min =0.5m, the highest liquid level is L max =6 meters, then the normalized liquid level value is:

[0034]

[0035] Where L is the actual liquid level value.

[0036] Similarly, other characteristic data such as pressure, pollutant concentration, COD, etc. are also normalized according to the corresponding range, and all characteristic data are converted to the [0,1] interval. Finally, an adaptive learning algorithm is used to randomly split the data set into 70% training set, 15% validation set, and 15% test set, providing a reasonable data division for subsequent model training, evaluation, and optimization.

[0037] Furthermore, the adaptive learning algorithm is used to calculate the Pearson correlation coefficient between each feature and the target variable. The remediation effect index (such as pollutant removal rate) is used as the target variable to calculate the correlation coefficient between the liquid level and the pollutant removal rate. By calculating the correlation coefficient, the strength of the linear relationship between each feature and the target variable can be understood. Assuming that the correlation threshold is set to r th =0.3 (this threshold can be adjusted according to actual conditions), and the features with absolute correlation values ​​greater than 0.3 are selected as an important basis for subsequent model construction. These selected features are considered to have a strong correlation with the repair effect and can provide more valuable information for the model.

[0038] Perform recursive feature elimination (RFE) algorithm. Logistic regression model:

[0039] y=sigmoid(β0+β1x1+β2x2+.....β n x n )

[0040] In the formula, β is the model parameter.

[0041] Use the training data set and the target variable to train the model to determine the weight or coefficient of each feature. Calculate the absolute value of the weight or coefficient of each feature and sort the features by absolute value. Start with the feature with the smallest weight, remove one feature at a time, and then re-train the logistic regression model with the remaining features and evaluate the performance of the model (accuracy, mean square error, etc.). Repeat this process until the preset number of features is reached. In this way, the most representative feature combination with the greatest impact on the target variable is gradually screened out, reducing the interference of irrelevant or redundant features on the model, and improving the predictive ability and efficiency of the model.

[0042] Furthermore, the input layer of the neural network model is constructed according to the number of features determined by feature engineering. Assuming that after feature screening, the number of features is determined to be, the number of neurons in the input layer is set to n = 6. Design the hidden layer structure, for example, set two hidden layers, the number of neurons in the first layer is m1 = 10, the number of neurons in the second layer is m2 = 5, and the activation function uses the ReLU function f(x) = max(0, x) to introduce nonlinear factors and enhance the model's ability to handle complex data relationships. Determine the number of neurons in the output layer according to the number of parameters that need to be optimized.

[0043] Mean Square Error (MSE):

[0044]

[0045] In the formula, y i is the actual value, is the predicted value, and N is the number of samples.

[0046] Select MSE as the loss function, Adam as the optimizer, set the initial learning rate α0=0.001 and the batch size batch_size0=32. Use the training set data for forward propagation to calculate the predicted value, pass the input data into the neural network through the input layer, and finally get the predicted result at the output layer after calculation and transformation in the hidden layer. Calculate the loss based on the difference between the predicted value and the actual value, calculate the gradient of the loss function to the model weight through the back propagation algorithm, and then update the model weight according to the gradient descent method. After each training epoch=10 rounds, use the validation set to evaluate the model performance, and record the accuracy, mean square error and other indicators. When the performance of the model on the validation set does not improve for k=5 consecutive rounds or reaches the preset maximum number of iterations max_iter=100, stop training and get the trained neural network model.

[0047] Input the real-time monitoring data into the trained neural network model to obtain the predicted optimization parameters.

[0048] Adjust the operating parameters of the extraction well according to the predicted parameters. Assume that the current extraction rate is R old =10m 3 / h, the extraction rate adjustment value predicted by the model is ΔR = 0.2, and the adjustment coefficient is λ = 0.5, then the new extraction rate R new =10×(1-0.5×0.2)=9m 3 / h. Other operating parameters such as the negative pressure setting value are adjusted in a similar way. The adjusted parameters are fed back to the PLC control system, which controls the speed of the submersible pump, the opening degree of the electric valve, etc., to achieve real-time automatic adjustment of the operating status of the two-phase extraction well.

[0049] Furthermore, the PLC control system receives the liquid level sensor data in real time. When the liquid level is higher than the set high liquid level threshold L high =5 meters, a start signal is sent to the submersible pump through the intelligent control module or directly according to the preset logic to start extracting groundwater and lower the liquid level; when the liquid level is lower than the set low liquid level threshold L low =5 meters, a stop signal is sent to stop extraction to prevent excessive extraction of low-concentration groundwater and ensure effective extraction of soil gas in the aeration zone. For example, if the liquid level sensor detects that the liquid level reaches 5.2 meters, the PLC control system immediately starts the submersible pump; when the liquid level drops to 0.9 meters, the submersible pump is stopped.

