Downhole self-adaptive intelligent sampling system and method based on dynamic response strategy

Abstract

The invention discloses an underground self-adaptive intelligent sampling system and method based on a dynamic response strategy, and the system comprises a multi-source sensing module which is used for obtaining underground environment parameters and underground water quality information in real time; the edge calculation and intelligent decision module is used for fusing the data of the multi-source sensing module; the self-adaptive execution module is used for receiving an instruction of the intelligent decision-making module; the sample pretreatment and temporary storage module is used for carrying out online pretreatment and independent packaging on the obtained underground water sample and carrying out associated coding with sampling process data; and the ground cooperation and communication module is used for realizing data interaction, instruction receiving and state reporting between the underground system and the ground monitoring center. According to the method, the transformation from'program fixing 'to'strategy dynamic' and from'remote control 'to'self-adaption' is realized, and the intelligent level and the operation reliability of obtaining high-quality and high-representativeness underground samples under complex geological conditions are remarkably improved.

Description

Technical Field

[0001] This invention relates to the fields of groundwater monitoring, environmental surveys and water resource management. Specifically, it relates to an intelligent groundwater sampling system and method that can automatically determine the optimal sampling time, dynamically adjust the sampling strategy and execute it accurately based on real-time hydrogeological conditions and water quality information in the well. Background Technology

[0002] Groundwater resources refer to water resources existing underground that can be utilized by humans. They are a part of global water resources and are closely linked to and mutually transform with atmospheric water resources and surface water resources. They possess both underground storage space and participate in the natural water cycle, exhibiting characteristics of fluidity and recoverability. The formation of groundwater resources mainly originates from precipitation infiltration and surface water infiltration in modern and earlier geological eras. The abundance of resources is related to climate, geological conditions, and other factors. Before utilizing groundwater resources, both water quality and quantity assessments must be conducted.

[0003] Existing groundwater sampling methods typically rely on manual operation or simple timed automatic control, which have significant technical limitations and efficiency bottlenecks.

[0004] (1) Sampling timing relies on manual labor, which is inflexible and prone to missing the optimal window: Traditional sampling methods, especially when using existing water wells (equipped with irrigation or water supply pumps that operate at fixed times), require technicians to go to the site in person and operate during specific time periods when the pumps are running. This not only consumes a lot of manpower and resources, but also, if technicians are unable to arrive at the designated time for any reason, the entire sampling window will be missed, resulting in interruption of the monitoring plan and missing data sequences, making it difficult to meet the needs of high-frequency and continuous modern groundwater monitoring.

[0005] (2) Sample representativeness is difficult to guarantee, and intelligent judgment is lacking: When pumping begins, the "old water" or disturbed water in the wellbore usually does not represent the true condition of the target aquifer, and often contains a lot of suspended impurities or chemical components that do not belong to the target layer. To ensure sample representativeness, traditional methods require technicians to test multiple parameters such as water level, turbidity, temperature, conductivity, and dissolved oxygen on-site, and judge whether the water quality is "stable" or "fresh" based on experience before deciding whether to take a sample. This process is time-consuming, highly subjective, and cannot achieve automated high-quality sampling without human intervention.

[0006] Therefore, how to develop a downhole sampling system that can automatically sense and make intelligent decisions, enabling it to autonomously determine the best sampling time and dynamically adjust the sampling operation to ensure the high quality and representativeness of the samples, thereby completely eliminating the reliance on manual operation and on-site experience judgment, has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] The purpose of this invention is to provide a downhole adaptive intelligent groundwater sampling system and method based on a dynamic response strategy, so as to solve the technical problems of excessive reliance on manual labor, inaccurate timing, and difficulty in automatically ensuring sample representativeness in existing groundwater sampling technologies.

[0008] A downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy includes:

[0009] The multi-source sensing module is used to acquire environmental parameters and groundwater quality information in the well in real time;

[0010] The edge computing and intelligent decision-making module is used to integrate data from the multi-source sensing module and automatically generate or adjust adaptive sampling strategies based on predefined water quality stability criteria through the built-in dynamic strategy engine.

[0011] An adaptive execution module is used to receive instructions from the intelligent decision-making module and drive the sampling pump, valve and filter assembly to complete precise pumping and sampling actions.

