A simulation experimental system for the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil

Abstract

The present invention provides a simulation experiment system for the migration pathways of polycyclic aromatic hydrocarbons (PAHs) in petroleum-contaminated soil. The system comprises a main reaction chamber, a robotic arm collaborative operation module, a multimodal sensing module, a fully natural environment control module, and an edge-based intelligent decision-making platform. A circular magnetic levitation track is installed on top of the main reaction chamber, and a soil sample chamber is located at its base. The multimodal sensing module, including an EIT electrical impedance imaging module, a LIBS laser spectroscopy module, and a terahertz time-domain spectroscopy module, measures the global conductivity distribution of the soil within the sample chamber, inverts the porosity of the soil sample and the pollutant concentration field in the pore water, analyzes the carbon, hydrogen, and oxygen concentrations of the pollutants, and specifically detects PAH molecules. The edge-based intelligent decision-making platform interacts with the multimodal sensing module, the fully natural environment control module, and the robotic arm collaborative operation module to actively adjust system parameters and perform remediation interventions on soil degradation. This system enables both automated preparation of contaminated soil and migration monitoring under various practical engineering application conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of petroleum-contaminated soil monitoring, and more particularly to a polycyclic aromatic hydrocarbon migration path simulation experimental system for petroleum-contaminated soil. Background Art

[0002] Amidst the booming petroleum industry, the problem of oil-contaminated soil has become increasingly severe. Oil-contaminated soil refers to soil contaminated after an oil spill. The primary pollutant in oil is polycyclic aromatic hydrocarbons, which can cause severe damage to both humans and the soil. Oil is highly volatile and mobile, easily migrating under the influence of temperature and environmental factors. Therefore, monitoring the migration of pollutants throughout their lifecycle and developing quantitative migration models are crucial indicators for controlling oil pollutant migration and remediating contaminated soil, providing scientific support for environmental risk management. This has become a key area of ​​current research and development.

[0003] Existing technologies for monitoring the migration of pollutants in petroleum-contaminated soil, such as Chinese patent ZL202211700526.1, which discloses a "field-phase" coupled migration simulation system for soil pollutants in petrochemical sites, can simulate the soil pollutant migration process that fits the actual conditions of the petrochemical site and couples multiple influencing factors. However, the system cannot automatically complete the simulation process and requires human intervention to prepare soil samples. It relies on offline sampling and analysis of pollutant data and cannot capture the dynamic migration process of pollutants in real time. Moreover, the device can only perform environmental simulation and cannot conduct research on various simulation scenarios such as bioremediation intervention, dynamic response, and solidification of petroleum-contaminated soil. Therefore, it is necessary to propose a fully automatic closed-loop control simulation experimental system that can complete various simulation scenarios. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a simulation experimental system for the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil, which can fully automatically realize efficient, accurate and real-time monitoring of the migration of polycyclic aromatic hydrocarbons in petroleum-contaminated soil.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A simulation experiment system for the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil includes a main reaction chamber, a robotic arm collaborative operation module, a multimodal sensing module, a full natural environment control module, and an edge intelligent decision-making platform. A circular magnetic levitation track is provided on the top of the main reaction chamber, and a soil sample chamber is provided at the bottom of the main reaction chamber.

[0007] The multimodal sensing module includes an EIT electrical impedance imaging module, a LIBS laser spectroscopy module, and a terahertz time-domain spectroscopy module, which are used to measure the global conductivity distribution of the soil in the soil sample room and invert the porosity of the soil sample and the pollutant concentration field in the pore water, analyze the carbon, hydrogen and oxygen element concentrations of the pollutants, and specifically detect polycyclic aromatic hydrocarbon molecules;

[0008] The full natural environment control module is used to simulate different natural environments and is located inside the main reaction chamber;

[0009] The robotic arm collaborative operation module includes a main robotic arm, a micro-manipulator arm and a quick-change end-type tool library. The main robotic arm is installed on a circular magnetic levitation track and adopts a six-degree-of-freedom serial configuration for macro-operation. The end of the main robotic arm is connected to the quick-change end-type tool library through a flange. The quick-change end-type tool library has nine types of tool heads. Different types of tool heads are in a storage state when not in use, including a biological printing head, a microwave thawing head, a multi-channel injection head for injecting additives and soil materials, an acoustic compaction head for increasing the compaction of soil samples during soil sample preparation, a laser sintering head for enhancing the consolidation of soil samples during soil sample preparation, an ultrasonic dispersion head for mixing soil during soil sample preparation, an ultraviolet disinfection head for providing a sterile environment for the main reaction chamber, a fiber grating sensor implantation head for implanting fiber grating sensors during soil sample preparation, and a self-healing agent injection head for injecting nutrients and self-healing agents.

[0010] One end of the micromanipulator is nested inside the third joint of the main robotic arm and is used for microscopic manipulation. A variety of integrated components are fixed to the end of the micromanipulator. These components are stored when not in use. These include a LIBS laser focusing lens for assisting the LIBS laser spectroscopy module in emitting laser light, a six-dimensional force sensor for monitoring soil stress, a photoacoustic monitoring head for monitoring microbial activity, and a capacitive humidity sensor for monitoring the humidity in the main reaction chamber.

[0011] The edge intelligent decision-making platform uses real-time data obtained by the multimodal sensing module, the full natural environment control module and the robotic arm collaborative operation module to actively adjust system parameters and apply engineering measures to repair and intervene in soil deterioration.

[0012] Furthermore, the edge intelligent decision-making platform includes an LSTM-GRU hybrid model and a digital twin module;

[0013] The LSTM-GRU hybrid model is used to predict the migration behavior of pollutants based on the time-varying data obtained by the multimodal sensing module, the full natural environment control module, and the robotic arm collaborative operation module;

[0014] The digital twin module includes a multi-physics field coupling solver, a PPO reinforcement learning algorithm, and a three-dimensional visualization interface;

[0015] A digital twin model is established using a multi-physics coupling solver based on COMSOL software. The migration behavior of pollutants predicted by the LSTM-GRU hybrid model is input into the digital twin model to obtain simulation data. The simulation data and the predicted data are then input into the PPO reinforcement learning algorithm. The PPO reinforcement learning algorithm is used to dynamically optimize the control parameter vector to obtain targeted control instructions. After the PPO reinforcement learning algorithm is optimized, the targeted control instructions are sent to the full natural environment control module and the robotic arm collaborative operation module.

[0016] The three-dimensional visualization interface is used to map and display the migration status of pollutants in real time.

[0017] Furthermore, the input data of the LSTM-GRU hybrid model includes pollutant characteristic data, environmental parameters and soil state data, and the output is the predicted migration behavior of pollutants;

[0018] The pollutant characteristic data include conductivity distribution, C element concentration, H element concentration, O element concentration, polycyclic aromatic hydrocarbon molecule concentration and pollutant phase;

[0019] The environmental parameters include temperature gradient, humidity, O2 concentration, N2 concentration, CO2 concentration, gas flow rate, light intensity and rainfall intensity;

[0020] The soil state data include soil sample porosity, permeability coefficient, soil stress and microbial activity;

[0021] The predicted migration behavior of pollutants includes future pollutant concentration field, migration distance, migration rate, risk level, control parameter vector, diffusion coefficient, permeability and thermal conductivity;

[0022] The control parameter vector includes gas flow rate adjustment amount, heating area weight distribution and rainfall intensity;

[0023] The targeted control instructions include gas flow rate adjustment amount, heating area weight distribution, rainfall intensity, tool head type, tool head action coordinates and action volume.

[0024] Furthermore, the EIT electrical impedance imaging module includes an EIT electrode ring consisting of a 128-electrode ring array, which surrounds the outside of the soil sample chamber;

[0025] The LIBS laser spectroscopy module is mounted at the lower end of the main robotic arm. The LIBS laser spectroscopy module is vertically aligned with the soil sample surface and coaxially arranged with the EIT electrode ring.

[0026] The transmitting end and receiving end of the terahertz time-domain spectroscopy module are symmetrically distributed on the left and right sides of the soil sampling room to form a transmission detection path.

[0027] Furthermore, the full natural environment control module includes an air path control submodule, a temperature and humidity control submodule, and a light / rainfall simulation submodule;

[0028] The gas path control submodule includes a piezoelectric ceramic micropump arranged at the bottom of the main reaction chamber, a MOX gas sensor array arranged inside the main reaction chamber, and a dynamic gas distribution algorithm, which are used to simulate the air in the natural environment;

[0029] The temperature and humidity control submodule includes a graphene heating film installed on the inner surface of the soil sample chamber, an infrared temperature measurement array electrically connected to the graphene heating film, a semiconductor cooling plate and an ultrasonic humidifier arranged in the soil sample chamber, which are used to simulate the temperature and humidity in a natural environment.

[0030] The illumination / rainfall simulation submodule includes a full-spectrum LED array and a rainfall nozzle array arranged on the top of the main reaction chamber to simulate illumination and rainfall in a natural environment.

[0031] Furthermore, the semiconductor refrigeration plate and the microwave thawing head form a freeze-thaw cycle module, which is used to simulate the freeze-thaw cycle contaminated soil working conditions, control the main reaction chamber to achieve a rapid temperature change of -50°C to 80°C, and perform pre-freezing and directional thawing of the soil;

[0032] When the LIBS laser spectrum module detection value is outside the preset bacterial community dispersion index value range, the bacterial community distribution is discrete, and the freeze-thaw temperature is adjusted through the semiconductor refrigeration chip and microwave thawing head to adapt to the life of the bacterial community;

[0033] The microbial remediation module consists of a UV sterilization head, a photoacoustic monitoring head, and a bioprinting head, which are used to provide a sterile environment and monitor microbial metabolic activity.

