Underground water environment monitoring system

Abstract

The invention belongs to the technical field of environment monitoring, and discloses an underground water environment monitoring system, which comprises a monitoring well body; the packers can be remotely controlled and are arranged in the depth direction of the monitoring well body; the optical fiber array can invert the seepage velocity field of the stratum around the well by applying controlled thermal pulse and monitoring temperature response; the robot flying shuttle can move in the well, and an in-situ chemical analysis module or a geophysical imaging module can be carried on the robot flying shuttle; the artificial intelligence model runs on the remote cloud platform; the artificial intelligence model can fuse multi-source data, takes a seepage field obtained by active detection and a stratum structure obtained by imaging as physical constraints, and can autonomously dispatch a robot shuttle to execute minute-level in-situ chemical diagnosis or scan inter-well strata. According to the invention, the leap from passive observation to active intelligent exploration is realized, and the monitoring precision and timeliness are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring technology, and specifically relates to a groundwater environmental monitoring system. Background Technology

[0002] Groundwater is a vital strategic resource, and its environmental monitoring is crucial for water resource protection and pollution control. Current groundwater monitoring technologies primarily rely on deploying point sensors (such as pressure, temperature, and pH meters) in monitoring wells and utilizing distributed fiber optic sensing technology to transmit data to a cloud platform for analysis via wireless networks. Some systems are also attempting to incorporate artificial intelligence (AI) models for data processing and trend prediction.

[0003] However, in the process of realizing this invention, the inventors discovered a fundamental limitation in existing technical systems: they are essentially passive observation systems. They can only wait for environmental changes to occur and be passively captured by sensors, leading to the following deep-seated technical challenges:

[0004] 1. Whether using point sensors or passive optical fibers, the measured physical quantities (such as temperature and pressure) have a nonlinear and non-unique indirect relationship with core hydrogeological parameters (such as water content, permeability coefficient, seepage velocity, and stratigraphic structure). This results in artificial intelligence models often being black-box models, with prediction results lacking physical interpretability and limited accuracy and reliability.

[0005] 2. Once an anomaly is detected, it usually relies on manual on-site inspection or sampling analysis, which has a long response cycle and cannot capture instantaneous or rapidly changing pollution events. Even the mobile detector in the US patent application with publication number US20170044894A1 is limited to passive point sensing along a profile and cannot provide diagnostic information.

[0006] 3. The formation parameters relied upon by artificial intelligence models are typically derived from one-time sampling during drilling, which is static and low-resolution. When groundwater flow alters the formation structure (e.g., erosion creates dominant channels), the model cannot update itself, leading to distorted predictions. Summary of the Invention

[0007] The present invention aims to at least partially solve the aforementioned technical problems. Therefore, the present invention aims to provide a groundwater environment monitoring system that can actively acquire key physical parameters, dynamically update the geological model, and achieve rapid response and diagnosis of abnormal events.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A groundwater environment monitoring system includes: a monitoring well; multiple remotely controllable packers arranged along the depth direction of the monitoring well to divide the monitoring well into multiple independent monitoring segments, each monitoring segment containing a point sensor for monitoring conventional water quality or hydrological parameters; an optical fiber array arranged along the depth direction of the monitoring well, the optical fiber array including a sensing optical fiber and a heating element arranged alongside the sensing optical fiber, the sensing optical fiber being used to acquire distributed temperature and / or strain data, the heating element being used to apply controlled heat pulses to the strata surrounding the monitoring well; a robotic shuttle configured to move along the depth direction of the monitoring well; an edge computing module communicatively connected to the packers, point sensors, optical fiber array, and robotic shuttle; and a remote cloud platform on which an artificial intelligence model runs, the artificial intelligence model being used to fuse and analyze the monitoring data from the point sensors and optical fiber array, and to issue control commands to the robotic shuttle to the edge computing module based on the analysis results.

[0010] Furthermore, the artificial intelligence model is also configured to: calculate the thermophysical parameters and seepage velocity field of the formation around the monitoring well body along the depth based on the temperature change curve monitored by the sensing fiber during and after the application of the heat pulse to the heating element.

[0011] Furthermore, the artificial intelligence model is a physical information neural network model, which uses the seepage velocity field derived from the temperature change curve as a physical constraint to improve the accuracy of predicting groundwater flow and solute transport.

[0012] Furthermore, the robotic shuttle is equipped with a replaceable task module.

[0013] Furthermore, the task module is an in-situ chemical analysis module, which includes a microfluidic chip and a sampling interface. The sampling interface is used to extract a trace amount of water sample from the monitoring section to the microfluidic chip to achieve in-situ real-time detection of chemical components in the water sample.

