Underground water pollution early warning and evaluation method suitable for oil and gas field produced water reinjection process

Abstract

This invention discloses a groundwater pollution early warning and assessment method applicable to the produced water reinjection process in oil and gas fields, belonging to the field of groundwater environmental protection technology. It collects shallow groundwater quality indicators and deep microseismic monitoring data through shallow monitoring wells and microseismic monitoring stations deployed around the reinjection wells. A system is constructed comprising three categories of indicators: shallow conventional water analysis, microseismic monitoring, and hydrogeological data. An ARIMA model is used to dynamically set thresholds, and an improved AHP-LightGBM model is used to calculate the deep formation stability and wellbore integrity indices. A fuzzy comprehensive evaluation method is used to calculate the shallow pollution risk value. A comprehensive risk value is obtained through weighted fusion, and four early warning levels are established, along with differentiated response measures. This invention achieves collaborative monitoring of shallow water quality and deep structure, improves risk assessment accuracy through multi-algorithm fusion, provides technical support for the prevention and control of groundwater pollution in deep produced water reinjection, and is applicable to groundwater environmental safety assurance scenarios in produced water reinjection operations.

Description

Technical Field

[0001] This invention relates to the field of groundwater environmental protection technology, and is applicable to the deep treatment of produced water in the oil and gas field industry. It establishes an assessment method for groundwater pollution early warning system during the deep reinjection of produced water. Background Technology

[0002] In the oil and gas field development industry, deep reinjection of produced water has become a key means of balancing environmental protection and resource utilization. However, the risk of pollutant leakage during the reinjection process poses a serious threat to surrounding shallow and deep groundwater. Currently, the field of groundwater monitoring in reinjection wells presents a situation where shallow monitoring methods lack timeliness, deep monitoring technologies face bottlenecks, and wellbore monitoring accuracy is limited. Shallow groundwater monitoring often employs a combination of manual sampling and conventional water analysis, while deep groundwater, due to its great depth and complex geological conditions, is difficult for conventional monitoring equipment to access. Wellbore integrity monitoring relies on traditional technologies such as pressure testing, which cannot capture the development of minute fractures in real time, making early risk warning difficult.

[0003] Chinese patent (publication number CN113176791A) discloses a groundwater pollution prevention and control system for contaminated sites. The patent proposes to collect water quality indicators by deploying shallow monitoring wells and combine them with reinjection pressure data to judge the pollution risk. However, it only focuses on shallow groundwater and reinjection pressure and does not involve the system monitoring of deep groundwater and well integrity.

[0004] Chinese patent (publication number CN116403092A) discloses a method and system for determining the degree of groundwater NAPL pollution based on image learning. This patent integrates parameters such as shallow water quality and groundwater flow velocity, and uses deep learning to construct a groundwater NAPL pollution degree determination model. It is mainly aimed at shallow groundwater pollution early warning, but does not involve deep groundwater monitoring, does not consider the fusion application of microseismic monitoring data, and the model construction is simplistic, with insufficient risk assessment accuracy and dynamic adaptability.

[0005] Against this backdrop, the development of a groundwater pollution early warning system for reinjection wells that enables multi-dimensional monitoring, high timeliness, high precision assessment, and strong operability has become an urgent need for the industry. Summary of the Invention

[0006] Based on the aforementioned technical problems, this application discloses a groundwater pollution early warning and assessment method applicable to the produced water reinjection process in oil and gas fields, specifically as follows:

[0007] By deploying shallow monitoring wells around the reinjection wells, conventional water analysis equipment is used to monitor shallow groundwater in real time and collect water quality indicators.

[0008] By deploying microseismic monitoring stations around the reinjection wells, microseismic monitoring technology is used to continuously monitor deep groundwater and well integrity, and collect monitoring data.

[0009] Based on the collected data, an index system was established, including shallow conventional water analysis index layer, microseismic monitoring index layer and hydrogeological index layer, and normal thresholds for each index were set.

[0010] Based on the collected data, an improved AHP-LightGBM model for the correlation between microseismic events and risks was constructed by combining the Analytic Hierarchy Process (AHP) with the LightGBM gradient boosting tree algorithm. The deep formation stability index and wellbore integrity index were calculated respectively, and the model prediction error was verified using an incremental dataset. The shallow contamination risk value was calculated using the fuzzy comprehensive evaluation method.