[0012] The sample pretreatment and temporary storage module is used to pretreatment the obtained groundwater samples online, package them independently, and associate and encode them with the sampling process data.

[0013] It also includes a surface coordination and communication module, used to enable data interaction, command reception, and status reporting between the downhole system and the surface monitoring center.

[0014] The multi-source sensing module includes an environmental sensor group and a water quality sensor group;

[0015] The environmental sensor group includes at least a pump status sensor, a water level sensor, a well pressure sensor, and a temperature sensor for monitoring the pump's start-up and shutdown status and operating time.

[0016] The water quality sensor group includes at least a turbidity sensor, a conductivity sensor, a dissolved oxygen sensor, and a pH sensor, and is used to monitor changes in groundwater quality indicators in real time.

[0017] The dynamic strategy engine includes a pre-trained water quality assessment model and an expert rule base. The water quality assessment model is trained based on historical water quality parameter time series data and is used to predict water quality stability trends. The expert rule base defines a "water quality stability criterion", which is a logical combination of multiple water quality parameters whose fluctuation amplitude is less than a set threshold within a set time. The dynamic strategy engine evaluates the sampling timing and sample representativeness based on real-time data and optimizes sampling parameters under environmental constraints.

[0018] The evaluation criteria for the water quality stability criterion are as follows: for each key water quality parameter... Calculate its relative rate of change:

[0019] =

[0020] Where Δt is the sampling interval time, when all parameters are... All less than the threshold If the duration exceeds T0, the stability condition is determined to be met, that is, the current water quality has reached a stable state, and the sampling procedure is automatically triggered.

[0021] The adaptive execution module integrates at least two different hydraulic sampling mechanisms, including but not limited to constant-speed suction, variable-speed suction, and pulse-washing suction, and can automatically switch or combine them according to the instructions of the dynamic strategy engine to adapt to different water quality conditions and sampling targets. The variable-speed suction uses an adaptive PID algorithm for flow rate control, adjusting the pumping rate in real time based on water quality parameter deviations. The basic control formula is:

[0022]

[0023] in, The pumping rate at the current moment;

[0024] This refers to the deviation of water quality parameters (the deviation between the target value and the actual value).

[0025] , , These are the proportional, integral, and differential coefficients, respectively.

[0026] Unlike traditional PID systems with fixed parameters, the PID parameters in this system are dynamically adjusted according to water quality conditions. For example, when turbidity is high, the system will automatically reduce the PID parameters. Reduce the control intensity to prevent the turbidity from increasing further due to over-adjustment.

[0027] The sample pretreatment and temporary storage module includes:

[0028] An online filtration unit is used to remove large particulate impurities before the sample enters the storage bottle;

[0029] A multi-channel sample storage unit is used to store multiple independent samples at different times or depths during a single well operation;

[0030] An automated sealing and labeling unit is used to seal and print or mark unique codes on stored sample vials;

[0031] The data association unit is used to bind environmental data, water quality data, and sampling strategy parameters at the sampling time with the unique code to form a holographic data packet.

[0032] An adaptive intelligent groundwater sampling method for wells includes the following steps:

[0033] S1: System deployment and initialization, lowering the sampling system to the target aquifer depth and establishing a communication link with the ground;

[0034] S2: Real-time sensing and data fusion, continuously acquiring real-time data on pump status, water level, pressure, temperature and multiple water quality parameters, and performing data cleaning and fusion;

[0035] S3: Intelligent assessment and strategy triggering. Based on the fused real-time data, it uses the water quality assessment model and expert rule base in the dynamic strategy engine for continuous analysis. When the predefined "water quality stability criteria" are met and / or a specific water quality change event is identified, the sampling strategy generation process is automatically triggered.

[0036] S4: Dynamically generate and optimize sampling strategies. Based on the current water quality conditions, environmental constraints and sampling objectives, dynamically generate or select the optimal sampling scheme from the strategy library. The sampling strategy includes at least the sampling start time, pumping rate, sampling duration, whether to enable online filtration and the selection of sample storage channels.