[0034] The dynamic response module includes a hydraulic servo vibration table and a fiber Bragg grating sensor. The hydraulic servo vibration table is installed at the bottom of the main reaction chamber, and the soil sample chamber is placed on the hydraulic servo vibration table to achieve vibration control. The fiber Bragg grating sensor is implanted into the contaminated soil sample during the soil sample preparation process through the fiber Bragg grating sensor implant head to monitor the strain of the contaminated soil sample.

[0035] During freeze-thaw cycle simulations, when the EIT electrical impedance imaging module detects that the ice crystal ratio is greater than 70%, the dynamic response module vibrates to assist in crack expansion. When the fiber grating sensor detects that the soil strain is lower than the first strain setting value, the semiconductor refrigeration plate is activated for emergency pre-freezing.

[0036] After the user sets the earthquake / traffic load through the edge intelligent decision-making platform, the vibration control of the hydraulic servo excitation table of the dynamic response module is used to simulate the redistribution of pollutants after the earthquake / traffic load;

[0037] Before the microbial remediation simulation, when the fiber Bragg grating sensor detects that the soil strain is lower than the second strain setting value, the hydraulic servo excitation table is controlled to perform pre-vibration before the microbial remediation is performed;

[0038] The freeze-thaw cycle module, microbial remediation module, and dynamic response module all interact with the edge intelligent decision-making platform.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] (1) Through the deep coupling of multiple disciplines of environmental engineering, artificial intelligence and precision machinery and the closed-loop design of the entire chain of perception-decision-execution-control, this invention provides a new generation of intelligent, standardized and full-scenario technical equipment system for the field of soil pollution prevention and control, which has significant industrial application value and technological leadership significance.

[0041] (2) The system of the present invention introduces a robotic arm collaborative operation module, a multimodal sensing module and a full natural environment control module and interacts with the edge intelligent decision-making platform. It can not only realize the automatic preparation of high homogeneity of complex soil samples such as petroleum contaminated soil, solidified petroleum contaminated soil, freeze-thaw cycle contaminated soil and microbial remediation contaminated soil, but also realize the simulation of actual working conditions of freeze-thaw cycle, dynamic response and microbial remediation. The simulation conditions match the actual working conditions, and the information of the migration behavior of pollutants in various petroleum contaminated soil samples throughout the whole cycle is accurately obtained. t = 0 to capture the initial migration process of the volatile / liquid phase.

[0042] (3) The present invention collects and analyzes the molecular composition of pollutants, concentration changes of elements and phase changes through a multimodal sensing module, and constructs X / Y / Z (space), time ( t ), physical phase (gas phase / liquid phase / adsorption phase), obtain multi-dimensional pollutant information, comprehensively perceive pollutant characteristics, improve the selectivity and stability of sensors in monitoring polycyclic aromatic hydrocarbons molecules, enhance the efficiency of capturing key pollutant information, and truly reflect the migration law of pollutants.

[0043] (4) The system of the present invention can input the required contaminated soil working condition setting values ​​on the edge intelligent decision-making platform, such as oil pollutant concentration, oil polluted soil composition, oil polluted soil density, strength, ambient temperature, humidity, light intensity, and rainfall intensity, and control the robotic arm collaborative operation module to automatically prepare the contaminated soil. At the same time, the full natural environment control module is started to control the environmental parameters in the main reaction chamber. For example, when simulating the oil leakage pollution scene of Dagang Oilfield in Binhai New Area, Tianjin, it is necessary to prepare oil polluted soil with the same actual oil pollution concentration and meet the working conditions of the oil polluted soil in Dagang Oilfield under the freeze-thaw cycle. In the simulation system of the present invention, it is possible to fully realize the automatic preparation of various contaminated soil samples according to actual needs and simulate the full-cycle migration behavior monitoring of pollutants under different working conditions, thereby solving the shortcomings of the existing technology that can only monitor the pollutant transformation process by simulating precipitation, atmospheric deposition and other pollution methods, and cannot realize the automatic preparation of contaminated soil online. The present invention can realize the contaminated soil migration monitoring under different actual engineering application working conditions.

[0044] (5) The present invention transforms pollutant information into a quantitative migration model through an edge intelligent decision-making platform, and previews the migration path of pollutants in the next 72 hours, achieving the effect of predictive remediation and accurately reproducing the migration path of pollutants from t =0, and dynamically adjusts environmental parameters and soil conditions based on real-time monitoring data, solving the difficult problem of precise control and precise management in the field of soil remediation.

[0045] (6) The present invention simulates various outdoor environmental conditions and soil sample requirements and simultaneously reproduces the migration path of petroleum pollutants indoors. It is an integrated equipment from in-situ preparation to in-situ monitoring, and also a complete integrated system of automatic preparation, monitoring and control. It provides a good solution to the current problem of being unable to timely predict the migration degree of pollutants in different soil bodies. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a structural schematic diagram of an embodiment of the experimental system for simulating the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil according to the present invention.

[0047] Figure 2 This is a control logic block diagram of the components of an embodiment of the system of the present invention.

[0048] Figure 3 Schematic diagram of the interaction between the components in the robotic arm collaborative operation module and the edge intelligent decision-making platform in the present invention.

[0049] Figure 4 This is a schematic diagram of the interaction logic between the freeze-thaw cycle module, microbial remediation module, and power response module and the edge intelligent decision-making platform in the present invention.

[0050] In the figure: 1. Main reaction cabin; 2. Circular magnetic levitation track; 3. Robotic arm collaborative operation module; 4. Soil sample room; 5. Multimodal sensing module; 6. Full natural environment control module; 7. Edge intelligent decision-making platform. DETAILED DESCRIPTION

[0051] The present invention is further explained below with reference to the embodiments and drawings, but they are not intended to limit the scope of protection of the present application.

[0052] The present invention can simulate pollution scenarios of different contaminated sites through a complete set of simulation equipment to collect and analyze the molecular composition of pollutants, concentration changes of elements, and phase changes. It can also automatically prepare contaminated soil, regulate environmental parameters to adapt to the actual conditions of contaminated sites, generate quantitative models of pollutant migration behavior, provide full-cycle migration data of pollutants and predict migration effects, and carry out remediation interventions.

[0053] The present invention is a simulation experimental system for the migration path of polycyclic aromatic hydrocarbons in oil-contaminated soil (see Figure 1 and Figure 2 ), including a main reaction cabin 1, a robotic arm collaborative operation module 3, a multimodal sensing module 5, a full natural environment control module 6, and an edge intelligent decision-making platform 7. The top of the main reaction cabin 1 is provided with a circular magnetic levitation track 2, and the robotic arm collaborative operation module 3 is arranged at the bottom of the circular magnetic levitation track 2. The robotic arm collaborative operation module can move in space on the circular magnetic levitation track. A soil sample room 4 is arranged at the bottom of the main reaction cabin 1, and the robotic arm collaborative operation module 3 prepares contaminated soil samples in the soil sample room 4. The full natural environment control module 6 is arranged inside the main reaction cabin 1, and the edge intelligent decision-making platform 7 is arranged on the side of the main reaction cabin 1.

[0054] The full natural environment control module is used to simulate different natural environments; the full natural environment control module interacts with the edge intelligent decision-making platform;

[0055] The multimodal sensing module includes an EIT electrical impedance imaging module, a LIBS laser spectroscopy module, and a terahertz time-domain spectroscopy module, which are used to measure the global conductivity distribution of the soil in the soil sample room and invert the soil sample porosity and pollutant concentration field in the pore water, analyze the C / H / O element concentration of pollutants, and specifically detect polycyclic aromatic hydrocarbon molecules;

[0056] The robotic arm collaborative operation module includes a main robotic arm, a micro-manipulator arm and a quick-change end-type tool library. The main robotic arm is installed at the bottom of the circular magnetic levitation track and adopts a six-degree-of-freedom serial configuration to achieve macro-operation. The quick-change end-type tool library equipped with a flange at the end of the main robotic arm has nine types of tool heads. The different types of tool heads are in a storage state when not in use, including a biological printing head, a microwave thawing head, a multi-channel injection head for injecting additives and soil materials, an acoustic compaction head for increasing the compaction of soil samples during soil sample preparation, a laser sintering head for enhancing the consolidation of soil samples during soil sample preparation, an ultrasonic dispersion head for mixing soil during soil sample preparation, an ultraviolet disinfection head for providing a sterile environment for the main reaction chamber, a fiber grating sensor implantation head for implanting fiber grating sensors during soil sample preparation, and a self-healing agent injection head for injecting nutrients and self-healing agents;

[0057] The multi-channel injection head includes channels for injecting solid soil material, which are connected to the solid soil storage tank via pipes. It also includes channels for injecting various additives (such as curing agents and freeze-thaw media), which are connected to the corresponding additive storage tanks for material supply. During the automated sampling process, additives and soil material are injected through the multi-channel injection head. The soil sample is mixed using an ultrasonic dispersion head, and the fiber Bragg grating sensor is implanted using a fiber Bragg grating sensor implant head. During the sample preparation process, the sample is compacted using an acoustic wave compaction head.

