Polluted site remediation system and method based on artificial intelligence

Abstract

The invention discloses a polluted site remediation system and method based on artificial intelligence, and relates to the technical field of environmental engineering. According to the system, a physical flow field with a specific carrier frequency is established for a stratum through a flow field modulation injection subsystem, stratum response signals are collected through a synchronous response monitoring subsystem, and cross-correlation phase-locked demodulation operation is executed through an intelligent control terminal. The system extracts transmission time-lag characteristic parameters of stratum response signals relative to injected carrier waves, the transmission time-lag characteristic parameters serve as the unique physical criterion for identifying the stratum flow channel type, plugging materials are switched to be injected when fracture short-circuit flow is judged, and repairing agents are switched to be injected when matrix seepage flow is judged. According to the method, an active physical excitation and synchronous signal detection mechanism is established, and quantitative sensing and self-adaptive regulation and control of the underground micropore repairing state are achieved under the strong noise background.

Description

Technical Field

[0001] This invention relates to the field of environmental engineering technology, and in particular to an artificial intelligence-based system and method for remediating contaminated sites. Background Technology

[0002] In in-situ chemical oxidation or reduction remediation projects for groundwater and contaminated sites, the prevalent heterogeneity of the underground medium leads to a disconnect between the physical transport process and the information sensing data. At the physical level, underground aquifers are typically composed of a highly permeable fracture network and low-permeability matrix micropores. Under constant flow field conditions, remediation agents tend to penetrate rapidly along the fracture channels of least resistance, forming a dominant flow short-circuit and failing to effectively reach contaminants within the matrix micropores. At the information level, existing monitoring methods mainly rely on static measurements of solute concentration. Due to the masking effect of environmental background noise and the inherent ambiguity of concentration readings, high concentration readings may originate from matrix infiltration through effective diffusion of the agent to the monitoring point, or from a short-circuit process where the agent directly reaches the monitoring point along the fracture. This phenomenon, where monitoring data fails to accurately reflect the physical flow channel properties, makes it difficult for the control system to accurately determine the actual formation state, thus hindering the correct decision between injection and plugging. This results in the ineffective loss of remediation agents along the short-circuit channels, while contaminants within the micropores continue to be released after remediation, causing a rebound. Summary of the Invention

[0003] This invention provides an artificial intelligence-based contaminated site remediation system and method to address the technical problem that existing in-situ remediation technologies cannot accurately control heterogeneous strata by utilizing static monitoring data to detect short-circuit flow and matrix infiltration flow in the presence of strong noise.

[0004] In view of the above problems, in a first aspect, the present invention provides an artificial intelligence-based contaminated site remediation system, comprising: The flow field modulation injection subsystem is configured to inject a fluid medium into an underground formation and, in response to an input modulation control signal, drive an actuator to cause periodic fluctuations in the pressure or flow velocity of the fluid medium, thereby establishing a physical carrier flow field in the formation. A synchronous response monitoring subsystem is deployed in the downstream influence area of ​​the flow field modulation injection subsystem and is configured to continuously acquire the formation response signal after the physical carrier flow field is transmitted through the formation medium. The intelligent control terminal is communicatively connected to the flow field modulation injection subsystem and the synchronous response monitoring subsystem, respectively. The intelligent control terminal is configured to perform the following operations: A reference signal is generated, the frequency of which is synchronized with the fluctuation frequency of the physical carrier flow field; Phase-locked demodulation is performed on the formation response signal and the reference signal to extract the transmission time delay characteristic parameters of the formation response signal relative to the physical carrier flow field; The flow channel type of the current formation medium is identified based on the transmission time delay characteristic parameters, and feedback control commands are generated based on the identification results. The feedback control command is used to adjust the type of fluid medium injected by the flow field modulation injection subsystem.