[0050] According to the extraction negative pressure signal collected by the pressure sensor, the intelligent control module determines the actual extraction negative pressure in each extraction well. low =-0.05MPa, the intelligent algorithm module performs fault diagnosis, analyzes the cause, and prompts the staff to check the corresponding parts through the human-machine interface and make repairs and adjustments; when the negative pressure is higher than the set high negative pressure threshold P high =-0.08MPa, it may be that the well screen extraction section is blocked or the water level control parameters in the well are set too high, resulting in a small exposed depth of the aeration zone, etc., and the staff will be prompted to take corresponding measures to keep the two-phase extraction well in an efficient operation range. For example, if the pressure sensor of a well detects a negative pressure of -0.04MPa, the system will prompt the staff to check whether the extraction pipeline is leaking; if the negative pressure is -0.09MPa, it will prompt the staff to check the well screen extraction section and water level control parameters.

[0051] The online water quality detector detects the COD data of groundwater in real time, and the intelligent control module determines the change pattern of the pollution degree around each repair well with the operation time of the repair equipment based on the data. When the water quality COD index of one or some wells in the repair well group is reduced to the repair target value of 50mg / L or even lower, and the exhaust gas concentration is also lower than the set exhaust gas concentration threshold of 100ppm, the intelligent algorithm module reduces the operation time of the well according to the preset strategy. For example, if the original operation time of the well is 12 hours a day, it can be adjusted to 6 hours a day; or when the COD concentration of groundwater in the well is higher than the set COD concentration of 80mg / L, the groundwater extraction pump is started and the exhaust gas extraction pipeline electric valve is opened, and the repair well operates normally; when the COD concentration is lower than the set concentration of 30mg / L, the groundwater extraction pump and the electric valve are turned off to reduce the load of the repair equipment occupied by the low-concentration repair well area, strengthen the exhaust gas and groundwater extraction capacity in the high-concentration repair well area, and improve the repair efficiency. For example, if the COD concentration of a well drops to 25 mg / L and the tail gas concentration is 80 ppm, the system will shut down the groundwater extraction pump and electric valve of the well.

[0052] The tail gas concentration detector detects the tail gas concentration of the extraction pipeline of the two-phase extraction repair well and transmits the tail gas concentration detection signal to the PLC control system. The intelligent control module judges the extraction effect of soil pollution in the aeration zone of the two-phase extraction well based on this. If the tail gas concentration continues to be higher than the set high concentration threshold of 500ppm, it may indicate that the soil pollution is serious or the extraction effect is poor. The intelligent algorithm module can adjust the extraction parameters, such as increasing the extraction rate or extending the extraction time, to ensure the effectiveness of the repair work; if the tail gas concentration is lower than the set low concentration threshold of 50ppm and continues for a period of time, the extraction intensity can be appropriately reduced to save energy. For example, if the tail gas concentration of a well is higher than 500ppm for three consecutive days, the system automatically increases the extraction rate by 20%; if the tail gas concentration is lower than 50ppm for a week, the system automatically reduces the extraction rate by 10%.

[0053] Furthermore, the system is connected to the cloud platform through wired and wireless networks, 4G and other communication methods. Users can log in to the cloud platform account through a computer or mobile phone APP to remotely monitor and control the two-phase extraction system. Remote users can view system operation data, status curves and other information in real time, including liquid level, pressure, pollutant concentration, COD data, etc., to fully understand the operating status of the repair well. For example, users can view the liquid level change curve in the past 24 hours on the mobile phone APP to intuitively understand the water level fluctuations. They can also view the changing trend of tail gas concentration in different time periods and analyze the repair effect.

[0054] Users can remotely issue control instructions, such as adjusting the operating parameters of the submersible pump, the switch status of the electric valve, etc. If a user finds that the liquid level in a well rises too fast through a mobile phone APP, he can remotely manually start the submersible pump of the well for early extraction; or remotely close the electric valves of some low-concentration repair wells according to the repair progress to stop their operation. After receiving the remote instructions, the intelligent control module performs the corresponding operations and feeds back the execution results to the remote user, realizing unmanned on-site operation and remote expert guidance, saving labor costs, improving the operating environment, and improving the repair effect of the two-phase extraction equipment.