[0011] By weighted fusion of the shallow contamination risk value, deep formation stability index, and wellbore integrity index, a comprehensive risk value is obtained. Based on the comprehensive risk value, four warning levels are established, and corresponding countermeasures are formulated.

[0012] Preferably, the data establishment index system is specifically as follows: the shallow monitoring wells are deployed within a range of 50 meters to 2000 meters around the reinjection wells, and include multiple monitoring wells to form a ring monitoring network; the microseismic monitoring stations are deployed within a range of 1 kilometer to 3 kilometers around the reinjection wells, and include multiple monitoring stations to ensure three-dimensional coverage of deep groundwater and well integrity; the index system is established based on conventional water analysis indicators collected from shallow layers and microseismic monitoring indicators collected from deep layers.

[0013] Preferably, the specific composition of the index system is as follows: the shallow conventional water analysis index layer includes: chloride, sulfide, petroleum hydrocarbons, strontium, pH, bromide, nitrate, sulfate, ammonia nitrogen, hexavalent chromium, total hardness, iron, manganese, and permanganate index; the microseismic monitoring index layer includes microseismic event frequency, average magnitude, b-value, event spatial clustering, and minimum distance from the wellbore; the hydrogeological index layer includes aquifer thickness, permeability coefficient, and groundwater flow direction.

[0014] Preferably, the normal threshold and the weight coefficient of the indicator system are specifically as follows: based on historical data, the threshold and weight coefficient are dynamically adjusted through a time series analysis algorithm. The time series analysis algorithm adopts the autoregressive integral moving average (ARIMA) model, and the formula is:

[0015]

[0016] in, For the shift operator, and It is a polynomial. Let be the difference order. For time series data, It is white noise.

[0017] Preferably, the construction of the microseismic event-risk correlation model using the Analytic Hierarchy Process (AHP) combined with the LightGBM gradient boosting tree algorithm specifically involves:

[0018] Based on microseismic monitoring data, characteristic parameters are extracted, including event frequency, magnitude, b-value, spatial clustering, and distance from the wellbore.

[0019] The weights of each feature parameter are determined using the Analytic Hierarchy Process (AHP). The AHP weights are calculated by constructing a judgment matrix and solving for the eigenvectors. The formula is as follows:

[0020]

[0021] in, To determine the elements of a matrix. To represent the importance of the i-th parameter relative to the j-th parameter, a scaling method is used. The weight vector is obtained by solving the eigenvalue problem. It is the largest eigenvalue;

[0022] Using feature parameters and AHP weights as input, a LightGBM gradient boosting tree algorithm model is trained to predict deep formation stability index and wellbore integrity index. The objective function is:

[0023]

[0024] in Let be the objective function. For loss function, This is a regularization term used to prevent overfitting; it calculates the prediction error of the trained model and triggers model updates based on the prediction error.

[0025] Preferably, the step of triggering model updates based on prediction errors specifically involves: using the incremental dataset D to perform performance validation on the existing improved AHP-LightGBM model for the association between microseismic events and risks, and calculating the model prediction error rate. The formula is:

[0026]

[0027] in, The number of samples in the incremental dataset. These are actual monitoring values, including measured values ​​of deep formation stability index and measured values ​​of shallow contamination risk. These are the model's predicted values;

[0028] When the prediction error rate When the value is too large, a model update is triggered, the feature parameter weights are readjusted using AHP, and an incremental dataset is used. Dataset merged with historical datasets Retrain the LightGBM model.

[0029] Preferably, the calculation of shallow contamination risk value using the fuzzy comprehensive evaluation method specifically includes:

[0030] Establish factor set ,in Indicates various water quality indicators;

[0031] Create a collection of comments ,in The risk level is indicated by categories such as low risk, medium risk, and high risk.

[0032] Construct a membership matrix M, where each element Indicators This is a comment. The degree of membership is calculated using a trapezoidal function.

[0033] Determine the weight vector W using the entropy weighting method or AHP;

[0034] Fuzzy synthesis is performed to calculate the shallow contamination risk value P. The specific formula is as follows:

[0035]

[0036] in, This represents the risk value for shallow contamination. This represents a fuzzy synthesis operator, using a weighted average model.

[0037] Preferably, the step of basing the comprehensive risk value The four-level early warning system is divided as follows: the comprehensive risk value is calculated using the following formula:

[0038]

[0039] in, , This represents the risk value for shallow contamination. This is a stability index for deep strata. The wellbore integrity index. , , These are the weighting coefficients;

[0040] The warning levels are divided into four levels based on the comprehensive risk value Rtotal, and three risk level thresholds are preset. And formulate corresponding countermeasures.