[0005] Secondly, the present invention also provides an artificial intelligence-based method for remediating contaminated sites, comprising the following steps: The flow field modulation injection subsystem is used to drive the injection equipment to generate physical pressure pulse waves with a specific carrier frequency, thereby actively modulating the underground flow field. The downstream stratum response signal is collected using the synchronous response monitoring subsystem, and a reference signal synchronized with the carrier frequency is generated using the intelligent control terminal. The formation response signal and the reference signal are subjected to phase-locked demodulation to extract transmission time delay characteristic parameters; The transmission time delay characteristic parameter is used as a physical fingerprint to distinguish the flow channel type: when the parameter indicates that the time delay is close to zero, it is determined to be a short-circuit flow; when the parameter indicates that the time delay is within a preset range, it is determined to be a seepage flow. Closed-loop control is implemented based on the determined flow channel type: if it is determined to be a short-circuit flow, the injection medium is switched to a sealing material; if it is determined to be a permeable flow, the injection medium is switched to a repair agent.

[0006] The technical solution provided in this application has at least the following technical effects: By introducing an active frequency domain modulation and phase-locked detection mechanism, a deterministic mapping relationship is established that directly characterizes the physical properties of underground flow channels using transmission time delay characteristic parameters. Cross-correlation phase-locked operation utilizes the orthogonality principle to cancel out non-co-frequency environmental background noise in the integral domain, enabling the system to accurately capture the phase lag characteristics caused by the micropore mass transfer process without relying on the absolute magnitude of the concentration amplitude. This mechanism transforms the identification of formation flow channel types from fuzzy inference to quantitative determination based on physical fingerprints, ensuring that the system only performs chemical injection when effective matrix infiltration is confirmed, and immediately performs sealing when a short circuit occurs. This achieves precise closed-loop control of the heterogeneous formation remediation process, avoiding ineffective chemical dissipation and the risk of contamination rebound. Attached Figure Description

[0007] Figure 1 A schematic diagram of the overall hardware architecture of an artificial intelligence-based contaminated site remediation system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a contaminated site remediation method based on artificial intelligence, provided as an embodiment of the present invention. Detailed Implementation

[0008] The above technical solutions will now be described in detail with reference to the accompanying drawings and specific embodiments to provide a better understanding of them. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments used only to explain the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.

[0009] For examples, please refer to Figure 1 An artificial intelligence-based contaminated site remediation system includes: The flow field modulation injection subsystem is configured to inject a fluid medium into an underground formation and, in response to an input modulation control signal, drive an actuator to cause periodic fluctuations in the pressure or flow velocity of the fluid medium, thereby establishing a physical carrier flow field in the formation. A synchronous response monitoring subsystem is deployed in the downstream influence area of ​​the flow field modulation injection subsystem and is configured to continuously acquire the formation response signal after the physical carrier flow field is transmitted through the formation medium. The intelligent control terminal is communicatively connected to the flow field modulation injection subsystem and the synchronous response monitoring subsystem, respectively. The intelligent control terminal is configured to perform the following operations: A reference signal is generated, the frequency of which is synchronized with the fluctuation frequency of the physical carrier flow field; Phase-locked demodulation is performed on the formation response signal and the reference signal to extract the transmission time delay characteristic parameters of the formation response signal relative to the physical carrier flow field; The flow channel type of the current formation medium is identified based on the transmission time delay characteristic parameters, and feedback control commands are generated based on the identification results. The feedback control command is used to adjust the type of fluid medium injected by the flow field modulation injection subsystem.

[0010] The contaminated site remediation system based on active flow field modulation and phase-locked loop identification uses an intelligent control terminal as the core physical command center. The intelligent control terminal establishes a closed-loop control network via an industrial fieldbus protocol. The downlink of the closed-loop control network connects to the flow field modulation injection subsystem, and the uplink connects to the synchronization response monitoring subsystem. The intelligent control terminal is electrically hardwired to the source-end noise cancellation circuit.

[0011] The flow field modulation injection subsystem is constructed using a hydraulic pulse shaping architecture. A variable frequency constant pressure pump serves as the fluid power source, providing continuous and stable fluid pressure. The fluid outlet of the variable frequency constant pressure pump is physically connected to the inlet of a three-way switching valve. The injection port of the three-way switching valve is physically connected to an underground injection well via a high-pressure resistant pipeline. The return port of the three-way switching valve is connected to an atmospheric pressure return container via a low-pressure pipeline. The high-speed input / output module inside the intelligent control terminal is connected to the electromagnetic drive coil of the three-way switching valve via a signal cable. The intelligent control terminal sends a frequency-adjustable pulse width modulation signal or a square wave drive signal to the three-way switching valve.