[0058] One end of the micromanipulator is nested inside the third joint of the main robotic arm to achieve microscopic operations. Different types of integrated components are fixed to the end of the micromanipulator. The different types of integrated components are in a storage state when not in use. The integrated components at the end of the micromanipulator include a LIBS laser focusing lens for assisting the LIBS laser spectroscopy module to emit laser, a six-dimensional force sensor for monitoring soil stress, a photoacoustic monitoring head for monitoring microbial activity, and a capacitive humidity sensor for monitoring the humidity in the main reaction chamber.

[0059] During the soil sampling process, the main robotic arm can extend into the soil sampling chamber and can move freely in the working space with the micro-manipulator arm.

[0060] The edge intelligent decision-making platform proactively adjusts system parameters based on real-time data analysis from its multimodal sensing module, all-natural environment control module, and robotic arm collaborative operation module. This allows for dynamic control and implements engineering measures to address soil degradation and remediate it. Acting as the system's "intelligent central nervous system," it not only monitors pollutants but also implements remediation measures.

[0061] As an embodiment, the edge intelligent decision-making platform includes an LSTM-GRU hybrid model and a digital twin module, and the digital twin module includes a multi-physics field coupling solver, a PPO reinforcement learning algorithm and a three-dimensional visualization interface.

[0062] The LSTM-GRU hybrid model is used to predict the migration behavior of pollutants based on the time-varying data obtained by the multimodal sensing module, the full natural environment control module, and the robotic arm collaborative operation module;

[0063] The input data of the LSTM-GRU hybrid model includes pollutant characteristic data, environmental parameters and soil state data, and outputs the predicted migration behavior of pollutants;

[0064] The pollutant characteristic data include conductivity distribution, C element concentration, H element concentration, O element concentration, polycyclic aromatic hydrocarbon molecule concentration, pollutant phase state, etc.; among them, the conductivity distribution is measured by the EIT electrical impedance imaging module; the C element concentration, H element concentration, and O element concentration are measured by the LIBS laser spectroscopy module; the polycyclic aromatic hydrocarbon molecule concentration and pollutant phase state are obtained by the terahertz time-domain spectroscopy module.

[0065] The environmental parameters include temperature gradient, humidity, O2 concentration, N2 concentration, CO2 concentration, gas flow rate, light intensity, rainfall intensity, etc.;

[0066] The soil state data include soil sample porosity, permeability coefficient, soil stress, microbial activity, etc.

[0067] The LSTM-GRU hybrid model outputs the predicted migration behavior of pollutants, which includes the future pollutant concentration field. C pred (x,y,z,t) , migration distance (unit: cm), migration rate (unit: cm / h), risk level (level I to IV), control parameter vector, diffusion coefficient, permeability, thermal conductivity, etc.; Among them, x,y,z represents three-dimensional space, t Indicates time;

[0068] The control parameter vector includes gas flow rate adjustment amount, heating area weight distribution, rainfall intensity, etc.

[0069] A digital twin model was established using COMSOL software using a multiphysics coupling solver. This solver transforms interfering physical processes into a computable set of coupled equations. The pollutant migration behavior predicted by the LSTM-GRU hybrid model was input into the digital twin model to generate simulation data. This data and the predicted data were then fed into a PPO reinforcement learning algorithm, which dynamically optimizes the control parameter vector to generate targeted control instructions. The control parameter vector is the executable variable directly optimized by the PPO reinforcement learning algorithm.

[0070] After the PPO reinforcement learning algorithm is optimized, targeted control instructions are sent to the full natural environment control module and the robotic arm collaborative operation module. The targeted control instructions include gas flow rate adjustment, heating area weight distribution, rainfall intensity, tool head type, tool head action coordinates and action volume;

[0071] The three-dimensional visualization interface is used to map and display the migration status of pollutants in real time.

[0072] As an example, the input data of the multiphysics coupling solver includes the diffusion coefficient D , permeability K , thermal conductivity k ,temperature T , gas flow rate v , rainfall intensity R rain , initial concentration field C 0 (x,y,z) , phase distribution, soil porosity, permeability coefficient, soil strain, etc. The output data of the multi-physics coupling solver include pollutant concentration field C(x,y,z,t) , temperature field T(x,y,z,t) , fluid seepage velocity field v s (x,y,z,t) , stress field σ(x,y,z,t) , strain field ε(x,y,z,t) , latent heat of phase change .

[0073] As an embodiment, the PPO reinforcement learning algorithm is used to dynamically optimize the control parameter vector, such as the gas flow rate adjustment amount, the heating area weight distribution, and the rainfall intensity, and generate targeted control instructions based on real-time environmental parameters, soil state data, and prediction data. T ),humidity( RH ), gas flow rate ( v ))、Porosity of soil sample ( ) and the pollutant concentration field for the next 72 hours output by the LSTM-GRU hybrid model C pred(x,y,z,t) Input into the PPO reinforcement learning algorithm and use Construct a state space for the state vector and set the optimization target parameter to the gas flow rate adjustment (Range: ±50mL / min), heating area weight distribution (Range: 0-1), rainfall intensity R rain (Range: 0-200mm / h), set the reward function Directly link prediction paths and regulatory effects:

[0074]

[0075] in, =0.1 (energy consumption weight), =0.05 (temperature gradient smoothness weight); C pred The output of the LSTM-GRU hybrid model is the predicted pollutant concentration field for the next 72 hours; C real is the measured value of the multimodal sensing module; is the L2 norm, which is used to perform a quadratic penalty on the prediction deviation, and the path tracking term Force the model predictions to be close to the real physical process; It is the inverse of the optimal gas flow rate of the system. The slower the gas flow rate, the greater the penalty. The energy consumption penalty term Used to prevent overloading of system resources; Temperature gradient smoothness, control temperature parameter change smoothness, stability reward Used to punish the violent fluctuation of temperature field when the freeze-thaw interface changes suddenly.

[0076] As an embodiment, the EIT electrical impedance tomography module includes an EIT electrode ring consisting of a 128-electrode annular array, which surrounds the outside of the soil sample chamber and is used to measure the global conductivity distribution of the soil, invert the porosity of the soil sample and the pollutant concentration field in the pore water, and the 128-electrode annular array is arranged at equal intervals;

[0077] The LIBS laser spectroscopy module is installed at the lower end of the main robotic arm. The LIBS laser spectroscopy module is vertically aligned with the soil sample surface and coaxially arranged with the EIT electrode ring to analyze the C / H / O element concentrations of pollutants.

[0078] The transmitting and receiving ends of the terahertz time-domain spectroscopy module are symmetrically distributed on the left and right sides of the soil sample room, 10 cm away from the soil sample surface, forming a transmission detection path for the specific detection of polycyclic aromatic hydrocarbon molecules.

[0079] As an embodiment, the full natural environment control module includes an air path control submodule, a temperature and humidity control submodule, and a light / rainfall simulation submodule;

[0080] The gas circuit control submodule includes a piezoelectric ceramic micropump arranged at the bottom of the main reaction chamber, a MOX gas sensor array arranged inside the main reaction chamber, and a dynamic gas distribution algorithm, which is used to simulate the air in the natural environment and support dynamic gas distribution;

[0081] The temperature and humidity control submodule includes a graphene heating film installed on the inner surface of the soil sample chamber, an infrared temperature measurement array electrically connected to the graphene heating film, a semiconductor cooling plate and an ultrasonic humidifier set in the soil sample chamber, which are used to simulate the temperature and humidity in the natural environment and achieve gradient temperature control;

[0082] The illumination / rainfall simulation submodule includes a full-spectrum LED array and a rainfall nozzle array arranged on the top of the main reaction chamber to simulate illumination and rainfall in a natural environment.

[0083] As an embodiment, the edge intelligent decision-making platform interacts with the robotic arm collaborative operation module. For example, in the process of preparing contaminated soil samples, the set values ​​of soil strength and soil stress of the contaminated soil samples are preset. The six-dimensional force sensor at the end of the micro-manipulator arm transmits real-time soil stress data to the edge intelligent decision-making platform (i.e., force feedback data process). The real-time soil strength is calculated based on the real-time soil stress data. The edge intelligent decision-making platform calibrates the real-time data of the soil sample with the set value, generates a control instruction based on the difference between the real-time data and the set value, and transmits it to the robotic arm collaborative operation module to control the operation of the ultrasonic compaction head and the laser sintering head to change the soil strength and soil stress until the difference is within the preset range.

[0084] The UV disinfection and sterilization head in the terminal tool library is quickly replaced to perform UV disinfection in the main reaction chamber to obtain a sterile environment. The bioprinting head is used to transfer the bacterial solution to the set location of the contaminated soil under sterile operation. The photoacoustic monitoring head is located at the end of the micromanipulation arm and is used to monitor the metabolic activity of the microbial flora. It is suitable for simulating the working conditions of microbial remediation of contaminated soil.

[0085] During the soil sample preparation process, a fiber Bragg grating sensor is implanted through the fiber Bragg grating sensor implantation head. When the fiber Bragg grating sensor detects that the width of the main crack is less than the crack setting value, the biological printing head is immediately started to accurately print the bacterial liquid along the crack; when the photoacoustic monitoring head detects that the microbial migration rate is less than the migration rate setting value, the hydraulic servo excitation table of the power response module is immediately started to provide vibration stimulation to increase the microbial migration rate.