[0037] S5: Adaptive and precise execution, based on the generated sampling strategy, drives the water pump and controls the valve to complete the groundwater extraction in a specified manner and delivers the water sample to the designated pretreatment and storage unit;

[0038] S6: Sample post-processing and information encapsulation. The obtained water sample is pre-processed, encapsulated independently, and a unique code is established to associate the data of the entire sampling process with the sample.

[0039] S7: Data upload and system reset. Uploads key process data and holographic data packets to the ground, and prepares for the next sampling or performs self-cleaning operation according to instructions or preset programs.

[0040] The method further includes an adaptive closed-loop control sub-step in step S5: during the sampling process, key water quality parameters (such as turbidity) are continuously monitored and compared with the strategy expectation; if the actual parameters deviate from the expectation, the dynamic strategy engine adjusts the execution parameters (such as pumping rate) in real time, or switches to the backup sampling scheme (such as starting stronger filtration or extending the pumping time) according to preset rules to ensure that representative samples are obtained.

[0041] In step S7, the complete data package for each sampling (including triggering conditions, execution strategy, process data, and final sample analysis results) is uploaded to a ground-based big data platform or the cloud for continuous training and optimization of the water quality assessment model in the dynamic strategy engine, thereby enabling case learning and model iteration.

[0042] The beneficial effects of the present invention are as follows:

[0043] (1) Fully automatic and unattended operation: By intelligently judging the sampling time in real time downhole, the dependence on manual on-site operation is completely eliminated, realizing true 7x24-hour automated sampling, which is especially suitable for remote areas or occasions requiring high-frequency monitoring.

[0044] (2) Ensure high representativeness of samples: Based on the "water quality stability criterion" of multi-parameter fusion, the water samples that are not representative at the beginning of the pump start-up are scientifically and objectively excluded, ensuring that each sample can truly reflect the water quality of the target aquifer, which greatly improves the quality and reliability of the monitoring data.

[0045] (3) Highly adaptive and reliable: The system can dynamically adjust the sampling action (such as flow rate and filtration) according to real-time water quality changes, and has self-cleaning and fault response strategies, which significantly improves the long-term working reliability and adaptability in complex and harsh groundwater environments.

[0046] (4) Data deep association and traceability: The innovative “sample-data” holographic binding mode makes each water sample carry a complete “birth certificate”, which greatly facilitates subsequent laboratory analysis, data interpretation and quality control, and realizes the full life cycle management of sample information.

[0047] (5) Intelligent evolution potential: Through data accumulation and model iteration on cloud or ground platforms, the decision-making ability of the entire system can be continuously optimized, gradually adapting to the sampling needs under different geological and hydrological conditions, and possessing the ability to evolve over a long period of time. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating the core steps of steps S0-S2 in this invention.

[0049] Figure 2 This is a flowchart illustrating the core process steps S3-S4 of the present invention.

[0050] Figure 3 This is a flowchart illustrating the core process steps S5-S6 of the present invention.

[0051] Figure 4 This is a flowchart illustrating the core steps of steps S7-S8 of the present invention. Detailed Implementation

[0052] To more clearly illustrate the technical features of this solution, the following detailed implementation method will be used to explain the solution.

[0053] See Figures 1-4 A downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy includes:

[0054] The multi-source sensing module is used to acquire environmental parameters and groundwater quality information in the well in real time;

[0055] The edge computing and intelligent decision-making module is used to integrate data from the multi-source sensing module and automatically generate or adjust adaptive sampling strategies based on predefined water quality stability criteria through the built-in dynamic strategy engine.

[0056] The adaptive execution module receives instructions from the intelligent decision-making module and drives the sampling pump, valves, and filter components to complete precise pumping and sampling actions.

[0057] The sample pretreatment and temporary storage module is used to pretreatment the obtained groundwater samples online, package them independently, and associate and encode them with the sampling process data.

[0058] It also includes a surface coordination and communication module, used to enable data interaction, command reception, and status reporting between the downhole system and the surface monitoring center.

[0059] The multi-source sensing module includes an environmental sensor group and a water quality sensor group;

[0060] The environmental sensor group includes at least a pump status sensor, a water level sensor, a well pressure sensor, and a temperature sensor for monitoring the pump's start-up and shutdown status and operating time.