[0012] When the three-way switching valve receives an opening command, its internal valve core actuates to connect the inlet and outlet, allowing pressurized fluid from the variable frequency constant pressure pump to enter the underground injection well and establish a high-pressure state within it. When the three-way switching valve receives a closing or reset command, its internal valve core actuates to cut off the outlet and connect the inlet and return outlet, unloading the fluid pressure in the underground injection well through the return outlet to the atmospheric pressure return container. The mechanical action response time of the three-way switching valve's valve core is set to less than 50 milliseconds. In conjunction with the continuous pressure supply from the variable frequency constant pressure pump, the three-way switching valve generates a square wave pressure pulse with both vertical rising and falling edges at the outlet. This square wave pressure pulse introduces harmonic components into the underground flow field.

[0013] The variable frequency constant pressure pump has a first feed line connected in parallel to the first storage tank of the repair agent and a second feed line connected to the second storage tank of the sealing gel material. A first electric feed valve is installed on the first feed line, and a second electric feed valve is installed on the second feed line. An intelligent control terminal is connected to both the first and second electric feed valves via control signal lines. During system operation, the variable frequency constant pressure pump operates continuously. Based on calculation results, the intelligent control terminal sends commands to open the first electric feed valve and close the second electric feed valve, or vice versa. By switching the opening and closing of the electric feed valves, the flow field modulation injection subsystem changes the type of fluid medium injected into the formation without stopping the pulse modulation operation.

[0014] The source-end noise cancellation circuit is physically embedded in the system's electrical connection architecture. A current transformer is snap-fitted onto any phase of the three-phase power supply cable of the variable frequency constant pressure pump drive motor. The current transformer senses the electromagnetic interference waveform generated during the operation of the variable frequency constant pressure pump in real time. The signal output terminal of the current transformer is connected to the input terminal of the inverting amplifier circuit via a shielded cable. The inverting amplifier circuit performs a 180-degree phase flip and gain adjustment on the acquired electromagnetic interference waveform. The output terminal of the inverting amplifier circuit is connected to the front end of the signal conditioning board of the synchronous response monitoring subsystem via electrical wiring. The front end of the signal conditioning board of the synchronous response monitoring subsystem is equipped with an analog adder circuit or a differential amplifier circuit. The underground electrochemical response signal and the inverted electromagnetic interference waveform are superimposed in the analog adder circuit or differential amplifier circuit. The superposition operation utilizes the wave interference cancellation principle to attenuate the common-mode interference component mixed in the underground electrochemical response signal. The intelligent control terminal reads the denoised digital signal and performs phase-locked demodulation operations in conjunction with the pulse modulation timing generated internally by the intelligent control terminal.

[0015] For examples, see Figure 2 An artificial intelligence-based method for remediating contaminated sites includes the following steps: The flow field modulation injection subsystem is used to drive the injection equipment to generate physical pressure pulse waves with a specific carrier frequency, thereby actively modulating the underground flow field. The downstream stratum response signal is collected using the synchronous response monitoring subsystem, and a reference signal synchronized with the carrier frequency is generated using the intelligent control terminal. The formation response signal and the reference signal are subjected to phase-locked demodulation to extract transmission time delay characteristic parameters; The transmission time delay characteristic parameter is used as a physical fingerprint to distinguish the flow channel type: when the parameter indicates that the time delay is close to zero, it is determined to be a short-circuit flow; when the parameter indicates that the time delay is within a preset range, it is determined to be a seepage flow. Closed-loop control is implemented based on the determined flow channel type: if it is determined to be a short-circuit flow, the injection medium is switched to a sealing material; if it is determined to be a permeable flow, the injection medium is switched to a repair agent.