[0086] As an embodiment, the semiconductor refrigeration plate and the microwave thawing head constitute a freeze-thaw cycle module. The microwave thawing head is a tool head of one of the quick-change end-type tool libraries. The freeze-thaw cycle module is used to achieve rapid temperature changes of -50°C to 80°C and microwave directional thawing in the main reaction chamber. It is suitable for simulating freeze-thaw cycle contaminated soil working conditions, providing an optimal temperature window for bioremediation, and can also achieve pre-freezing of soil bodies and enhance the stability of the dynamic loading structure.

[0087] When the LIBS laser spectrum module detection value is outside the preset bacterial community dispersion index value range, the bacterial community distribution is discrete, and the semiconductor refrigeration chip and microwave thawing head are used to immediately adjust the freeze-thaw temperature to adapt to the bacterial community life;

[0088] The microbial remediation module consists of a UV sterilization head, a photoacoustic monitoring head, and a bioprinting head, which are used to provide a sterile environment and monitor microbial metabolic activity.

[0089] The dynamic response module includes a hydraulic servo vibration table and a fiber Bragg grating sensor. The hydraulic servo vibration table is installed at the bottom of the main reaction chamber. The soil sample chamber is placed on the hydraulic servo vibration table to achieve vibration control. The fiber Bragg grating sensor is installed inside the soil sample chamber through the fiber Bragg grating sensor implant head to monitor the strain of the contaminated soil sample.

[0090] During freeze-thaw cycle simulations, when the EIT electrical impedance imaging module detects that the ice crystal ratio is greater than 70%, the dynamic response module vibrates to assist in crack expansion. When the fiber grating sensor detects that the soil strain is lower than the first strain setting value, the semiconductor refrigeration chip is immediately controlled to start emergency pre-freezing to enhance the stability of the soil structure.

[0091] After the user sets the earthquake / traffic load through the edge intelligent decision-making platform, the vibration control of the hydraulic servo excitation table of the dynamic response module is used to simulate the redistribution of pollutants after the earthquake / traffic load;

[0092] Before the microbial remediation simulation, when the fiber Bragg grating sensor detects that the soil strain is lower than the second strain setting value, the hydraulic servo excitation table is controlled to perform pre-vibration before the microbial remediation is performed;

[0093] The freeze-thaw cycle module, microbial remediation module, and dynamic response module all interact with the edge intelligent decision-making platform. The freeze-thaw cycle module feeds back the crack distribution map generated by the freeze-thaw process to the edge intelligent decision-making platform. The edge intelligent decision-making platform controls the biological printing head of the microbial remediation module to perform 3D printing along the cracks, enhances the penetration of the bacterial community through the freeze-thaw cracks, and obtains the bacterial community activity in the microbial metabolic process through the photoacoustic monitoring head. The edge intelligent decision-making platform adjusts the thawing rate of the freeze-thaw cycle module based on the feedback of the bacterial community activity to accurately match the bacterial community metabolic temperature requirements; when the photoacoustic monitoring head detects insufficient bacterial community metabolic activity, it controls the self-healing agent injection head at the end of the main robotic arm to inject nano-Fe3O4 enhancer to improve the pollutant degradation efficiency. The edge intelligent decision-making platform monitors whether the pollutant degradation rate meets the standard by obtaining the C element concentration changes obtained by the LIBS laser spectroscopy module in real time, thereby achieving soil remediation.

[0094] In the present invention, the first strain setting value, the second strain setting value, etc. can be set according to actual conditions.

[0095] Example 1:

[0096] This embodiment provides a simulation experimental system for the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil, including a main reaction chamber 1, a robotic arm collaborative operation module 3, a multimodal sensing module 5, a full natural environment control module 6, and an edge intelligent decision-making platform 7. The top of the main reaction chamber 1 is provided with a circular magnetic levitation track 2, the robotic arm collaborative operation module 3 is arranged at the bottom of the circular magnetic levitation track 2, a soil sample chamber 4 is provided at the bottom of the main reaction chamber 1, the robotic arm collaborative operation module 3 prepares contaminated soil samples in the soil sample chamber 4, the full natural environment control module 6 is arranged inside the main reaction chamber 1, and the edge intelligent decision-making platform 7 is arranged on the side of the main reaction chamber 1;

[0097] The multimodal sensing module includes an EIT electrical impedance imaging module, a LIBS laser spectroscopy module, and a terahertz time-domain spectroscopy module.

[0098] The EIT electrical impedance imaging module consists of a 128-electrode ring array embedded in the main reaction chamber at equal intervals, surrounding the outside of the soil sample chamber. The electrode spacing is 2 cm, and the AC excitation frequency is 10 kHz. The conductivity detection range is 0.01-10 S / m and the spatial resolution is 0.1 mm. 3 The EIT electrical impedance tomography module measures the global conductivity distribution of the soil, inverts the porosity of the soil sample and the pollutant concentration field in the pore water, and outputs the data in the form of a three-dimensional conductivity cloud map. The soil sample chamber is a cylindrical structure made of transparent quartz material.

[0099] The LIBS laser spectroscopy module is located at the lower end of the main robotic arm, vertically aligned with the soil sample surface and coaxially arranged with the EIT electrode ring. It emits a 50mJ pulsed laser to ablate the soil sample surface, collects plasma emission spectra, and analyzes the C / H / O element concentrations. The data is output as a heat map of the element concentration distribution.

[0100] The terahertz time-domain spectroscopy module's transmitter and receiver are symmetrically located on the left and right sides of the soil sample chamber, 10 cm from the sample surface, forming a transmission detection path. It features a 0.1-4 THz broadband emission source with a temporal resolution of 100 fs. It utilizes a molecular signature spectral database containing fingerprints for over 200 pollutants and employs dual-mode transmission / reflection detection. Polycyclic aromatic hydrocarbons (PAHs) are identified through molecular vibrational-rotational spectral fingerprints, such as the characteristic peak of benzopyrene at 2.5 THz. Data is output as molecular species and concentration distribution, achieving a specific PAH identification rate of >95%.

[0101] The terahertz time-domain spectroscopy module preferably includes a spectral feature extraction model that divides the terahertz spectrum into multiple frequency bands (e.g., 0.1 THz intervals). A self-attention mechanism is used to model inter-band associations, calculate correlations between bands, and convert similarities into probability distributions. This is then used to adaptively focus on key frequency bands (e.g., the 2.5 THz fingerprint region of benzopyrene, a major pollutant in polycyclic aromatic hydrocarbons), suppressing irrelevant noise. Convolution kernels of varying scales (e.g., 1×1, 3×3, and 5×5) are used to extract multi-granularity terahertz spectral features. The fused results are then obtained through weighted fusion of channel attention, enabling the simultaneous capture of both narrow peaks (molecular vibrations) and broad peaks (lattice vibrations), enhancing feature representation.

[0102] The multimodal sensing module achieves millisecond-level synchronization through trigger signals from physical circuit components. The EIT electrical impedance tomography module's data communication interface transmits conductivity distribution data via an RS-485 bus; the LIBS laser spectroscopy module's data communication interface transmits spectral data via optical fiber; and the terahertz time-domain spectroscopy module's data communication interface transmits time-domain waveform data via a PCIe interface. The main robotic arm moves the LIBS laser spectroscopy module along a pre-set trajectory, triggering a full-section scan from the EIT electrical impedance tomography module. The terahertz time-domain spectroscopy module then initiates transmission detection.

[0103] The robotic arm collaborative operation module includes a main robotic arm, a micro-manipulator arm, and a quick-change end-of-tool library.

[0104] The main robotic arm is installed at the bottom of the circular magnetic levitation track and adopts a 6-degree-of-freedom serial configuration to achieve macro-operation. The quick-change end-type tool library equipped with a flange at the end of the main robotic arm has nine types of tool heads. Different types of tool heads are in storage when not in use, including biological printing heads, microwave thawing heads, multi-channel injection heads for injecting additives and soil materials, ultrasonic compaction heads for increasing the compaction of soil samples during soil sample preparation, laser sintering heads for enhancing the consolidation of soil samples during soil sample preparation, ultrasonic dispersion heads for mixing soil during soil sample preparation, ultraviolet disinfection heads for providing a sterile environment for the main reaction chamber, fiber optic Bragg grating sensor implantation heads for implanting fiber optic Bragg grating sensors during soil sample preparation, and self-healing agent injection heads for administering nutrients and self-healing agents.

[0105] One end of the micromanipulator is nested inside the third joint of the main robotic arm to achieve microscopic operations. The end of the micromanipulator is fixed with different types of integrated components that are integrated with the micromanipulator. The different types of integrated components are in a storage state when not in use. They include a LIBS laser focusing lens for assisting the LIBS laser spectroscopy module to emit laser, a six-dimensional force sensor for monitoring soil stress, a photoacoustic monitoring head for monitoring microbial activity, and a capacitive humidity sensor for monitoring the humidity in the main reaction chamber.

[0106] The interaction between the components in the robotic arm collaborative operation module and the edge intelligent decision-making platform can be seen in Figure 3 The edge intelligent decision-making platform directly controls the operation of the main robotic arm, micro-manipulator arm and quick-change end-of-tool library. The main robotic arm realizes macro-operation and the micro-manipulator arm realizes micro-operation to meet the macro-micro motion coupling work. The main robotic arm is synchronized with the tool head status of the quick-change end-of-tool library, and the micro-manipulator arm sends force feedback data to the edge intelligent decision-making platform.