[0061] The water quality sensor array includes at least a turbidity sensor, a conductivity sensor, a dissolved oxygen sensor, and a pH sensor, and is used to monitor changes in groundwater quality indicators in real time.

[0062] Among them, the environmental sensor group monitors environmental parameters directly related to the sampling operation in real time, including but not limited to: water level gauge (monitoring dynamic water level changes), pressure sensor (monitoring well pressure), temperature sensor, and key pump status sensor (used for non-invasive or direct sensing of the pump's start / stop status and cumulative running time).

[0063] The water quality sensor suite integrates multiple online water quality analysis sensors, including at least: a turbidity sensor (monitoring the degree of turbidity in the water), a conductivity sensor (monitoring changes in total dissolved solids), a dissolved oxygen sensor, and a pH sensor. Optionally, it can also integrate specific ion sensors (such as nitrate and chloride ions) or organic matter fluorescence sensors for monitoring specific contaminated sites.

[0064] The dynamic strategy engine includes a pre-trained water quality assessment model and an expert rule base. The water quality assessment model is trained based on historical water quality parameter time series data and is used to predict water quality stability trends. The expert rule base defines a "water quality stability criterion", which is a logical combination of multiple water quality parameters whose fluctuation amplitude is less than a set threshold within a set time. The dynamic strategy engine evaluates the sampling timing and sample representativeness based on real-time data and optimizes sampling parameters under environmental constraints.

[0065] The evaluation criteria for water quality stability are as follows: for each key water quality parameter... Calculate its relative rate of change:

[0066] =

[0067] Where Δt is the sampling interval time, when all parameters are... All less than the threshold If the duration exceeds T0, the stability condition is determined to be met, that is, the current water quality has reached a stable state, and the sampling procedure is automatically triggered.

[0068] Edge Computing and Intelligent Decision-Making Module: This module, with a corrosion-resistant and water-pressure-resistant embedded industrial computer at its core, is deployed underground. The data fusion center receives asynchronous data streams from all sensors, performs time synchronization, filtering, cleaning, and standardization processing, and constructs a real-time dynamic digital profile of the underground hydrological and water quality status.

[0069] The dynamic strategy engine is the "brain" of this invention. It embeds a pre-trained water quality assessment model (trained based on a large amount of historical water quality time-series data, capable of predicting stable trends in water quality parameters) and a configurable expert rule base. The core of the rule base is a user-definable "water quality stability criterion," a set of complex logical conditions consisting of multiple water quality parameters and their change thresholds and durations. The engine continuously analyzes the fused data, and once it detects the water pump starting, it enters a "monitoring-assessment" loop. When real-time data meets the "water quality stability criterion," it automatically determines the optimal sampling time and immediately triggers strategy generation. Furthermore, the engine can optimize specific sampling parameters in real time based on the objective (such as routine monitoring or pollution event capture) and environmental constraints (such as battery power and sample bottle remaining volume).

[0070] The adaptive execution module integrates at least two different hydraulic sampling mechanisms, including but not limited to constant-speed suction, variable-speed suction, and pulse-washing suction, and can automatically switch or combine them according to instructions from the dynamic strategy engine to adapt to different water quality conditions and sampling targets. Among them, the flow rate control of variable-speed suction adopts an adaptive PID algorithm, adjusting the pumping rate in real time according to the deviation of water quality parameters. The basic control formula is:

[0071]

[0072] in, The pumping rate at the current moment;

[0073] This refers to the deviation of water quality parameters (the deviation between the target value and the actual value).

[0074] , , These are the proportional, integral, and differential coefficients, respectively.

[0075] Unlike traditional PID systems with fixed parameters, the PID parameters in this system are dynamically adjusted according to water quality conditions. For example, when turbidity is high, the system will automatically reduce the PID parameters. Reduce the control intensity to prevent the turbidity from increasing further due to over-adjustment.