[0041] Preferably, the formulation of corresponding countermeasures specifically involves: based on risk level thresholds. , , The four warning levels are divided into four levels, and different measures are taken for each level:

[0042] The comprehensive risk value for Level 1 early warning is 0- The shallow indicators are normal, the deep microseismic events are normal, and the wellbore is intact. Measures to be taken: maintain the current monitoring frequency and submit an assessment report every quarter.

[0043] Level II Early Warning Comprehensive Risk Value - The following measures were taken: increase the monitoring frequency of shallow layers, strengthen the analysis of microseismic data, submit daily reports, and investigate the parameters of the reinjection operation.

[0044] Level 3 Early Warning Comprehensive Risk Value - If the shallow core indicators are close to the threshold, or the frequency and magnitude of deep microseismic events increase sharply, with 3-5 microseismic events per week around the well, the following measures should be taken: suspend some reinjection operations, initiate emergency monitoring, and organize experts to formulate prevention and control plans.

[0045] Level IV Early Warning Comprehensive Risk Value -10 points indicates that shallow indicators exceed the standard, or that deep microseismic events occur frequently and with high magnitudes, with more than 5 microseismic events per week around the wellbore. Measures to be taken include: immediately stopping reinjection, initiating emergency response, carrying out pollution control and wellbore repair, and reporting to the environmental protection department.

[0046] Compared with the prior art, the technical solution of this application has the following technical effects:

[0047] This invention achieves precision through stratified monitoring. For shallow layers, conventional water analysis is used to monitor water quality in real time; for deeper layers and wellbore, microseismic technology is used, breaking through the bottleneck of deep monitoring technology and accurately capturing anomalies in the formation and wellbore.

[0048] This invention integrates the risks of shallow contamination, deep formations, and wellbore integrity to establish a collaborative early warning system, solving the fragmentation problem of existing methods and achieving comprehensive risk early warning.

[0049] This invention utilizes a multi-algorithm collaboration approach. For shallow contamination risk assessment, it employs a fuzzy comprehensive evaluation method, combined with AHP to determine weights, and quantifies the risk contribution of each water quality indicator through a membership matrix. For deep contamination and wellbore risk assessment, it adopts an improved AHP-LightGBM model. First, it determines the weights of microseismic characteristic parameters through AHP, and then inputs them into the LightGBM model for training, thereby improving timeliness and accuracy.

[0050] This invention automatically completes data collection, analysis, and early warning through the integration of supporting data and platforms. It supports remote transmission, is suitable for rapid operation by on-site personnel, and is highly practical.

[0051] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the preferred embodiments of this application are described in detail below with reference to the accompanying drawings.

[0052] The above and other objects, advantages and features of this application will become more apparent to those skilled in the art from the following detailed description of specific embodiments in conjunction with the accompanying drawings. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0054] Based on the description of the figures and their corresponding technical content in the document, the titles of the figures are as follows:

[0055] Figure 1 A flowchart of a groundwater pollution early warning assessment method applicable to the produced water reinjection process in oil and gas fields;

[0056] Figure 2 A general framework diagram for a groundwater pollution early warning and assessment method applicable to the produced water reinjection process in oil and gas fields;

[0057] Figure 3 A schematic diagram showing the arrangement of substations along the dominant axis of microseismic events;

[0058] Figure 4 The neural network structure diagram of the improved AHP-LightGBM model;

[0059] Figure 5 Experimental architecture diagram for testing this method in the T7 reinjection well of the natural gas field;

[0060] Figure 6 This is a comparison chart of abnormal warning time data of the three methods in the embodiments of this application in the experiment;

[0061] Figure 7 This is a comparison chart of pH data from experiments for the three methods described in the embodiments of this application.

[0062] Figure 8 This is a comparison chart of potassium permanganate data in experiments for the three methods described in the embodiments of this application. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. In the following description, specific details such as specific configurations and components are provided merely to help fully understand the embodiments of this application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. In addition, for clarity and brevity, descriptions of known functions and structures are omitted in the embodiments.

[0064] It should be understood that the phrase "an embodiment" or "this embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "an embodiment" or "this embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0065] Furthermore, reference numerals and / or letters may be repeated in different examples within this application. Such repetition is for the purpose of simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or settings discussed.