[0016] Before sending a start command to the flow field modulation injection subsystem, the intelligent control terminal prioritizes running a full-field silent scan program. The intelligent control terminal sends a sampling command to the synchronous response monitoring subsystem and simultaneously locks the flow field modulation injection subsystem into a shutdown state. Under conditions of no artificial active flow field interference, the synchronous response monitoring subsystem continuously acquires electrochemical and pressure signals from the underground aquifer at a sampling rate higher than 100 Hz. The synchronous response monitoring subsystem encapsulates the acquired continuous time-series signals into an environmental baseline noise data sequence. The intelligent control terminal receives the environmental baseline noise data sequence via the communication bus.

[0017] The digital signal processing module inside the intelligent control terminal performs a Fast Fourier Transform (FFT) operation on the received environmental baseline noise data sequence. The FFT transforms the time-varying environmental baseline noise data sequence into complex spectral data that varies with frequency. The intelligent control terminal calculates the square of the modulus of the complex spectral data, thereby generating power spectral density (PSD) distribution data reflecting the frequency distribution of environmental noise energy. The PSD distribution data covers the entire frequency band from 0 Hz to the Nyquist frequency. The intelligent control terminal performs spectral entropy calculation on the PSD distribution data to quantify the degree of disorder in the environmental background noise. First, the intelligent control terminal normalizes the PSD distribution data by dividing the energy value of each frequency component by the total energy value of all frequency components, obtaining a normalized probability distribution vector. The intelligent control terminal iterates through each element in the normalized probability distribution vector, calculating the product of that element and its natural logarithm. Finally, the intelligent control terminal sums all the product results and takes the negative of the sum to obtain the Shannon spectral entropy value, which characterizes the complexity of the current underground environmental noise.

[0018] The intelligent control terminal executes frequency domain obstacle avoidance search logic based on power spectral density distribution data. This logic presets a low-frequency cutoff threshold to shield against groundwater level fluctuations caused by tidal effects below 0.01 Hz. It also presets a power frequency notch range to shield against industrial power electromagnetic interference at 50 Hz or 60 Hz. Within the remaining spectrum after removing frequency bands below the low-frequency cutoff threshold and the power frequency notch range, the intelligent control terminal executes a minimum value search algorithm. This algorithm traverses all frequency points within the remaining spectrum to find the frequency point with the lowest power spectral density amplitude. The intelligent control terminal identifies this frequency point as the energy valley frequency. This energy valley frequency is then locked as the optimal modulation frequency for the system. The intelligent control terminal writes the optimal modulation frequency into the control register of the flow field modulation injection subsystem and uses it as the frequency reference for subsequently generating the digital reference signal.

[0019] The intelligent control terminal reads the optimal modulation frequency value locked in the storage register. The microprocessor inside the intelligent control terminal calls either a direct digital frequency synthesis algorithm or a lookup table algorithm. Based on the optimal modulation frequency value, the microprocessor constructs two sets of discrete digital sequences with strict orthogonality. The first set of sequences is an in-phase reference sequence vector following a sine function. The second set of sequences is an orthogonal reference sequence vector following a cosine function. The in-phase reference sequence vector and the orthogonal reference sequence vector have completely identical time axis resolution and zero-phase starting point. These two sets of digital sequences are cached in the intelligent control terminal's fast memory, serving as the sole mathematical demodulation key for subsequent cross-correlation operations.

[0020] Within the same instruction cycle that generates the digital reference sequence, the intelligent control terminal initiates the level-flipping logic of the general-purpose input / output interface (GPIO). This logic maps the optimal modulation frequency value to a timing signal that switches between high and low levels. The GPIO outputs a physical voltage pulse train with a specific duty cycle. The rising edge of the physical voltage pulse train is aligned nanoseconds with the start of the positive half-cycle of the in-phase reference sequence vector. The frequency of the physical voltage pulse train is strictly equal to the optimal modulation frequency value. The physical voltage pulse train is then transmitted through a power amplifier circuit to the three-way switching valve of the flow field modulation injection subsystem.