[0107] The edge intelligent decision-making platform includes an LSTM-GRU hybrid model and a digital twin module.

[0108] The LSTM-GRU hybrid model is used to predict the migration behavior of pollutants based on the time-varying data obtained from the multimodal sensing module, the full natural environment control module, and the robotic arm collaborative operation module.

[0109] The digital twin module includes a multi-physics field coupling solver, a PPO reinforcement learning algorithm, and a three-dimensional visualization interface.

[0110] A digital twin model was established using COMSOL software using a multi-physics coupling solver, transforming interfering physical processes into a computable set of coupled equations. The pollutant migration behavior predicted by the LSTM-GRU hybrid model was input into the digital twin model to obtain simulation data. This data and the predicted data were then fed into the PPO reinforcement learning algorithm, which dynamically optimized the control parameter vector to obtain targeted control instructions.

[0111] After the PPO reinforcement learning algorithm is optimized, targeted control instructions are sent to the full natural environment control module and the robotic arm collaborative operation module. The targeted control instructions include gas flow rate adjustment, heating area weight distribution, rainfall intensity, tool head type, tool head action coordinates and action volume;

[0112] The three-dimensional visualization interface is used to map and display the migration status of pollutants in real time.

[0113] The interaction process between the edge intelligent decision-making platform and the multimodal sensing module is as follows: the multimodal sensing module transmits the pollution source data to the LSTM-GRU hybrid model, the LSTM-GRU hybrid model transmits the data to the digital twin module to generate control instructions, and sends the control instructions to the full natural environment control module and the robotic arm collaborative operation module for actual control, providing status feedback to the multimodal sensing module to form a closed-loop control system.

[0114] The construction process of the digital twin model includes:

[0115] A digital twin model is established based on COMSOL software using a multi-physics coupling solver. The multi-physics coupling solver solves the thermal-gas-mechanical three-field coupling equations, which include the convection-diffusion equation, the energy conservation equation, and Darcy's law.

[0116] The convection-diffusion equation describes the concentration of substances. C As time goes by, two transport mechanisms, diffusion and convection, are considered:

[0117]

[0118] in: C is the pollutant concentration; t It’s time; D is the diffusion coefficient, which indicates the diffusion ability of a substance in a medium; v s It is the fluid seepage velocity field, which indicates the convection velocity of the material moving with the fluid, and is related to the gas flow rate. v The relationship can be seen as ; is the divergence operator, which is used here to describe the flux of diffusion and convection.

[0119] The energy conservation equation describes the temperature T Changes over time, taking into account the effects of heat conduction and heat sources:

[0120]

[0121] in: T It is the temperature; t It’s time; is the density of the medium; is the specific heat capacity of the medium; is thermal conductivity, which indicates the ability of a medium to conduct heat; It is the heat source term per unit volume, which represents the heat generated by various reasons (such as chemical reaction, electric heating, etc.); is the divergence operator.

[0122] Darcy's law describes the flow of fluid in porous media and gives the fluid seepage velocity field v s The expression:

[0123]

[0124] in: K is the permeability of the porous medium, which represents the resistance of the medium to fluid flow; is the dynamic viscosity of the fluid; P is the pressure of the fluid; is the density of the fluid; g is the acceleration due to gravity; z is the vertical coordinate; is the divergence operator, which is used here to consider the influence of gravity.

[0125] This embodiment introduces cross-coupling terms into the above three classical equations to reflect the interaction of multiple physical fields, specifically:

[0126] Add a temperature dependency to the convection-diffusion equation: ;

[0127] in: D 0 is the diffusion coefficient reference value at the reference temperature; R is the ideal gas constant; Ea is the activation energy, reflecting the effect of temperature on the diffusion coefficient; T Indicates temperature.

[0128] Then the convection-diffusion equation is:

[0129]

[0130] Darcy's law introduces dynamic correction of permeability:

[0131] in: is the temperature / porosity dependent permeability; is the porosity of the soil sample; is the coefficient of thermal expansion; K 0 is the reference state permeability benchmark value; T 0 is the reference temperature.

[0132] Then Darcy's law is:

[0133]

[0134] The energy conservation equation adds the latent heat of phase change:

[0135] in: L is the latent heat of phase change; is the ice phase ratio.

[0136] Then the energy conservation equation is:

[0137]

[0138] In addition, through the edge intelligent decision-making platform, traditional fixed parameters are upgraded to dynamic prediction parameters of the LSTM-GRU hybrid model, realizing real-time updates of pollution source data of the multimodal sensing module, environmental parameters of the full natural environment control module, soil stress, soil strength, soil sample porosity, diffusion coefficient, permeability, thermal conductivity, etc.

[0139] In this example, a five-dimensional pollutant migration prediction simulation was conducted for the digital twin model simulation verification. The input parameters were soil sample porosity 0.3, permeability coefficient 1×10 -5 m / s, initial concentration 5000 mg / kg, and comparison between simulation and measured data:

[0140]

[0141] The full natural environment control module includes an air path control submodule, a temperature and humidity control submodule, and a light / rainfall simulation submodule.

[0142] The gas control submodule includes a piezoelectric ceramic micropump located at the bottom of the main reaction chamber, a MOX gas sensor array located within the chamber, and a dynamic gas distribution algorithm. These components simulate natural air and support dynamic gas distribution. The piezoelectric ceramic micropump has 16 channels, a flow range of 0.1-800 mL / min, and a response speed of <10 ms. The MOX gas sensor array supports CO2, O2, and N2 with a detection accuracy of ±0.1%. The dynamic gas distribution algorithm utilizes PID adaptive regulation.

[0143] The temperature and humidity control submodule consists of a graphene heating film installed on the inner surface of the soil sample chamber, an infrared temperature measurement array electrically connected to the graphene heating film, a semiconductor cooling plate installed in the soil sample chamber, and an ultrasonic humidifier. These components simulate the temperature and humidity found in natural environments and achieve gradient temperature control. The graphene heating film is 0.2mm thick and has a resistance gradient of 0.5-2Ω / cm. The semiconductor cooling plate can be set to a minimum of -50°C. The ultrasonic humidifier's atomized particles have a diameter of 1-5μm and a humidity control range of 10-100% RH.

[0144] The illumination / rainfall simulation submodule includes a full-spectrum LED array and a rainfall nozzle array installed on the top of the main reaction chamber to simulate the illumination and rainfall in the natural environment. The full-spectrum LED array has a spectral coverage range of 280-1000nm and an irradiance of 0-2000μmol / m 2 / s; the rainfall intensity range of the rainfall nozzle array is 0-200mm / h, the raindrop particle size range is 0.5-5mm, and the terminal velocity range is 3-9m / s.

[0145] The installation of each device shall be based on the principle of not interfering with each other.

[0146] The edge intelligent decision-making platform issues control commands to the gas control submodule, temperature and humidity control submodule, and light / rainfall simulation submodule, putting them into operation. Gas composition feedback data is transmitted to the gas control submodule and the temperature and humidity control submodule. Temperature and humidity compensation is used to coordinate the control between the temperature and humidity control submodule and the light / rainfall simulation submodule. The light / rainfall simulation submodule transmits photothermal compensation to the gas control submodule.

[0147] The submodules within the full natural environment control module are connected to the edge intelligent decision-making platform via the CAN bus. The gas flow control submodule supports dynamic O2 / N2 / CO2 ratios, adjusts gas flow rates from 0.1 to 800 mL / min, and compensates for gas density based on temperature and humidity changes. For example, carrier gas flow rates are increased during high temperatures or low humidity to prevent uncontrolled pollutant volatilization. The temperature and humidity control submodule implements gradient temperature control (from -50°C to 80°C) and precise humidity control (from 10% to 100% RH). It also provides thermodynamic boundary conditions for light and rainfall, such as pre-cooling to prevent condensation before rainfall. The light and rainfall simulation submodule supports full-spectrum irradiance and rainfall intensity control. Synergistic effects are reflected in photothermal effect compensation, which automatically reduces the power of the graphene heating film when the light intensity is high, and in raindrop kinetic energy regulation, which matches wind speed to prevent trajectory deviation.

[0148] The process of pollution diffusion after simulating tropical rainstorms in the full natural environment control module is:

[0149] a. Start the rainstorm mode: The edge intelligent decision-making platform sends a control instruction to the light / rainfall simulation sub-module to start the rainstorm mode, setting the rainfall intensity to 200 mm / h and the raindrop size to 5 mm.

[0150] b. Request cooling compensation: After the light / rainfall simulation submodule is started, the ambient temperature may rise, so a cooling compensation request is made to the temperature and humidity control submodule to prevent condensation.

[0151] c. Adjust the intake air temperature: The temperature and humidity control submodule responds to the request and sends a command to the air path control submodule to adjust the intake air temperature from 40°C to 35°C to adapt to environmental changes.

[0152] d. Feedback on O2 concentration changes: The gas path control submodule monitors the O2 concentration in the environment and finds that the O2 concentration has decreased due to dissolved oxygen in rainfall, so it feeds this information back to the edge intelligent decision-making platform.

[0153] e. Increase O2 supply: Based on the feedback, the edge intelligent decision-making platform issues instructions to the gas path control submodule to increase the O2 supply from 21% to 25% to balance the dissolved oxygen content.

[0154] f. Increase humidity to simulate evaporation: The edge intelligent decision-making platform simultaneously sends instructions to the temperature and humidity control submodule, requiring the ambient humidity to be increased to 95% RH to simulate the evaporation process after rainfall.