[0066] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" in this article describes another type of relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the related objects before and after it are in an "or" relationship.

[0067] In this article, the term "at least one" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, "at least one of A and B" can mean: A exists alone, A and B exist simultaneously, or B exists alone.

[0068] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion.

[0069] Example 1 describes a groundwater pollution early warning assessment method applicable to the produced water reinjection process in oil and gas fields, such as... Figure 1 - Figure 2 As shown, it specifically includes:

[0070] By deploying shallow monitoring wells around the reinjection wells, conventional water analysis equipment is used to monitor shallow groundwater in real time and collect water quality indicators.

[0071] By deploying microseismic monitoring stations around the reinjection wells, microseismic monitoring technology is used to continuously monitor deep groundwater and well integrity, and collect monitoring data.

[0072] Based on the collected data, an index system was established, including shallow conventional water analysis index layer, microseismic monitoring index layer and hydrogeological index layer, and normal thresholds for each index were set.

[0073] Based on the collected data, an improved AHP-LightGBM model for the correlation between microseismic events and risks was constructed by combining the Analytic Hierarchy Process (AHP) with the LightGBM gradient boosting tree algorithm. The deep formation stability index and wellbore integrity index were calculated respectively, and the model prediction error was verified using an incremental dataset. The shallow contamination risk value was calculated using the fuzzy comprehensive evaluation method.

[0074] By weighted fusion of the shallow contamination risk value, deep formation stability index, and wellbore integrity index, a comprehensive risk value is obtained. Based on the comprehensive risk value, four warning levels are established, and corresponding countermeasures are formulated.

[0075] Furthermore, the data establishment indicator system specifically includes: shallow monitoring wells are deployed within a range of 50 to 2000 meters around the reinjection wells, and multiple monitoring wells are included to form a ring monitoring network; microseismic monitoring stations are deployed within a range of 1 to 3 kilometers around the reinjection wells, and multiple monitoring stations are included to ensure three-dimensional coverage of deep groundwater and well integrity; an indicator system is established based on conventional water analysis indicators collected from shallow layers and microseismic monitoring indicators collected from deep layers; the hydrogeological indicator layers include water layer thickness, permeability coefficient, and groundwater flow direction.

[0076] Furthermore, such as Figure 3 As shown, the monitoring system is configured as follows: with the wellhead as the center, a monitoring radius of 400m is set to monitor the background conditions and normal water injection conditions of each reinjection layer and non-reinjection layer. The geophones are deployed with the wellhead as the center, and each substation is about 50m to 100m apart.

[0077] The first monitoring center was located 750m from the wellhead in the direction of the dominant water flow. Twenty-four substations were deployed along this axis in three rows, with a spacing of 50m to 100m between each substation. The second monitoring was conducted with the same center located 750m from the first monitoring center in the same direction. Again, 24 substations were deployed along this axis in three rows, with a spacing of 50m to 100m between each substation, until the monitoring reached the leading edge of the water body.

[0078] Additional monitoring will be conducted in areas outside the direction of the dominant water body to confirm whether the reinjected water has reached the leading edge position. The deployment method will be consistent with the water direction dominance tracking and monitoring plan.

[0079] Furthermore, microseismic monitoring technology is used to monitor the source location, magnitude, and occurrence time of microseismic events. The distribution of events is used to determine the development of deep strata fractures and the migration channels of pollutants. Microseismic events around the wellbore are used to identify integrity issues such as wellbore damage and cement sheath failure.

[0080] Furthermore, the construction of the indicator system is specifically as follows:

[0081] Shallow water analysis indicators reflect the pollution status of shallow water. Exceeding the standards of these indicators suggests a pollution risk. These indicators include pH (6.5-8.5), chloride (≤250mg / L), sulfide (≤0.05mg / L), petroleum hydrocarbons (≤0.05mg / L), strontium (≤0.2mg / L), bromide (≤0.1mg / L), nitrate (≤10mg / L), sulfate (≤250mg / L), ammonia nitrogen (≤1.0mg / L), hexavalent chromium (≤0.05mg / L), total hardness (≤450mg / L), iron (≤0.3mg / L), manganese (≤0.1mg / L), and permanganate index (≤6mg / L).