[0014] Furthermore, the mission module is a geophysical imaging module, which includes a signal transmitting unit and / or a signal receiving unit.

[0015] Furthermore, the system includes at least two monitoring wells, one of which contains a robotic shuttle carrying the signal transmitting unit, and the other contains a robotic shuttle or a fixed sensor as the signal receiving unit. The two are configured to work together to perform cross-hole fault scanning, thereby acquiring structural images of the formation between wells.

[0016] Furthermore, the packer is provided with a sampling valve that cooperates with the sampling interface of the robot shuttle, and the sampling valve is used to allow the robot shuttle to extract water samples after docking.

[0017] Furthermore, the artificial intelligence model is also configured to: autonomously generate task instructions when it identifies abnormal monitoring data or predicts potential risks, and schedule the robot shuttle to move to a specified depth through the edge computing module to perform in-situ chemical analysis or geophysical imaging tasks.

[0018] Furthermore, the system also includes a composite energy harvesting module and an intelligent power management unit; the artificial intelligence model is also configured to predict future task energy consumption and optimize the scheduling of energy harvested by the composite energy harvesting module through the intelligent power management unit.

[0019] The beneficial effects of this invention are as follows:

[0020] This invention can actively apply thermal disturbance to the formation through the set fiber optic array, and quantitatively invert the seepage velocity field around the well based on the thermal response. It provides key physical inputs for artificial intelligence models that cannot be obtained by passive sensing technology, and transforms prediction from correlation guessing to physical calculation.

[0021] This invention designs the robotic shuttle as a mobile geophysical platform, and through a cross-well collaborative working mode, it can perform high-resolution scanning of the wellbore or the formation between wells as needed, dynamically capturing the evolution of the formation structure, thus solving the fundamental defect of static geological parameters in traditional models.

[0022] This invention integrates a microfluidic analysis chip into a robotic shuttle, enabling in-situ chemical diagnosis of pollutants within minutes of an AI-generated warning, providing unprecedented timeliness for capturing and tracking sudden pollution events. Through an artificial intelligence model, the aforementioned active detection, dynamic imaging, and real-time diagnostic capabilities are organically integrated to construct a complete intelligent closed loop. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall architecture of a groundwater environment monitoring system according to the present invention.

[0024] Figure 2 This is a schematic diagram of the inter-well collaborative operation of the present invention.

[0025] Figure 3 This is a functional block diagram of the robot shuttle and its task module of the present invention. Detailed Implementation

[0026] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0027] It should be understood that, and also noted, in the embodiments, the functions / actions may appear in a different order than those shown in the figures. For example, depending on the functions / actions involved, they may actually be performed substantially concurrently, or sometimes the two figures shown consecutively may be performed in reverse order.

[0028] like Figure 1 , Figure 2 and Figure 3 As shown, the groundwater environment monitoring system of this embodiment consists of three main parts: underground equipment, a ground control unit, and a remote cloud platform. The underground equipment is installed inside the monitoring well and includes multiple controllable packers and attached point sensors, a fiber optic array, and a robotic shuttle. The ground control unit is responsible for on-site data processing and equipment control, and includes an edge computing module, a composite energy and power management module, and a data transmission module. The ground control unit is connected to the underground equipment via cables and communicates bidirectionally with the remote cloud platform via a data transmission module (such as 4G / 5G / satellite).

[0029] The packer can be an inflatable rubber bladder. The packer has two states: a default state and a primary state, which is the working state. In this state, the packer is inflated, its outer wall tightly adhering to the well wall. It divides the well into multiple independent monitoring sections, preventing crosstalk between them and ensuring that the point sensors in each section measure the true hydrological and water quality information for that section. In this state, the robotic shuttle cannot pass through.

[0030] The packer also has a passage mode. When the system determines that a robotic shuttle needs to be dispatched to perform a task (e.g., moving from section A to section C), the system executes a passage sequence. Upon receiving the command, the packer in section B, located between sections A and C, performs a deflation / contraction action. The packer retracts towards its central axis, creating a sufficient passage for the robotic shuttle to pass through in the center of the wellbore. After the robotic shuttle passes through, the packer can be re-inflated to restore its isolation function.

[0031] like Figure 2As shown, the fiber optic array is deployed along the full depth of the monitoring well. The array consists of tightly coupled sensing fibers and heating elements. The sensing fibers are standard single-mode fibers connected to a ground demodulator to achieve distributed temperature sensing (DTS) and / or distributed acoustic / strain sensing (DAS / DSS) functions, used for passively monitoring the natural temperature field, stress-strain, or sound field around the well. The heating elements are heating wires or conductive coatings that run alongside or are integrated with the sensing fibers. They are connected to a precision power supply in the ground control unit, which can apply precisely timed and powerful electrothermal pulses to the fiber optic array according to instructions from an artificial intelligence model.