[0155] g. Close local sprinklers to prevent oversaturation: After adjusting the humidity, the temperature and humidity control submodule sends a command to the light / rainfall simulation submodule to close local sprinklers to prevent oversaturation of the environment.

[0156] In special implementation cases, the present invention supports the implementation of a freeze-thaw cycle module, a microbial remediation module, and a dynamic response module.

[0157] The freeze-thaw cycle module includes a semiconductor refrigeration unit and a microwave thawing head; the microbial remediation module includes an ultraviolet disinfection head, a photoacoustic monitoring head and a biological printing head; and the dynamic response module includes a hydraulic servo excitation table and a fiber grating sensor.

[0158] The interaction logic between the freeze-thaw cycle module, microbial remediation module, and dynamic response module and the edge intelligent decision-making platform can be found in Figure 4 Each module sends its own data (such as temperature gradient data, bacterial activity data, and vibration spectrum data) to the edge intelligent decision-making platform. The edge intelligent decision-making platform generates optimization instructions based on this data and feeds the optimization instructions back to the corresponding module to achieve system optimization and coordination.

[0159] The independent function of the freeze-thaw cycle module is to achieve rapid temperature changes from -50°C to 80°C and microwave-directed thawing. The synergistic function is to provide an optimal temperature window for bioremediation, such as activating bacterial communities at 25°C and pre-freezing soil to enhance the stability of dynamically loaded structures. The independent function of the microbial remediation module is to 3D print bacterial liquid to the corresponding site using a bioprinter head and monitor microbial metabolic activity using a photoacoustic monitoring head. The synergistic function is to utilize freeze-thaw cracks to enhance bacterial penetration and increase microbial migration rate through vibration stimulation. The independent function of the dynamic response module is to achieve vibration control through a hydraulic servo excitation table and real-time strain monitoring through fiber grating sensors. The synergistic function is to simulate the redistribution of pollutants after earthquake / traffic loads and assist the expansion of freeze-thaw cracks through vibration.

[0160] Example 2:

[0161] The specific steps of the working process of this embodiment, taking the freeze-thaw-microbial collaborative restoration of petroleum pollution in permafrost areas as an example, are as follows:

[0162] a. Start the freeze-thaw cycle: The edge intelligent decision-making platform instructs the freeze-thaw cycle module to start the freeze-thaw cycle, raising the temperature from -30°C to 25°C for 5 cycles.

[0163] b. Feedback of crack distribution map: The freeze-thaw cycle module feeds back the soil crack distribution map generated by the freeze-thaw process to the edge intelligent decision-making platform.

[0164] c. 3D printed bacterial colonies: The edge intelligent decision-making platform instructs the microbial repair module to 3D print along the cracks, with a bacterial colony density of 1000 CFU / cm 3 .

[0165] d. Monitoring of metabolites: The microbial remediation module monitors and feeds back metabolites (such as increased CO2 concentration) to the edge intelligent decision-making platform.

[0166] e. Adjusting the thawing rate: The edge intelligent decision-making platform adjusts the thawing rate of the freeze-thaw cycle module based on the feedback of bacterial activity.

[0167] f. Inject enhancer: The edge intelligent decision-making platform instructs the self-healing agent injection head at the end of the main robotic arm to inject nano-Fe3O4 enhancer to improve the efficiency of pollutant degradation.

[0168] g. Return degradation rate: The microbial remediation module feeds back the pollutant degradation rate to the edge intelligent decision-making platform.

[0169] The interaction between the multimodal sensing module and the edge intelligent decision-making platform is reflected in the following steps:

[0170] In the first step, the EIT electrical impedance tomography module scans the soil sample's electrical conductivity in real time, the LIBS laser spectroscopy module collects elemental spectra, and the terahertz time-domain spectroscopy module detects the vibrational characteristics of polycyclic aromatic hydrocarbons (PAHs). The collected pollutant characteristic data is transmitted to the edge intelligent decision-making platform.

[0171] In the second step, the Kalman filter algorithm is used to align the timestamps of multi-source data, and PCA dimensionality reduction is used to extract key features, such as pollutant concentration gradient and diffusion rate.

[0172] The third step is to trigger the real-time control instructions of the edge intelligent decision-making platform when a sudden change in pollutant concentration is detected.

[0173] Example 3:

[0174] The working process of the petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system of this embodiment is:

[0175] S1. After the system is started, the robotic arm collaborative operation module first automatically prepares soil samples. By quickly replacing the multi-channel injection head in the terminal tool library, raw materials such as petroleum pollutants, curing agents, freeze-thaw media, and soil materials are transported to the soil sample chamber. Then, the combined action of the ultrasonic dispersion head and the laser sintering head is used to eliminate the agglomeration effect of soil particles. The EIT electrical impedance imaging module, LIBS laser spectroscopy module, and terahertz time-domain spectroscopy module are used to verify the in-situ quality of the soil. It is verified whether the porosity of the soil sample, the homogeneity of the pollutants, and the molecular phase state of the pollutants meet the target requirements, and a standardized soil sample that meets the target parameters is prepared, achieving high homogeneity sample preparation and completing the preparation quality verification.

[0176] S2. After the sample preparation is completed, the full natural environment control module is turned on to perform environmental simulation. During the environmental simulation process: the full-spectrum LED array outputs irradiation with a spectrum coverage range of 280-1000nm to simulate the natural environment lighting factors, and the rainfall nozzle array sprays according to the preset rainfall intensity. The raindrop particle size is 0.5-5mm and the terminal speed is 3-9m / s to simulate the influence of rainfall in nature. The graphene heating film and the semiconductor cooling plate cooperate to control the temperature to simulate the natural temperature factors. The ultrasonic humidifier controls the humidity to simulate the natural humidity influence. The piezoelectric ceramic micropump adjusts the O2 / N2 / CO2 partial pressure ratio to simulate the natural air partial pressure influence.

[0177] S3. After completing soil sample preparation and environmental simulation and control, the multimodal sensing module begins to dynamically monitor the five-dimensional migration data. The EIT electrical impedance imaging module scans once every 5 seconds to reconstruct the pore water conductivity field, and inverts the soil sample porosity and the pollutant concentration field in the pore water to collect spatial dimension data; the LIBS laser spectroscopy module penetrates the sampling point every 30 seconds to obtain the C / H / O element concentration ratio and element dimension data; the terahertz time-domain spectroscopy module scans continuously to identify characteristic peaks such as benzopyrene of polycyclic aromatic hydrocarbons and collect molecular dimension data.

[0178] The data identified by the multimodal sensing module are fused to construct a five-dimensional tensor model of X / Y / Z / time / phase. The characteristic vibration frequencies of polycyclic aromatic hydrocarbon molecules in the terahertz band are calculated through quantum chemistry, and these quantum-level spectral features are embedded in the neural network as physical constraints. The model is pre-trained on the QM9 dataset to learn the mapping relationship between molecular structure and spectrum. In practical applications, only terahertz time-domain spectral data needs to be input to infer the molecular structure and phase changes of pollutants, and analyze the phase migration path of pollutants.

[0179] S4. Finally, dynamic regulation and repair intervention are carried out. Based on the five-dimensional data of step S3, the LSTM-GRU hybrid model in the edge intelligent decision-making platform is used to predict the migration path in the next 72 hours. Then, the effect of the regulation strategy is simulated in real time through COMSOL software. Finally, the simulation results obtained by COMSOL software are compared with the soil state data set values ​​of the edge intelligent decision-making platform, and the environmental temperature, humidity and rainfall intensity parameters of step S2 are dynamically corrected to achieve closed-loop execution and regulation.

[0180] Based on the monitoring data of the multimodal sensing module, the real-time pollutant concentration field, migration rate, etc. are predicted. The instructions issued by the edge intelligent decision-making platform will be fed back in real time to the full natural environment control module and the robotic arm collaborative operation module for repair intervention. The data on the repair effect after the execution of the full natural environment control module and the robotic arm collaborative operation module will be returned to the edge intelligent decision-making platform for update calculation to complete the dynamic control of the system.

[0181] Example 4:

[0182] During the pollution remediation process of an oil field using the system of this application, the EIT electrical impedance imaging module detected an abnormal movement speed of the pollutant front (0.8cm / h→2.3cm / h). The edge intelligent decision-making platform immediately reduced the power of the graphene heating film from 500W to 300W to avoid thermal runaway.

[0183] Example 5:

[0184] In the freeze-thaw cycle simulation, the edge intelligent decision-making platform automatically switches to the "low temperature-low oxygen" mode (-30℃+5%O2) based on the simulation results of COMSOL software, inhibiting the effect of ice crystals on the diffusion of pollutants.

[0185] Example 6:

[0186] The interaction between the edge intelligent decision-making platform and the full natural environment control module in this embodiment is reflected in the following steps:

[0187] In the first step, the pollutant migration behavior is predicted by inputting the pollutant characteristic data obtained in real time by the multimodal sensing module, the real-time feedback of the infrared temperature measurement array of the graphene heating film, the mass flow sensor data of the piezoelectric ceramic micropump air path, and the capacitive humidity sensor into the LSTM-GRU hybrid model. The migration behavior of the pollutants predicted by the LSTM-GRU hybrid model is input into the digital twin model to obtain simulation data, and the spatiotemporal distribution of future pollutant concentrations and control parameter vectors, such as gas flow rate adjustment amount, heating area weight distribution, rainfall intensity, etc., are output.