[0082] The microseismic monitoring index layer reflects the integrity of the wellbore. An abnormal increase indicates a failure of the wellbore integrity. This includes the frequency of deep microseismic events (normal ≤ 5 times / day), focal depth (100-4500 meters), and magnitude (normal ≤ 0). A sudden increase in frequency or magnitude indicates the development of deep strata fractures. The number of microseismic events around the wellbore (normal ≤ 2 times / week) is also considered.

[0083] The reinjection process and hydrogeological indicators help determine the sources of risk, including reinjection pressure (not exceeding formation fracturing pressure), reinjection volume fluctuation (≤10%), shallow groundwater flow velocity (0.1-1m / d), and deep formation permeability coefficient.

[0084] Furthermore, the normal threshold and weight coefficient of the indicator system are specifically as follows: based on historical data, the threshold and weight coefficient are dynamically adjusted through a time series analysis algorithm. The time series analysis algorithm adopts the autoregressive integral moving average (ARIMA) model, and the formula is:

[0085]

[0086] in, For the shift operator, and It is a polynomial. Let be the difference order. For time series data, It is white noise.

[0087] Furthermore, such as Figure 4 As shown, the microseismic event-risk correlation model constructed by combining the Analytic Hierarchy Process (AHP) with the LightGBM gradient boosting tree algorithm is as follows:

[0088] Based on microseismic monitoring data, characteristic parameters are extracted, including event frequency, magnitude, b-value, spatial clustering, and distance from the wellbore.

[0089] The weights of each feature parameter are determined using the Analytic Hierarchy Process (AHP). The AHP weights are calculated by constructing a judgment matrix and solving for the eigenvectors. The formula is as follows:

[0090]

[0091] in, To determine the elements of a matrix. To represent the importance of the i-th parameter relative to the j-th parameter, a scaling method is used. The weight vector is obtained by solving the eigenvalue problem. It is the largest eigenvalue;

[0092] Using feature parameters and AHP weights as input, a LightGBM gradient boosting tree algorithm model is trained to predict deep formation stability index and wellbore integrity index. The objective function is:

[0093]

[0094] in Let be the objective function. For loss function, This is a regularization term used to prevent overfitting; it calculates the prediction error of the trained model and triggers model updates based on the prediction error.

[0095] Furthermore, the step of triggering model updates based on prediction errors specifically involves: using the incremental dataset D to perform performance validation on the improved AHP-LightGBM model that correlates existing microseismic events with risks, and calculating the model prediction error rate. The formula is:

[0096]

[0097] in, The number of samples in the incremental dataset. These are actual monitoring values, including measured values ​​of deep formation stability index and measured values ​​of shallow contamination risk. These are the model's predicted values;

[0098] When the prediction error rate When the value is too large, a model update is triggered, the feature parameter weights are readjusted using AHP, and an incremental dataset is used. Dataset merged with historical datasets Retrain the LightGBM model.

[0099] Furthermore, the calculation of shallow contamination risk values ​​using the fuzzy comprehensive evaluation method specifically includes:

[0100] Establish factor set ,in Indicates various water quality indicators;

[0101] Create a collection of comments ,in The risk level is indicated by categories such as low risk, medium risk, and high risk.

[0102] Construct a membership matrix M, where each element Indicators This is a comment. The degree of membership is calculated using a trapezoidal function.

[0103] Determine the weight vector W using the entropy weighting method or AHP;

[0104] Fuzzy synthesis is performed to calculate the shallow contamination risk value P. The specific formula is as follows:

[0105]

[0106] in, This represents the risk value for shallow contamination. This represents a fuzzy synthesis operator, using a weighted average model.

[0107] Furthermore, the statement based on the comprehensive risk value The four-level early warning system is divided as follows: the comprehensive risk value is calculated using the following formula:

[0108]

[0109] in, , This represents the risk value for shallow contamination. This is a stability index for deep strata. The wellbore integrity index. , , These are the weighting coefficients;

[0110] The warning levels are divided into four levels based on the comprehensive risk value Rtotal, and three risk level thresholds are preset. And formulate corresponding countermeasures.

[0111] Furthermore, the specific measures to formulate corresponding responses are as follows: based on risk level thresholds... , , The four warning levels are divided into four levels, and different measures are taken for each level:

[0112] The comprehensive risk value for Level 1 early warning is 0- The shallow indicators are normal, the deep microseismic events are normal, and the wellbore is intact. Measures to be taken: maintain the current monitoring frequency and submit an assessment report every quarter.