[0076] In the adaptive execution module, the drive and execution mechanisms mainly include a high-precision variable-speed submersible pump or a multi-channel solenoid valve controller. It receives instructions from the intelligent decision-making module and can precisely control the pumping rate (e.g., gradually increasing from a low flow rate), sampling duration, and valve switching for distributing water samples to different storage bottles. The mechanism switching capability can switch between different sampling modes according to strategy instructions. For example, in the initial stage, a "low-flow-rate well cleaning" mode is used, switching to a "standard flow-rate sampling" mode after the water quality stabilizes; if encountering high-turbidity water, it can automatically switch to a "pulse suction + enhanced filtration" mode to prevent clogging.

[0077] The sample preprocessing and temporary storage module includes:

[0078] An online filtration unit is used to remove large particulate impurities before the sample enters the storage bottle;

[0079] A multi-channel sample storage unit is used to store multiple independent samples at different times or depths during a single well operation;

[0080] An automated sealing and labeling unit is used to seal and print or mark unique codes on stored sample vials;

[0081] The data association unit is used to bind environmental data, water quality data, and sampling strategy parameters at the sampling time with the unique code to form a holographic data packet.

[0082] Specifically, the online filtration unit automatically uses membranes of varying precision to filter and remove suspended particles before the sample enters the storage bottle. The multi-channel sample storage unit, employing a rotary or multi-row design, can store multiple independent samples in a single well run for collection at different times or under different water quality conditions. After sampling, the automatic sealing and labeling unit automatically seals the sample bottle (e.g., by adding a sealing cap or liner) and generates a unique code corresponding to the sample using a micro-printer or RFID tag writer. The data association unit automatically binds all "process data" (such as water quality parameters at the time of triggering, strategy parameters used, and execution logs) and "environmental data" (such as sampling depth and temperature) of this sampling to the sample's unique code, forming an inseparable "sample-data" holographic package.

[0083] Preferably, the wellhead coordination and communication module achieves two-way communication between the well and the surface through technologies such as cable carrier communication, wireless radio frequency, or fiber optic microcable. It can compress and upload critical status information, alarm messages, and holographic data packets, and can receive mission parameter updates, model upgrade packages, or emergency intervention commands issued from the surface.

[0084] An adaptive intelligent groundwater sampling method for wells includes the following steps:

[0085] More ideally, S0: Task Configuration and Strategy Pre-installation: Ground operators define the objectives of this monitoring task (such as routine monthly monitoring or pollution plume tracking) using monitoring software, and set key parameters, including the target aquifer depth, the type of pollutant of interest, the specific threshold of the "water quality stability criterion" (e.g., turbidity <5 NTU and stable for more than 10 minutes, conductivity change rate <1% / min), and the expected number of samples. After configuration, the information is sent to the downhole system via the communication module.

[0086] S1: System deployment and initialization, the sampling system is lowered to the target aquifer depth, power-on self-test, each sensor and mechanism is initialized, and a stable communication link is established;

[0087] S2: Real-time sensing and data fusion. The system enters a low-power monitoring state and continuously acquires real-time data on pump status, water level, pressure, temperature, and multiple water quality parameters. The data fusion center processes the data in real time to form a fused data stream that can be used for decision-making.

[0088] S3: Intelligent assessment and strategy triggering. Based on the fused real-time data, it uses the water quality assessment model and expert rule base in the dynamic strategy engine for continuous analysis. When the predefined "water quality stability criteria" are met and / or a specific water quality change event is identified, the sampling strategy generation process is automatically triggered.

[0089] S4: Dynamically generate and optimize sampling strategies. Based on the current water quality conditions, environmental constraints and sampling objectives, dynamically generate or select the optimal sampling plan from the strategy library. The sampling strategy includes at least the sampling start time, pumping rate, sampling duration, whether to enable online filtration and the selection of sample storage channels.

[0090] S5: Adaptive and precise execution, based on the generated sampling strategy, drives the water pump and controls the valve to complete the groundwater extraction in a specified manner and delivers the water sample to the designated pretreatment and storage unit;

[0091] S6: Sample post-processing and information encapsulation. The obtained water sample is pre-processed, encapsulated independently, and a unique code is established to associate the data of the entire sampling process with the sample.

[0092] S7: Data upload and system reset. Uploads key process data and holographic data packets to the ground, and prepares for the next sampling or performs self-cleaning operation according to instructions or preset programs.