[0021] The solenoid coil of the three-way switching valve performs mechanical opening and closing actions driven by a physical voltage pulse train. The fluid pressure within the injection well generates periodic high-pressure-unloading fluctuations following these mechanical actions. This establishes a physical carrier flow field in the underground aquifer with the optimal modulation frequency as the fundamental frequency. Since the excitation source signal of the physical carrier flow field and the digital reference sequence vector in the memory both derive from the same system clock source of the intelligent control terminal, a mandatory co-originating and in-phase causal relationship is established between the physical carrier flow field and the digital reference sequence vector. This co-originating and in-phase causal relationship eliminates clock drift errors between the transmitter and receiver, establishing the physical premise for subsequent cross-correlation phase-locked demodulation operations.

[0022] The intelligent control terminal continuously reads the original discrete response signal sequence transmitted back by the synchronous response monitoring subsystem via a digital communication interface. The original discrete response signal sequence contains useful electrochemical response components and superimposed environmental background noise components. The digital signal processing unit inside the intelligent control terminal retrieves the cached in-phase reference sequence vector and orthogonal reference sequence vector.

[0023] The digital signal processing unit first performs in-phase component extraction. The original discrete response signal sequence is multiplied point-by-point with the in-phase reference sequence vector. The multiplied values ​​are then summed over a sampling window covering an integer multiple of the signal period. This summation operation corresponds to the discrete definite integral operation in mathematics. The result is assigned as the in-phase component value. The discrete mathematical model followed by the in-phase component extraction operation is expressed as:

[0024] In the above mathematical expression, Defined at the sampling time The acquired raw discrete response signal values; Defined as the optimal modulation frequency currently locked by the system; Defined as a discrete sampling time point sequence; symbol Defined as an accumulation operation of all data points within the sampling window.

[0025] During the parallel period of the in-phase component extraction operation, the digital signal processing unit performs the quadrature component extraction operation. The original discrete response signal sequence is multiplied point-by-point with the orthogonal reference sequence vector. The multiplied values ​​are accumulated and summed within the same sampling window. The result is assigned as the quadrature component value. The discrete mathematical model followed by the orthogonal component extraction operation is expressed as: The cross-correlation operation utilizes the orthogonality principle of trigonometric functions to achieve narrowband filtering. This applies whenever the frequency is not equal to the optimal modulation frequency. The expected value of the ambient background noise component, after being multiplied by a sine or cosine function of the reference frequency and then periodically integrated, approaches zero. This is only true when the frequency is strictly equal to the optimal modulation frequency. The useful electrochemical response components retain non-zero values ​​after integration. Through this integration filtering mechanism, non-co-frequency noise is removed from the original discrete response signal sequence, and the system retains pure co-frequency response components in a strong noise background.

[0026] The intelligent control terminal uses the calculated in-phase component values Values ​​of orthogonal components A complex vector of the formation response is constructed. The intelligent control terminal performs modulus calculation on the complex vector of the formation response to extract the electrochemical response amplitude. The formula for calculating amplitude is as follows: The intelligent control terminal performs a four-quadrant arctangent calculation on the ratio of the quadrature component value to the in-phase component value, thereby extracting the transmission time delay characteristic parameter, i.e., the phase lag value. The formula for calculating the phase lag value is as follows: In this calculation logic, the electrochemical response amplitude The intensity characteristics of underground micropore connectivity were quantified; phase hysteresis value The time delay effect characteristics generated by the transmission of physical carrier flow field through underground medium were quantified.

[0027] The intelligent control terminal retrieves the amplitude saturation processing module from the micropore connectivity efficiency index calculation model. This module receives the electrochemical response amplitude data output from the cross-correlation operation. The electrochemical response amplitude data is first multiplied by a preset sensitivity adjustment coefficient. The result of the multiplication is inversely represented and used as the exponent of the natural constant to calculate the exponential decay term. This exponential decay term is added to the first value as the denominator, and the second value is used as the numerator for division. The quotient of the division is subtracted from the first value to obtain the normalized amplitude saturation factor. This calculation logic, based on a variant of the Sigmoid function, numerically constructs a diminishing marginal effect mechanism. When the electrochemical response amplitude data is in a low-level range, the amplitude saturation factor increases approximately linearly with increasing amplitude; when the electrochemical response amplitude data exceeds a certain threshold, the growth rate of the amplitude saturation factor gradually slows down and asymptotically converges to the first value. This saturation mechanism prevents high-amplitude artifacts caused by localized high-concentration accumulation near the sensor probe from excessively interfering with the overall evaluation results, ensuring that the evaluation results reflect broad regional connectivity rather than single-point concentration extremes.