[0113] Level II Early Warning Comprehensive Risk Value - The following measures were taken: increase the monitoring frequency of shallow layers, strengthen the analysis of microseismic data, submit daily reports, and investigate the parameters of the reinjection operation.

[0114] Level 3 Early Warning Comprehensive Risk Value - If the shallow core indicators are close to the threshold, or the frequency and magnitude of deep microseismic events increase sharply, with 3-5 microseismic events per week around the well, the following measures should be taken: suspend some reinjection operations, initiate emergency monitoring, and organize experts to formulate prevention and control plans.

[0115] Level IV Early Warning Comprehensive Risk Value -10 points indicates that shallow indicators exceed the standard, or that deep microseismic events occur frequently and with high magnitudes, with more than 5 microseismic events per week around the wellbore. Measures to be taken include: immediately stopping reinjection, initiating emergency response, carrying out pollution control and wellbore repair, and reporting to the environmental protection department.

[0116] This implementation details a groundwater pollution early warning and assessment method applicable to the produced water reinjection process in oil and gas fields. Shallow monitoring wells and microseismic monitoring stations deployed around the reinjection wells collect shallow groundwater quality indicators and deep microseismic monitoring data, respectively. A system is constructed comprising three categories of indicators: shallow conventional water analysis, microseismic monitoring, and hydrogeological data. The ARIMA model is used to dynamically set thresholds, and the deep formation stability and wellbore integrity indices are calculated based on the AHP-LightGBM improved model. The fuzzy comprehensive evaluation method is used to calculate shallow pollution risk values. A weighted fusion is used to obtain a comprehensive risk value, which is then divided into four early warning levels, and differentiated response measures are formulated.

[0117] Example 2 details the experiment conducted on the T71 well, a deep reinjection well for produced water from a natural gas well, to test this method. The reinjection depth of this well is 2394m–2415m, the reinjection pressure is 0–5MPa, the average daily reinjection volume is 30–200m³, the surrounding area is hilly, and the shallow groundwater (0–100m) is the drinking water source for the surrounding residents. The hydrogeological conditions are clastic rock aquifers with a shallow permeability coefficient of 0.3–0.8m / d and a deep permeability coefficient of 0.005–0.01m / d. The specific experimental implementation is as follows:

[0118] like Figure 5 As shown, three shallow monitoring wells, numbered S1-S3, were deployed within a range of 0.5km, 1km, and 2km to monitor background values, midstream and downstream data, respectively. A HACHHQ40d water quality analyzer was installed to monitor chloride, sulfide, petroleum hydrocarbons, strontium, pH, bromide, nitrate, sulfate, ammonia nitrogen, hexavalent chromium, total hardness, iron, manganese, and permanganate index. These indicators were collected once per quarter.

[0119] Five monitoring stations, numbered M1-M5, were set up within a range of 1-3km and equipped with GS-20DX three-component geophones. Two near-wellpoints, M6 and M7, were set up around the well. Monitoring was conducted at a sampling rate of 2000Hz and continuous data was collected for 24 hours. After removing industrial noise, the effective microseismic events averaged 4 per week, and the events around the well averaged 1 per week.

[0120] The risk of shallow water pollution was calculated using the fuzzy comprehensive evaluation method. Based on the current water quality, a factor set U and a comment set V were constructed. The key influencing factors in factor set U, namely pH (7.2-7.4), potassium permanganate (2.5-4.5 mg / L), chloride ions (≤250 mg / L), petroleum (≤0.05 mg / L), and strontium (≤0.2 mg / L), were selected as representative water quality data to be recorded. The membership matrix was calculated using the trapezoidal membership function, and the weights were determined using the entropy weight method. Finally, the weekly average shallow water pollution risk value was obtained as 2.0-2.4 points.

[0121] The deep formation and wellbore indices were calculated using an improved AHP-LightGBM model. The microseismic feature weights were determined using AHP and input into the LightGBM model for training. The resulting mean deep formation stability index was 0.88-0.92 and the wellbore integrity index was 0.93-0.97.

[0122] The weighted fusion yielded a comprehensive risk score of 2.5, which was classified as a Level 1 warning. The experimental data shown in Table 1 below were obtained. The measures taken were to maintain the existing monitoring frequency, submit an assessment report every quarter, continuously track data changes, and ensure the safety of reinjection operations and groundwater.

[0123] Table 1 shows the experimental results of testing reinjection wells.