[0188] In the second step, the control parameter vector is dynamically optimized through the PPO reinforcement learning algorithm, and targeted control instructions are generated based on the predicted data and real-time environmental parameters and soil state data.

[0189] In the third step, targeted control instructions are sent to the full-natural environment control module to adjust the control parameter vector. After executing the control instructions, the multi-dimensional parameter data of the system are monitored again. The monitoring data is used to update the LSTM-GRU hybrid model, digital twin model and optimize the PPO reinforcement learning algorithm.

[0190] The fourth step is to send temperature gradient control instructions to the temperature and humidity control submodule through the CAT bus, such as axial 5℃ / cm, and send gas ratio instructions to the gas path control submodule, such as O2 / N2=3:7.

[0191] Example 7:

[0192] The interaction between the edge intelligent decision-making platform and the robotic arm collaborative operation module in this embodiment is reflected in the following steps:

[0193] In the first step, based on the pollutant distribution data, for example, if the LIBS laser spectroscopy module detects that a local element exceeds the standard, the main robotic arm is activated to move to the contaminated area, and the end-of-line tool library is quickly replaced to switch to the self-healing agent injection head to inject the repair agent.

[0194] In the second step, the operation data, such as soil stress, is transmitted back to the edge intelligent decision-making platform through the six-dimensional force sensor to form a closed-loop calibration.

[0195] Example 8:

[0196] In the microbial remediation task, the main robotic arm automatically adjusted the injection path of the bioprinter head based on the polycyclic aromatic hydrocarbon molecular hotspot area (benzopyrene characteristic peak 2.5THz) identified by the terahertz time-domain spectroscopy module, thereby increasing the bacterial liquid coverage rate to 98%.

[0197] Example 9:

[0198] During the soil sample preparation process, the vibration of the ultrasonic dispersion head causes local temperature fluctuations in the soil. The edge intelligent decision-making platform uses semiconductor refrigeration chips to compensate for the temperature in real time and maintain temperature stability.

[0199] Example 10:

[0200] Simulating the migration of contaminated soil with different oil concentrations under rainfall conditions to verify the multimodal sensing module's ability to dynamically monitor the migration of oil-contaminated sand:

[0201] The first step is to prepare soil samples. Petroleum pollutants are injected into standard sand with a particle size of 0.1-0.5 mm at concentrations of 50 mg / g, 100 mg / g, and 150 m / g each time using a multi-channel injection head. The injection speed is set to 0.1 mL / s.

[0202] In the second step, a rainfall environment simulation was carried out. The graphene heating film was turned on, and the temperature gradient in the soil sample room was maintained at 50℃→150℃ axial distribution. N2 carrier gas was introduced with a flow rate set to 800mL / min. The rainfall simulation setting parameters were a rainfall intensity of 50mm / h and a duration of 30 minutes.

[0203] In the third step, the multimodal sensing module performs data acquisition. The EIT electrical impedance imaging module scans once every 10 seconds, and the parameters are set in the 128-electrode full activation mode. The LIBS laser spectroscopy module moves and scans along the Z axis every 5 minutes, and the parameters are set to a step size of 2mm and a pulse energy of 50mJ.

[0204] The fourth step was to verify the effectiveness. The terahertz time-domain spectroscopy module was used to image and display the migration path of the pollutant front. The minimum identifiable pore diameter was 0.1 mm. The LIBS laser spectroscopy module detected the C element concentration as low as 0.05%, indicating that the multimodal sensing module is sensitive to monitoring petroleum pollutants.

[0205] Example 11:

[0206] In this embodiment, ultrasonic dispersion and laser consolidation are used to prepare solidified petroleum contaminated soil, and its homogeneity and strength are effectively improved.

[0207] In the first step, the raw materials for preparing petroleum-contaminated soil (petroleum pollutants and saline soil) were mixed using an ultrasonic disperser. Then, lime and fly ash were added according to the mass ratio of saline soil, lime, and fly ash of 70:10:20. The mixture was then mixed using an ultrasonic disperser at a frequency of 40 kHz. The parameters were set to a cavitation intensity of 5 MPa and a mixing time of 10 minutes.

[0208] In the second step, the soil is consolidated by irradiating the laser sintering head to enhance the consolidation degree of the soil sample.

[0209] The third step involves curing and monitoring, maintaining a constant temperature and humidity during curing to produce solidified petroleum-contaminated soil. During curing, the temperature and humidity are maintained at 30°C and 95% RH, respectively. Fiber Bragg grating sensors monitor soil strain in real time, and the edge intelligent decision-making platform predicts soil strength development curves every hour.

[0210] The fourth step was to verify the effectiveness. XRD revealed a fly ash homogeneity distribution CV of 2.8%. Testing also showed a compressive strength of 28.5 MPa after seven days of curing. This demonstrates that the robotic arm collaborative module improves homogeneity and strength for the production of solidified petroleum-contaminated soil, enabling automated production of soil that meets specific requirements.

[0211] Example 12:

[0212] This example simulates a freeze-thaw cycle to produce contaminated soil, achieving system control and contaminant containment at extremely low temperatures. The freeze-thaw cycle simulation parameters are set as follows: a semiconductor refrigeration element cools the soil to -30°C within 30 minutes and maintains this temperature for two hours. A microwave thawing head then heats the soil to 25°C. During the thawing phase, a nano-SiO2 suspension is injected through a self-healing agent injection head to prevent the migration of petroleum contaminants. A graphene heating film assists in heating the soil during the thawing process. Once the thawing is complete, the next freeze-thaw cycle is performed to produce contaminated soil.

[0213] This application can prepare freeze-thaw cycle soil well under extreme environments, and the system has strong control performance, which can achieve the control of pollutant migration in freeze-thaw cycle soil.

[0214] Example 13: Indoor simulation and control integration of the entire process of petroleum pollutant migration

[0215] Application scenario: Construct a contaminated soil column in a soil sample room (1m×1m×1.5m) to study the migration patterns of petroleum pollutants in a high-temperature-rainfall coupled environment, and achieve migration suppression through multimodal sensing and intelligent control.

[0216] Step 1: The robotic arm collaborative operation module automatically prepares the contaminated soil sample, injects sand with a standard particle size of 0.1-0.5 mm through the multi-channel injection head, and adjusts the density to 1.6 g / cm 3The soil sample chamber is filled with soil stress data monitored by the six-dimensional force sensor of the micro-manipulator arm to provide real-time feedback on the compaction degree (target 50kPa±2%). The multi-channel injection head injects petroleum pollutants with a petroleum content of 100mg / g according to the preset mode, and the injection speed is set to 0.08mL / s. After the injection is completed, the acoustic compaction head is switched to automatically adjust the amplitude according to the preset soil sample porosity of 0.32. The six-dimensional force sensor provides real-time feedback on the compaction degree and stops after reaching the target value. Finally, the soil sample preparation quality is verified by the multimodal sensing module. The LIBS laser spectroscopy module scans the initial element concentration. The EIT electrical impedance imaging module detects the global conductivity distribution of the soil in the soil sample chamber and inverts the soil sample porosity and the pollutant concentration field in the pore water. When the inverted soil sample porosity reaches =0.32±0.01, the required contaminated soil sample preparation is complete. The real-time monitoring data from the multimodal sensing module is transmitted to the LSTM-GRU hybrid model of the edge intelligent decision-making platform. The digital twin module optimizes the control parameter vector and generates control instructions. These control instructions are then sent to the submodules of the full natural environment control module for dynamic environmental simulation.

[0217] The second step: Environmental simulation was performed using the fully natural environment control module. The graphene heating film in the temperature and humidity control submodule controlled the temperature in six zones (axial gradients of 50°C / 60°C / 80°C / 110°C / 130°C / 150°C). The carrier gas system (N2) flow rate in the gas path control submodule was set to 800 mL / min, and the rainfall intensity in the illumination / rainfall simulation submodule was set to 50 mm / h (equivalent to torrential rain conditions). Finally, the edge intelligent decision-making platform performed dynamic optimization, adjusting the temperature every five minutes to suppress pollutant volatilization and reducing the gas flow rate from 800 mL / min to 650 mL / min to minimize gas-phase migration. The edge intelligent decision-making platform verified the rationality of the simulation parameters and output a simulation completion signal.

[0218] In the third step, the multimodal sensing module monitors the migration process in real time. The EIT electrical impedance imaging module scans every 10 seconds to reconstruct the three-dimensional conductivity field. A sudden increase in conductivity (4.2→6.8 S / m) at Z=0.6m is detected and marked as a pollution front. The LIBS laser spectroscopy module detects the characteristic peak of the C-H bond at 247.8nm and finds an abnormal concentration at Z=0.8m. The theoretical value is 100 mg / g, and the measured value is 132 mg / g. The terahertz time-domain spectroscopy module scans the 0.5-3 THz frequency band to identify the petroleum phase. The 2.5 THz absorption peak intensity is used to invert the migration flux. Finally, data fusion and transmission are performed. The multi-source data monitored by the multimodal sensing module is synchronously transmitted through the TSN network, and the edge intelligent decision-making platform generates a migration heat map.