[0093] The method also includes an adaptive closed-loop control sub-step in step S5: during the sampling process, key water quality parameters (such as turbidity) are continuously monitored and compared with the strategy expectation; if the actual parameters deviate from the expectation, the dynamic strategy engine adjusts the execution parameters (such as pumping rate) in real time, or switches to the backup sampling scheme (such as starting stronger filtration or extending the pumping time) according to preset rules to ensure that representative samples are obtained.

[0094] In step S7, the complete data package for each sampling (including triggering conditions, execution strategy, process data, and final sample analysis results) is uploaded to the ground big data platform or the cloud for continuous training and optimization of the water quality assessment model in the dynamic strategy engine, thereby realizing case learning and model iteration.

[0095] Preferably, during self-cleaning, a small cleaning pump and a storage tank are integrated. According to the instructions of the strategy engine, the sampling pipeline, filter screen and sensor probe can be automatically backwashed or chemically cleaned during sampling intervals or after the task is completed, which greatly extends the continuous working time of the system in complex water quality environments.

[0096] Example 1: Monthly automatic sampling of conventional groundwater monitoring wells

[0097] The system of this invention has been deployed in a groundwater monitoring network in a certain area. At the beginning of each month, the ground center wakes up the downhole system via remote command.

[0098] Mission pre-configuration: Ground-based instructions are issued to set the "water quality stability criterion" as follows: turbidity continuously below 3 NTU and conductivity change rate less than 0.5% / minute for a duration of 15 minutes.

[0099] Deployment and Waiting: The system is deployed to the main aquifer depth and enters a low-power monitoring state.

[0100] Automatic Trigger: On a certain day of the month, the agricultural irrigation pump starts as scheduled. The pump status sensor detects the pump's operation, and the system is immediately activated. The water quality sensor begins high-frequency monitoring. In the initial stage of pumping, the turbidity reaches as high as 20 NTU, and the conductivity changes rapidly, failing to meet the criteria.

[0101] Dynamic Decision-Making and Execution: After approximately 25 minutes, real-time data showed that the turbidity decreased to 2 NTU and remained stable, while the conductivity change rate decreased to 0.3% / min. The dynamic strategy engine determined that the conditions were met and immediately triggered. Based on the current state of sufficient battery power and empty bottle 1, the engine generated the strategy: "Pump water at a flow rate of 500 mL / min for 5 minutes, activate 5μm online filtration, and store the sample in bottle 1."

[0102] Closed-loop sampling: The adaptive execution module controls the water pump to operate at a specified flow rate, and the turbidity remains stable throughout the process. After 5 minutes, approximately 2.5 liters of water sample is accurately obtained, filtered, and stored in bottle No. 1.

[0103] Information Encapsulation and Upload: The automatic encapsulation unit seals bottle No. 1 and prints the label "MW-01-20231015-001". Simultaneously, the entire process curves of water level, temperature, turbidity, and conductivity for 15 minutes before and after the trigger, as well as the strategy parameters used, are all bound to the label code and uploaded to the ground center as a data packet.

[0104] Reset: After automatically performing a single pipeline flush with clean water, the system enters hibernation mode, awaiting the next month's task. The entire process requires no human intervention, and the obtained samples are representative water samples after the water quality has stabilized.

[0105] Example 2: Emergency Monitoring and Sampling of Contaminated Sites

[0106] This system was deployed in a monitoring well at an industrial site suspected of having a leak, and was equipped with a fluorescent sensor for characteristic pollutants (such as benzene compounds).

[0107] Pre-installed task: Two sets of triggering logic are set. One is the conventional "water quality stability criterion". The other is the "event triggering criterion": a sudden increase in the fluorescence index of benzene series compounds exceeding 50% of the background value.

[0108] Monitoring and event capture: During routine monitoring, one morning, the reading of the benzene series compound sensor rose sharply, increasing by more than 80% within 2 minutes, immediately triggering the "event sampling" mode.

[0109] Dynamic strategy generation: The dynamic strategy engine prioritizes responses to event triggers. Considering the possibility of an initial pollution front arriving, the engine generates a rapid capture strategy: "Immediately pump water at the maximum safe flow rate of 800 mL / min for 3 minutes. To capture potential high concentration peaks, filtration will not be used this time. The sample will be stored in a dedicated 'event sample' bottle No. 2 (brown glass bottle, pre-inertized)."