[0032] The fiber optic array's operating modes are controlled by an artificial intelligence model, including passive and active modes. In passive mode, the heating element is inactive, and the sensing fiber continuously monitors the natural state of the wellbore, providing background data. Active mode is triggered when the AI ​​model needs to acquire quantitative hydrogeological parameters. For example, the AI ​​command system applies a brief thermal pulse to the heating element. During and after this pulse, the sensing fiber records the temperature response curve of the entire well section at high temporal resolution. The AI ​​model on the remote cloud platform receives this curve data and, using a built-in heat conduction inversion algorithm, calculates the effective thermal conductivity and heat capacity of the formation surrounding the wellbore at each depth. Since the relevant thermal parameters of water are much higher than those of the rock and soil skeleton, these parameters can be directly and quantitatively converted into formation saturation and vertical / radial seepage velocities.

[0033] like Figure 3 As shown, the robotic shuttle is an intelligent platform capable of autonomous movement within the wellbore. It serves as the actuator for dynamic imaging and in-situ diagnostics, with its core being the replaceable mission modules. These modules can be either in-situ chemical analysis modules or geophysical imaging modules.

[0034] When the task module is the in-situ chemical analysis module, and the AI ​​model warns of potential contamination in a certain layer, a robotic shuttle equipped with this module is dispatched. It moves to the packer at the designated depth and automatically docks with the sampling valve pre-installed on the packer via its sampling interface. After successful docking, a trace amount of water sample is extracted and fed into the microfluidic chip inside the module. The chip contains pre-loaded reagents that react with the water sample. A miniature detector within the module reads the reaction signal, quantitatively determining the concentration of the target pollutant within minutes and uploading the results to the cloud platform in real time.

[0035] When the mission module is a geophysical imaging module, it can achieve high-precision imaging inside the well and collaborative fault scanning between wells. Specifically, the principle of achieving high-precision imaging inside the well is as follows: a robotic shuttle equipped with an acoustic or resistivity imaging probe can move at an extremely slow speed inside the well to perform high-resolution scanning of the well wall, clearly identifying minute fractures and pores.

[0036] The principle behind robotic shuttle-based collaborative fault scanning between wells is as follows: In adjacent monitoring wells A and B, an artificial intelligence model independently schedules two robotic shuttles. The robot in monitoring well A carries a signal transmitting unit (such as a high-frequency acoustic source), while the robot in monitoring well B carries a signal receiving unit (such as a high-sensitivity hydrophone array). Under the unified scheduling of the AI, both move synchronously and step-by-step within the well. At each depth step, the robot in monitoring well A transmits a signal, and the robot in monitoring well B receives the signal transmitted through the formation. After completing the scanning of the entire profile, the cloud platform AI processes all received signals using a fault scanning reconstruction algorithm to generate a high-resolution image showing the formation structure, fracture channels, or water-bearing anomalies between the two wells.

[0037] like Figure 1 As shown, the artificial intelligence model is a Physical Information Neural Network (PINN) or a similar fusion model deployed on a remote cloud platform. The artificial intelligence model can achieve multimodal data fusion, physical law constraints, and intelligent closed-loop decision-making.

[0038] Specifically, the principle of achieving multimodal data fusion is as follows: it integrates discrete water quality data from point sensors, continuous profile data from fiber optic arrays, seepage velocity field data from active detection modes, and chemical data and stratigraphic images collected on demand by robotic shuttles.

[0039] The principle behind achieving physical law constraints is as follows: Unlike traditional black-box AI, it incorporates groundwater movement equations and solute transport equations as penalty terms into the loss function of the neural network. The actively detected and inverted seepage velocity field, along with the formation structure obtained from inter-well CT imaging, are used as strong constraints and key input parameters for these physical equations.

[0040] The principle behind intelligent closed-loop decision-making is as follows: When an AI model predicts a potential pollution risk in a certain area, or when its prediction deviates significantly from actual monitoring data, it can autonomously generate exploration tasks. For example, it might issue commands such as, "Dispatch a robot shuttle from a certain well, equipped with a chemical analysis module, to the designated depth to detect nitrate concentration," or "Dispatch robots from two wells, equipped with imaging modules, to perform cross-hole acoustic scanning within a designated depth range." After completing the task, the robot feeds back new, high-value data to the AI, which then updates and corrects its own model, forming a complete intelligent closed loop of "prediction-verification-active exploration-model iteration."

[0041] The workflow of the groundwater environment monitoring system is as follows:

[0042] 1. Initial Phase: The system conducts all-weather passive monitoring. An artificial intelligence model integrates data from point sensors and passive fiber optic cables to establish a basic hydrogeological model.