[0219] In the fourth step, the edge intelligent decision-making platform conducts dynamic regulation and repair intervention. First, the 72-hour migration path is predicted through the LSTM-GRU hybrid model, and then the digital twin module generates control instructions, so that the temperature and humidity control submodule establishes a reverse gradient (+5℃ / cm) at Z=0.6m, and the gas path control submodule operates the injection of O2 (concentration from 0→15%) to activate microbial degradation. The self-healing agent injection head implants nano-Fe3O4 enhancer and nutrient nano-SiO2 at Z=0.6m.

[0220] The fifth step is to issue and execute dynamic control instructions, synchronously control the dynamic power distribution of the graphene heating film of the temperature and humidity control submodule, and adjust the piezoelectric ceramic micropump of the gas path control submodule to adjust the O2 / N2 mixing ratio from 18:82 to 25:75. The main robotic arm is positioned to X=1.0m, Y=1.0m, and Z=0.8m. The end-tool library is quickly replaced and switched to the self-healing agent injection head to inject the self-healing agent. The six-dimensional force sensor detects soil stress in real time. Finally, real-time effect feedback is provided through the multimodal sensing module. After 6 hours of detection by the LIBS laser spectroscopy module, the CH bond concentration dropped to 98mg / g. The EIT electrical impedance imaging module detects that the migration rate of the pollution front has dropped to 0.18cm / h (inhibition rate 28%). The terahertz time-domain spectroscopy module's spectral liquid phase ratio has returned to 74% (gas phase migration has decreased).

[0221] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A simulation experimental system for the migration path of polycyclic aromatic hydrocarbons in petroleum-contaminated soil, characterized by: It includes a main reaction cabin, a robotic arm collaborative operation module, a multimodal sensing module, a full natural environment control module and an edge intelligent decision-making platform. A circular magnetic levitation track is set on the top of the main reaction cabin, and a soil sample room is set at the bottom of the main reaction cabin. The edge intelligent decision-making platform includes an LSTM-GRU hybrid model and a digital twin module; The LSTM-GRU hybrid model is used to predict the migration behavior of pollutants based on the time-varying data obtained by the multimodal sensing module, the full natural environment control module, and the robotic arm collaborative operation module; The multimodal sensing module includes an EIT electrical impedance imaging module, a LIBS laser spectroscopy module, and a terahertz time-domain spectroscopy module, which are used to measure the global conductivity distribution of the soil in the soil sample room and invert the porosity of the soil sample and the pollutant concentration field in the pore water, analyze the carbon, hydrogen and oxygen element concentrations of the pollutants, and specifically detect polycyclic aromatic hydrocarbon molecules; The full natural environment control module is used to simulate different natural environments and is located inside the main reaction chamber; The robotic arm collaborative operation module includes a main robotic arm, a micro-manipulation arm and a quick-change end-type tool library. The main robotic arm is installed on a circular magnetic levitation track for macro-operation; the end of the main robotic arm is connected to the quick-change end-type tool library through a flange. The quick-change end-type tool library includes nine types of tool heads. Different types of tool heads are in a storage state when not in use, including a biological printing head, a microwave thawing head, a multi-channel injection head for injecting additives and soil materials, an acoustic compaction head for increasing the compaction of soil samples during soil sample preparation, a laser sintering head for enhancing the consolidation of soil samples during soil sample preparation, an ultrasonic dispersion head for mixing soil during soil sample preparation, an ultraviolet disinfection head for providing a sterile environment for the main reaction chamber, a fiber grating sensor implantation head for implanting fiber grating sensors during soil sample preparation, and a self-healing agent injection head for injecting nutrients and self-healing agents; One end of the micromanipulator is nested in the main robotic arm for microscopic manipulation. A variety of integrated components are fixed to the end of the micromanipulator. These components are stored when not in use. These include a LIBS laser focusing lens for assisting the LIBS laser spectroscopy module in emitting laser light, a six-dimensional force sensor for monitoring soil stress, a photoacoustic monitoring head for monitoring microbial activity, and a capacitive humidity sensor for monitoring the humidity in the main reaction chamber. The edge intelligent decision-making platform uses real-time data obtained by the multimodal sensing module, the full natural environment control module and the robotic arm collaborative operation module to actively adjust system parameters and apply engineering measures to repair and intervene in soil deterioration.

2. The petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system according to claim 1 is characterized in that: The digital twin module includes a multi-physics field coupling solver, a PPO reinforcement learning algorithm, and a three-dimensional visualization interface; A digital twin model is established using a multi-physics coupling solver based on COMSOL software. The migration behavior of pollutants predicted by the LSTM-GRU hybrid model is input into the digital twin model to obtain simulation data. The simulation data and the predicted data are then input into the PPO reinforcement learning algorithm. The PPO reinforcement learning algorithm is used to dynamically optimize the control parameter vector to obtain targeted control instructions. After the PPO reinforcement learning algorithm is optimized, the targeted control instructions are sent to the full natural environment control module and the robotic arm collaborative operation module. The three-dimensional visualization interface is used to map and display the migration status of pollutants in real time.

3. The petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system according to claim 2 is characterized in that: The input data of the LSTM-GRU hybrid model includes pollutant characteristic data, environmental parameters and soil state data, and the output is the predicted migration behavior of pollutants; The pollutant characteristic data include conductivity distribution, C element concentration, H element concentration, O element concentration, polycyclic aromatic hydrocarbon molecule concentration and pollutant phase; The environmental parameters include temperature gradient, humidity, O2 concentration, N2 concentration, CO2 concentration, gas flow rate, light intensity and rainfall intensity; The soil state data include soil sample porosity, permeability coefficient, soil stress and microbial activity; The predicted migration behavior of pollutants includes future pollutant concentration field, migration distance, migration rate, risk level, control parameter vector, diffusion coefficient, permeability and thermal conductivity; The control parameter vector includes gas flow rate adjustment amount, heating area weight distribution and rainfall intensity; The targeted control instructions include gas flow rate adjustment amount, heating area weight distribution, rainfall intensity, tool head type, tool head action coordinates and action volume.

4. The petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system according to claim 1 is characterized in that: The EIT electrical impedance imaging module includes an EIT electrode ring consisting of a 128-electrode ring array, which surrounds the outside of the soil sample chamber; The LIBS laser spectroscopy module is mounted at the lower end of the main robotic arm. The LIBS laser spectroscopy module is vertically aligned with the soil sample surface and coaxially arranged with the EIT electrode ring. The transmitting end and receiving end of the terahertz time-domain spectroscopy module are symmetrically distributed on the left and right sides of the soil sampling room to form a transmission detection path.

5. The petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system according to claim 1 is characterized by: The full natural environment control module includes an air path control submodule, a temperature and humidity control submodule, and a light / rainfall simulation submodule; The gas path control submodule includes a piezoelectric ceramic micropump arranged at the bottom of the main reaction chamber, a MOX gas sensor array arranged inside the main reaction chamber, and a dynamic gas distribution algorithm, which are used to simulate the air in the natural environment; The temperature and humidity control submodule includes a graphene heating film installed on the inner surface of the soil sample chamber, an infrared temperature measurement array electrically connected to the graphene heating film, a semiconductor cooling plate and an ultrasonic humidifier arranged in the soil sample chamber, which are used to simulate the temperature and humidity in a natural environment. The illumination / rainfall simulation submodule includes a full-spectrum LED array and a rainfall nozzle array arranged on the top of the main reaction chamber to simulate illumination and rainfall in a natural environment.

6. The petroleum-contaminated soil polycyclic aromatic hydrocarbon migration path simulation experimental system according to claim 5, characterized in that: The semiconductor refrigeration plate and the microwave thawing head form a freeze-thaw cycle module, which is used to simulate the freeze-thaw cycle contaminated soil working conditions, control the main reaction chamber to achieve a rapid temperature change of -50°C to 80°C, and perform pre-freezing and directional thawing of the soil; When the LIBS laser spectrum module detection value is outside the preset bacterial community dispersion index value range, the bacterial community distribution is discrete, and the freeze-thaw temperature is adjusted through the semiconductor refrigeration chip and microwave thawing head to adapt to the life of the bacterial community; The microbial remediation module consists of a UV sterilization head, a photoacoustic monitoring head, and a bioprinting head, which are used to provide a sterile environment and monitor microbial metabolic activity. The dynamic response module includes a hydraulic servo vibration table and a fiber Bragg grating sensor. The hydraulic servo vibration table is installed at the bottom of the main reaction chamber, and the soil sample chamber is placed on the hydraulic servo vibration table to achieve vibration control. The fiber Bragg grating sensor is implanted into the contaminated soil sample during the soil sample preparation process through the fiber Bragg grating sensor implant head to monitor the strain of the contaminated soil sample. During freeze-thaw cycle simulations, when the EIT electrical impedance imaging module detects that the ice crystal ratio is greater than 70%, the dynamic response module vibrates to assist in crack expansion. When the fiber grating sensor detects that the soil strain is lower than the first strain setting value, the semiconductor refrigeration plate is activated for emergency pre-freezing. After the user sets the earthquake / traffic load through the edge intelligent decision-making platform, the vibration control of the hydraulic servo excitation table of the dynamic response module is used to simulate the redistribution of pollutants after the earthquake / traffic load; Before the microbial remediation simulation, when the fiber Bragg grating sensor detects that the soil strain is lower than the second strain setting value, the hydraulic servo excitation table is controlled to perform pre-vibration before the microbial remediation is performed; The freeze-thaw cycle module, microbial remediation module, and dynamic response module all interact with the edge intelligent decision-making platform.

Images