[0124] A comparative experiment was conducted using the existing common methods SM+IPM and SMSM monitoring. The SM+IPM method collected water samples from wells S1-S3 every 15 days and tested water quality indicators in the laboratory, while recording the reinjection pressure every hour. The SMSM monitoring method collected microseismic data from near-wellpoints M6-M7 with the same acquisition parameters as the method, but only counted the frequency of microseismic events around the wellbore. The experimental comparison data are shown in Table 2 below.

[0125] Table 2 Experimental Comparison Data

[0126] As shown in Tables 1 and 2, this method achieves full-dimensional monitoring of shallow, deep, and wellbore layers, and is significantly superior to the other two existing methods in terms of data acquisition dimensions. Moreover, the abnormal warning time is ≤15 minutes, and the pH / potassium permanganate data deviation rate is low, which verifies the comprehensiveness, timeliness, accuracy and practicality of this method.

[0127] according to Figure 6 — Figure 8It can be seen that the average time of abnormal warning is lower than that of the SM+IPM method and the SMSM monitoring method. Moreover, the data of pH and potassium permanganate monitoring are more concentrated, and the data deviation rates of pH and potassium permanganate are as low as 1.27% and 1.34%, respectively, which are significantly better than the other two methods. This verifies the advantages of this method in terms of timeliness, practicality, accuracy and precision.

[0128] This embodiment describes in detail the experiment of testing this method in a natural gas produced water deep reinjection well T71. The experimental results fully verify that this method achieves low data deviation rate, high timeliness and practicality through full-dimensional monitoring of shallow, deep and wellbore layers and multi-algorithm collaboration, proving the outstanding advantages of this method in comprehensive monitoring, fast response and high accuracy.

[0129] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any changes, modifications, substitutions, integrations, and parameter changes made to these embodiments within the spirit and principles of the present invention, without departing from the principles and spirit of the present invention, through conventional substitutions or to achieve the same function, fall within the scope of protection of the present invention.

Claims

1. A groundwater pollution early warning and assessment method applicable to the produced water reinjection process in oil and gas fields, characterized in that, include: By deploying shallow monitoring wells around the reinjection wells, conventional water analysis equipment is used to monitor shallow groundwater in real time and collect water quality indicators. By deploying microseismic monitoring stations around the reinjection wells, microseismic monitoring technology is used to continuously monitor deep groundwater and well integrity, and collect monitoring data. Based on the collected data, an index system was established, including shallow conventional water analysis index layer, microseismic monitoring index layer and hydrogeological index layer, and normal thresholds for each index were set. Based on the collected data, an improved AHP-LightGBM model for the correlation between microseismic events and risks was constructed by combining the Analytic Hierarchy Process (AHP) with the LightGBM gradient boosting tree algorithm. The deep formation stability index and wellbore integrity index were calculated respectively, and the model prediction error was verified using an incremental dataset. The risk value of shallow contamination is calculated using the fuzzy comprehensive evaluation method. By weighted fusion of the shallow contamination risk value, deep formation stability index, and wellbore integrity index, a comprehensive risk value is obtained. Based on the comprehensive risk value, four warning levels are established, and corresponding countermeasures are formulated.

2. The method according to claim 1, characterized in that, The data establishes an indicator system, specifically as follows: shallow monitoring wells are deployed within a range of 50 to 2000 meters around the reinjection wells, and include multiple monitoring wells to form a ring monitoring network; microseismic monitoring stations are deployed within a range of 1 to 3 kilometers around the reinjection wells, and include multiple monitoring stations to ensure three-dimensional coverage of deep groundwater and well integrity; the indicator system is established based on conventional water analysis indicators collected from shallow layers and microseismic monitoring indicators collected from deep layers.

3. The method according to claim 2, characterized in that, The specific composition of the indicator system is as follows: the shallow conventional water analysis indicator layer includes: chloride, sulfide, petroleum hydrocarbons, strontium, pH, bromide, nitrate, sulfate, ammonia nitrogen, hexavalent chromium, total hardness, iron, manganese, and permanganate index; the microseismic monitoring indicator layer includes microseismic event frequency, average magnitude, b-value, event spatial clustering, and minimum distance from the wellbore; the hydrogeological indicator layer includes aquifer thickness, permeability coefficient, and groundwater flow direction.