[0110] Execution and sample processing: The system responded quickly, completing water sample collection and storage before the pollution peak could subside.

[0111] Enhanced data correlation: In addition to routine data, this study used real-time abrupt changes in benzene series concentrations as core process data, strongly correlated with the samples. The system also simultaneously sent a "Level 3 pollution event alarm" and snapshot data to ground monitoring personnel via the communication module.

[0112] Follow-up actions: After receiving the alarm, ground personnel can remotely instruct the system to collect a control sample according to standard criteria once the water quality has stabilized, for comparative research. This embodiment demonstrates the system's ability to automatically detect sudden pollution events in unattended conditions, providing crucial first-time samples and data for emergency response.

[0113] Example 3: Adaptive Maintenance and Sampling in Long-Term Observation Wells

[0114] In long-term monitoring wells for mineral water sources, the system needs to address the risks of sensor drift and pipeline blockage caused by potential microbial film growth or minor scaling.

[0115] Integrated self-cleaning module: The system is equipped with the optional self-cleaning and maintenance sub-module of this invention.

[0116] Intelligent maintenance strategy: The dynamic strategy engine has preset maintenance rules: a) After each successful sampling, a clean water backwash is automatically performed; b) If the time required to trigger three consecutive samplings (i.e., the water quality stabilization time) increases significantly, it is determined that the filter or sensor may be clogged, and the engine will decide to automatically perform an enhanced "citric acid cleaning cycle" before the next sampling.

[0117] Adaptive Execution: During the six-month observation period, the system automatically completed dozens of samplings. During the seventh sampling, the engine, based on historical data comparison, found that the water quality stabilization time had increased from an average of 30 minutes to 55 minutes, triggering maintenance rule b.

[0118] Automatic maintenance: Before the eighth water pump start-up, the system automatically pumps citric acid cleaning solution from the maintenance fluid tank to circulate and clean the sampling flow path and sensor chamber, followed by replacement with clean water. After cleaning, the water quality stabilization time returns to 32 minutes for the next cycle.

[0119] Benefits: The system achieves fully automated integrated management of "sampling and maintenance", ensuring long-term stability of data quality and reliable operation of the system itself during months of unattended observation, and avoiding sample distortion or mission failure due to equipment performance degradation.

[0120] The technical features of this invention not described can be implemented by or using existing technology, and will not be repeated here. Of course, the above description is not a limitation of this invention, and this invention is not limited to the examples above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of this invention should also be within the protection scope of this invention.

Claims

1. A downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy, characterized in that, include: The multi-source sensing module is used to acquire environmental parameters and groundwater quality information in the well in real time; The edge computing and intelligent decision-making module is used to integrate data from the multi-source sensing module and automatically generate or adjust adaptive sampling strategies based on predefined water quality stability criteria through the built-in dynamic strategy engine. An adaptive execution module is used to receive instructions from the intelligent decision-making module and drive the sampling pump, valve and filter assembly to complete precise pumping and sampling actions. The sample pretreatment and temporary storage module is used to pretreatment the obtained groundwater samples online, package them independently, and associate and encode them with the sampling process data. It also includes a surface coordination and communication module, used to enable data interaction, command reception, and status reporting between the downhole system and the surface monitoring center.

2. The downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy according to claim 1, characterized in that, The multi-source sensing module includes an environmental sensor group and a water quality sensor group; The environmental sensor group includes at least a pump status sensor, a water level sensor, a well pressure sensor, and a temperature sensor for monitoring the pump's start-up and shutdown status and operating time. The water quality sensor group includes at least a turbidity sensor, a conductivity sensor, a dissolved oxygen sensor, and a pH sensor, and is used to monitor changes in groundwater quality indicators in real time.