[0055] Furthermore, the system operation data is continuously collected, and the new data is used to optimize the adaptive learning algorithm module. Every month, the neural network model is retrained with a new data set, the model parameters are updated, and the prediction accuracy and adaptability of the model are improved. During the retraining process, the initial parameters of the system (such as liquid level threshold, pressure threshold, COD target value, etc.) are adjusted according to the actual repair effect and operation status. For example, if it is found that the liquid level fluctuation range of the repair well in a certain area has changed, the liquid level threshold can be adjusted appropriately; if the exhaust gas concentration is reduced overall, the exhaust gas concentration threshold can be lowered accordingly to optimize the repair strategy, adapt to different pollution conditions and repair requirements, and ensure that the system is always in an efficient operation state.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds, characterized in that: It includes soil gas extraction pipeline, liquid level sensor, pressure sensor, instrument signal circuit, tail gas concentration detector, groundwater extraction pipeline, ground hardening layer, solid pipe section of repair well, flower pipe section of repair well, filler of repair well, groundwater in repair well, groundwater extraction pump control circuit, online water quality detector, submersible pump, electric valve, intelligent control module and adaptive learning algorithm module; The liquid level sensor is used to detect the liquid level of groundwater in the two-phase extraction well and input the liquid level signal to the PLC. The pressure sensor is used to detect the vacuum degree in the two-phase extraction well and input the pressure signal to the PLC. The tail gas concentration detector is used to detect the pollutant concentration of the extracted soil gas and transmit the tail gas concentration detection signal to the PLC control system. The online water quality detector is used to detect the organic pollution of the polluted water in the well in real time and send the detection data signal to the PLC in real time. The submersible pump is controlled to start and stop according to the program by the PLC control system, and the electric valve is controlled to open and close by the PLC control system. The intelligent control module is connected to the PLC control system, receives the data of each sensor, and analyzes and processes it according to the preset rules and algorithms to realize the whole system. Intelligent control, the adaptive learning algorithm module is integrated in the intelligent control module, including data preprocessing, feature engineering, neural network model construction and training, parameter optimization, model verification and adjustment functions; the PLC control system realizes the functions of automatic start of high liquid level extraction pump and automatic stop of low liquid level extraction pump according to the data detected by the liquid level sensor, so as to control the liquid level in the two-phase extraction well within the set fluctuation range, and at the same time judges the operating effect of the two-phase extraction equipment to extract groundwater and soil gas according to the data detected by the pressure sensor, online water quality detector and exhaust gas concentration detector, and adjusts the operating status of the two-phase extraction well group. The system can also access the cloud platform through wired and wireless networks, 4G, etc., and realize remote monitoring and control of the two-phase extraction system through a computer or mobile phone APP.

2. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The data preprocessing in the adaptive learning algorithm module includes data cleaning, normalization and data segmentation. Normalization scales the features to the interval [0,1], and data segmentation is performed in the ratio of 70% training set, 15% validation set, and 15% test set.

3. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The feature engineering in the adaptive learning algorithm module includes calculating the correlation between features and target variables, using Pearson correlation coefficient calculation, adopting recursive feature elimination (RFE) for feature selection, using the logistic regression model as the base model to train the relationship between the data set and the target variable, calculating and sorting feature weights, and removing the features with the smallest weight until the preset conditions are met.

4. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The construction of the neural network model in the adaptive learning algorithm module includes determining the number of input layer neurons according to the number of features, designing the hidden layer, selecting the ReLU f(x)=max(0,x) function as the activation function, determining the number of output layer neurons according to the number of optimized parameters, and selecting the mean square error MSE as the loss function: In the formula, y i is the actual value, is the predicted value, N is the number of samples; The optimizer selects Adam, sets the learning rate α0 and the batch size batch_size0.

5. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The parameter optimization in the adaptive learning algorithm module includes inputting real-time monitoring data into the trained neural network model to obtain predicted optimization parameters, adjusting the operating parameters of the dual-phase extraction well, such as the extraction rate, and feeding back the adjusted parameters to the PLC control system.

6. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The model verification and adjustment in the adaptive learning algorithm module includes using the test set to evaluate the model performance, calculating the accuracy, recall and other indicators, adjusting the model hyperparameters such as learning rate, batch size, number of hidden layer neurons according to the results, and adjusting the model structure when necessary.

7. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The PLC control system determines the actual extraction negative pressure in each extraction well according to the extraction negative pressure signal collected by the pressure sensor. When the negative pressure of a single well is too low or too high, maintenance and adjustment are performed to keep the two-phase extraction well in an efficient operation range.

8. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: The PLC control system reduces the operating time of a well based on the groundwater COD data detected by the online water quality detector, when the water quality COD index of one or some wells in the repair well group is reduced to the repair target value or even lower, and the extracted exhaust gas concentration is also low, and controls the opening and closing of the groundwater extraction pump and the exhaust gas extraction pipeline electric valve according to the set COD concentration.

9. The automatic operation monitoring system for two-phase extraction and remediation wells contaminated by volatile organic compounds according to claim 1 is characterized in that: It includes an intelligent control module and an adaptive learning algorithm module; the intelligent control module is connected to the PLC control system, receives data from various sensors, analyzes and processes data based on preset rules and algorithms, and realizes intelligent control of the entire system, including automatically adjusting the working status of the submersible pump and electric valve, judging system operation failures and giving maintenance suggestions, etc. The adaptive learning algorithm module is integrated in the intelligent control module, which includes functions such as data preprocessing, feature engineering, neural network model construction and training, parameter optimization, model verification and adjustment. Through learning and analysis of a large amount of historical data and real-time monitoring data, it can predict and optimize the operating parameters of the repair wells, thereby improving the efficiency and accuracy of the repair.

Images