4. The method according to claim 1, characterized in that, The normal threshold and weight coefficients of the indicator system are specifically as follows: based on historical data, the thresholds and weight coefficients are dynamically adjusted using a time series analysis algorithm. The time series analysis algorithm uses the autoregressive integral moving average (ARIMA) model, and the formula is: in, For the shift operator, and It is a polynomial. It is the difference order. For time series data, It is white noise.

5. The method according to claim 1, characterized in that, The microseismic event-risk correlation model is constructed by combining the Analytic Hierarchy Process (AHP) with the LightGBM gradient boosting tree algorithm, specifically as follows: Based on microseismic monitoring data, characteristic parameters are extracted, including event frequency, magnitude, b-value, spatial clustering, and distance from the wellbore. The weights of each feature parameter are determined using the Analytic Hierarchy Process (AHP). The AHP weights are calculated by constructing a judgment matrix and solving for the eigenvectors. The formula is as follows: in, To determine the elements of a matrix. To represent the importance of the i-th parameter relative to the j-th parameter, a scaling method is used. The weight vector is obtained by solving the eigenvalue problem. It is the largest eigenvalue; Using feature parameters and AHP weights as input, a LightGBM gradient boosting tree algorithm model is trained to predict deep formation stability index and wellbore integrity index. The objective function is: in Let be the objective function. For loss function, This is a regularization term used to prevent overfitting; it calculates the prediction error of the trained model and triggers model updates based on the prediction error.

6. The method according to claim 5, characterized in that, The step of triggering model updates based on prediction errors specifically involves: using the incremental dataset D to perform performance validation on the existing improved AHP-LightGBM model that correlates microseismic events with risks, and calculating the model prediction error rate. The formula is: in, The number of samples in the incremental dataset. These are actual monitoring values, including measured values ​​of deep formation stability index and measured values ​​of shallow contamination risk. These are the model's predicted values; When the prediction error rate When the value is too large, a model update is triggered, the feature parameter weights are readjusted using AHP, and an incremental dataset is used. Dataset merged with historical datasets Retrain the LightGBM model.

7. The method according to claim 1, characterized in that, The calculation of shallow contamination risk value using the fuzzy comprehensive evaluation method specifically includes: Establish factor set ,in Indicates various water quality indicators; Create a collection of comments ,in The risk level is indicated by categories such as low risk, medium risk, and high risk. Construct a membership matrix M, where each element Indicators This is a comment. The degree of membership is calculated using a trapezoidal function. Determine the weight vector W using the entropy weighting method or AHP; Fuzzy synthesis is performed to calculate the shallow contamination risk value P. The specific formula is as follows: in, This represents the risk value for shallow contamination. This represents a fuzzy synthesis operator, using a weighted average model.

8. The method according to claim 1, characterized in that, Based on comprehensive risk value The warning levels are divided into four levels, and the comprehensive risk value is calculated using the following formula: in, , This represents the risk value for shallow contamination. This is a stability index for deep strata. The wellbore integrity index. , , These are the weighting coefficients; The warning levels are divided into four levels based on the comprehensive risk value Rtotal, and three risk level thresholds are preset. And formulate corresponding countermeasures.

9. The method according to claim 8, characterized in that, The formulation of corresponding response measures specifically involves: based on risk level thresholds. , , The four warning levels are divided into four levels, and different measures are taken for each level: The comprehensive risk value for Level 1 early warning is 0- The shallow indicators are normal, the deep microseismic events are normal, and the wellbore is intact. Measures to be taken: maintain the current monitoring frequency and submit an assessment report every quarter. Level II Early Warning Comprehensive Risk Value - If there are slight anomalies in some non-core indicators in the shallow layer, or a slight increase in the frequency of microseismic events in the deep layer, the following measures should be taken: increase the monitoring frequency of the shallow layer, strengthen the analysis of microseismic data, submit reports daily, and check the parameters of the reinjection operation. Level 3 Early Warning Comprehensive Risk Value - If the shallow core indicators are close to the threshold, or the frequency and magnitude of deep microseismic events increase sharply, with 3-5 microseismic events per week around the wellbore, the following measures should be taken: suspend some reinjection operations, initiate emergency monitoring, and organize experts to formulate prevention and control plans. Level IV Early Warning Comprehensive Risk Value -10 points indicates that shallow indicators exceed the standard, or that deep microseismic events occur frequently and with high magnitudes, with more than 5 microseismic events per week around the wellbore. Measures to be taken include: immediately stopping reinjection, initiating emergency response, carrying out pollution control and wellbore repair, and reporting to the environmental protection department.

Images