3. The downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy according to claim 1, characterized in that, The dynamic strategy engine includes a pre-trained water quality assessment model and an expert rule base. The water quality assessment model is trained based on historical water quality parameter time series data and is used to predict water quality stability trends. The expert rule base defines a "water quality stability criterion", which is a logical combination of multiple water quality parameters whose fluctuation amplitude is less than a set threshold within a set time. The dynamic strategy engine evaluates the sampling timing and sample representativeness based on real-time data and optimizes sampling parameters under environmental constraints. The evaluation criteria for the water quality stability criterion are as follows: for each key water quality parameter... Calculate its relative rate of change: = Where Δt is the sampling interval time, when all parameters are... All less than the threshold If the duration exceeds T0, the stability condition is determined to be met, that is, the current water quality has reached a stable state, and the sampling procedure is automatically triggered.

4. The downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy according to claim 1, characterized in that, The adaptive execution module integrates at least two different hydraulic sampling mechanisms, including but not limited to constant-speed suction, variable-speed suction, and pulse-washing suction, and can automatically switch or combine them according to the instructions of the dynamic strategy engine to adapt to different water quality conditions and sampling targets. The variable-speed suction uses an adaptive PID algorithm for flow rate control, adjusting the pumping rate in real time based on water quality parameter deviations. The basic control formula is: in, The pumping rate at the current moment; This refers to the deviation of water quality parameters (the deviation between the target value and the actual value). , , These are the proportional, integral, and differential coefficients, respectively. Unlike traditional PID systems with fixed parameters, this system dynamically adjusts the PID parameters based on water quality conditions. For example, when turbidity is high, the system will automatically reduce the PID parameters. Reduce the control intensity to prevent the turbidity from increasing further due to over-adjustment.

5. The downhole adaptive intelligent groundwater sampling system based on a dynamic response strategy according to claim 1, characterized in that, The sample pretreatment and temporary storage module includes: An online filtration unit is used to remove large particulate impurities before the sample enters the storage bottle; A multi-channel sample storage unit is used to store multiple independent samples at different times or depths during a single well operation; An automated sealing and labeling unit is used to seal and print or mark unique codes on stored sample vials; The data association unit is used to bind environmental data, water quality data, and sampling strategy parameters at the sampling time with the unique code to form a holographic data packet.

6. A downhole adaptive intelligent groundwater sampling method based on the system described in any one of claims 1-5, characterized in that, Includes the following steps: S1: System deployment and initialization, lowering the sampling system to the target aquifer depth and establishing a communication link with the ground; S2: Real-time sensing and data fusion, continuously acquiring real-time data on pump status, water level, pressure, temperature and multiple water quality parameters, and performing data cleaning and fusion; S3: Intelligent assessment and strategy triggering. Based on the fused real-time data, it uses the water quality assessment model and expert rule base in the dynamic strategy engine for continuous analysis. When the predefined "water quality stability criteria" are met and / or a specific water quality change event is identified, the sampling strategy generation process is automatically triggered. S4: Dynamically generate and optimize sampling strategies. Based on the current water quality conditions, environmental constraints and sampling objectives, dynamically generate or select the optimal sampling scheme from the strategy library. The sampling strategy includes at least the sampling start time, pumping rate, sampling duration, whether to enable online filtration and the selection of sample storage channels. S5: Adaptive and precise execution, based on the generated sampling strategy, drives the water pump and controls the valve to complete the groundwater extraction in a specified manner and delivers the water sample to the designated pretreatment and storage unit; S6: Sample post-processing and information encapsulation. The obtained water sample is pre-processed, encapsulated independently, and a unique code is established to associate the data of the entire sampling process with the sample. S7: Data upload and system reset. Uploads key process data and holographic data packets to the ground, and prepares for the next sampling or performs self-cleaning operation according to instructions or preset programs.

7. The adaptive intelligent groundwater sampling method for wells according to claim 6, characterized in that, The method further includes an adaptive closed-loop control sub-step in step S5: during the sampling process, key water quality parameters are continuously monitored and compared with the expected parameters; if the actual parameters deviate from the expected parameters, the dynamic strategy engine adjusts the execution parameters in real time, or switches to a backup sampling scheme according to preset rules, to ensure that representative samples are obtained.

8. The adaptive intelligent groundwater sampling method according to claim 7, characterized in that, In step S7, the complete data packet of each sampling is uploaded to the ground big data platform or the cloud for continuous training and optimization of the water quality assessment model in the dynamic strategy engine, so as to realize case learning and model iteration.

Images