[0043] 2. AI detects anomalies: The model detects a drop in the water level of an aquifer, but a small, persistent negative anomaly appears in the temperature profile of the overlying weakly permeable layer, which causes a large prediction error in the model.

[0044] 3. AI Autonomous Decision-Making and Active Detection: The AI ​​determined that this was a potential crossflow channel. It autonomously issued the following instructions: (1) Initiate active mode for the fiber optic array in this depth range. The inversion results confirmed the existence of a high-permeability channel. (2) Schedule the robotic shuttles of the two adjacent wells, equipped with geophysical imaging modules, to perform cross-hole scanning of the area.

[0045] 4. Robot Execution and Data Feedback: The robot shuttle completes the scan, and the scanned image clearly shows a micro-crack channel. The robot then uploads the image data.

[0046] 5. Model Correction and Risk Assessment: The AI ​​integrated seepage velocity and fracture images to correct the geological structure model and hydrological parameters. It re-run the prediction and determined that the overflow would carry upper-layer pollution into the lower drinking water aquifer within a certain period of time, issuing a high-level warning to management personnel and providing precise location of the pollution channel and sealing recommendations.

[0047] Through the above implementation methods, the present invention transforms groundwater monitoring from a static, passive data acquisition system into a dynamic, proactive, self-learning, and evolving intelligent exploration and diagnostic system.

[0048] This invention is not limited to the above-described optional embodiments. Anyone can derive other various forms of products under the guidance of this invention. However, regardless of any changes made in their shape or structure, any technical solution that falls within the scope of the claims of this invention shall be protected by this invention.

Claims

1. A groundwater environment monitoring system, characterized in that, include: Monitoring well body; Multiple remotely controllable packers are arranged along the depth direction of the monitoring well body to divide the monitoring well body into multiple independent monitoring sections. Each monitoring section is equipped with a point sensor for monitoring conventional water quality or hydrological parameters. An optical fiber array is deployed along the depth direction of the monitoring well. The optical fiber array includes a sensing fiber and a heating element disposed along with the sensing fiber. The sensing fiber is used to acquire distributed temperature and / or strain data, and the heating element is used to apply controlled thermal pulses to the formation surrounding the monitoring well. A robotic shuttle is configured to move in the depth direction of the monitoring well body; The edge computing module is communicatively connected to the packer, point sensors, fiber optic array, and robot shuttle; A remote cloud platform runs an artificial intelligence model, which is used to fuse and analyze the monitoring data of the point sensors and fiber optic array, and to issue control commands to the robot shuttle to the edge computing module based on the analysis results.

2. The system according to claim 1, characterized in that, The artificial intelligence model is also configured to: calculate the thermophysical parameters and seepage velocity field of the formation around the monitoring well body along the depth based on the temperature change curve monitored by the sensing fiber during and after the application of the heat pulse to the heating element.

3. The system according to claim 2, characterized in that, The artificial intelligence model is a physical information neural network model, which uses the seepage velocity field derived from the temperature change curve as a physical constraint to improve the accuracy of predicting groundwater flow and solute transport.

4. The system according to claim 1, characterized in that, The robotic shuttle is equipped with a replaceable task module.

5. The system according to claim 4, characterized in that, The task module is an in-situ chemical analysis module, which includes a microfluidic chip and a sampling interface. The sampling interface is used to extract a trace amount of water sample from the monitoring section to the microfluidic chip to achieve in-situ real-time detection of chemical components in the water sample.

6. The system according to claim 4, characterized in that, The mission module is a geophysical imaging module, which includes a signal transmitting unit and / or a signal receiving unit.

7. The system according to claim 6, characterized in that, The system includes at least two monitoring wells, one of which contains a robotic shuttle carrying the signal transmitting unit, and the other contains a robotic shuttle or a fixed sensor as the signal receiving unit. The two are configured to work together to perform cross-hole fault scanning to obtain structural images of the formation between wells.

8. The system according to claim 1, characterized in that, The packer is equipped with a sampling valve that cooperates with the sampling interface of the robot shuttle. The sampling valve is used to allow the robot shuttle to extract water samples after docking.

9. The system according to claim 1, characterized in that, The artificial intelligence model is also configured to: autonomously generate task instructions when it identifies abnormal monitoring data or predicts potential risks, and schedule the robot shuttle to move to a specified depth through the edge computing module to perform in-situ chemical analysis or geophysical imaging tasks.

10. The system according to claim 1, characterized in that, It also includes a composite energy harvesting module and an intelligent power management unit; the artificial intelligence model is further configured to predict future task energy consumption and optimize the scheduling of the energy harvested by the composite energy harvesting module through the intelligent power management unit.

Images