[0028] The intelligent control terminal executes the phase matching factor calculation process in parallel or serially. The phase lag value output from the cross-correlation operation is used as the input variable. The phase deviation value is obtained by subtracting the preset target phase angle from the phase lag value. The target phase angle represents the theoretical time delay when the underground fluid reaches the convection and diffusion equilibrium state in the microporous matrix. The phase deviation value is squared and then divided by twice the square of the resonant bandwidth parameter. The negative of this quotient is used as the exponent of the natural constant to calculate the phase matching factor. This calculation logic based on the Gaussian radial basis function constructs a mathematical bandpass filter in the phase domain. Only when the detected phase lag value falls within a specific interval centered on the target phase angle and with the resonant bandwidth parameter as the tolerance range, the phase matching factor output is a high evaluation coefficient close to one. For short-circuit flow phase characteristics approaching zero or excessively large blockage flow phase characteristics, the phase matching factor rapidly decays to near zero. This filtering mechanism filters out effective microporous mass transfer signals at the physical mechanism level and eliminates invalid fracture DC signals.

[0029] The intelligent control terminal performs the final synthesis calculation of the micropore connectivity efficiency index based on a preset nonlinear mathematical model. The calculation of the micropore connectivity efficiency index follows the analytical expression below: In mathematical expressions, Defined as the micropore connectivity efficiency index; Defined as the global gain coefficient, it is used to adjust the numerical range of the output value; Defined as the electrochemical response amplitude extracted by cross-correlation calculation; Defined as the sensitivity adjustment coefficient, its physical dimension is amplitude. The reciprocal of is used to control the slope of the amplitude saturation curve; Defined as the phase lag value, i.e., the transmission time delay characteristic parameter; Defined as the target phase angle, it represents the theoretical hysteresis value when convection and diffusion reach equilibrium. Defined as the resonant bandwidth parameter, it has the same angular dimension as the phase lag value and is used to set the tolerance range of the effective phase. The intelligent control terminal stores the calculated micropore connectivity efficiency index into the dynamic decision register as the quantitative basis for subsequent closed-loop control.

[0030] The intelligent control terminal compares the calculated phase lag value with a preset short-circuit threshold stored in its internal register. If the phase lag value is less than the preset short-circuit threshold (e.g., 5 degrees) and the electrochemical response amplitude is higher than the minimum effective signal limit, the intelligent control terminal determines that the current underground flow field is in a short-circuit state in the dominant fracture channel. Physically, this means that the injected fluid does not pass through the effective impediment of the microporous matrix and directly reaches the monitoring point through the high-permeability fracture. It should be noted that when the phase lag is below a certain threshold or exceeds the upper limit of the preset range, it is determined to be "blocked or disconnected," and the system executes an alarm or high-pressure unblocking operation.

[0031] In response to the determination of a short-circuit state in the dominant fracture channel, the intelligent control terminal generates a sealing mode control command. This command is sent to the flow field modulation injection subsystem. The subsystem then closes the first electrically operated feed valve connected to the repair agent tank and simultaneously opens the second electrically operated feed valve connected to the sealing gel tank. The high-viscosity sealing material is then pumped into the underground fracture channel. Simultaneously, the intelligent control terminal increases the flow field modulation frequency to utilize the skin effect of the fluid to assist in the deposition of the sealing material on the fracture wall. A significant decrease in the monitored electrochemical response amplitude indicates that the short-circuit channel has been effectively physically blocked.

[0032] If the phase lag value falls within a preset effective range (e.g., between 30 and 60 degrees) defined by the target phase angle plus or minus the resonance bandwidth parameter, the intelligent control terminal determines that the current underground flow field is in a microporous permeable state of the matrix. Physically, this means that the injected fluid has successfully entered the low-permeability matrix and established an effective convection-diffusion balance.

[0033] In response to the determination of the micropore permeability status of the matrix, the intelligent control terminal generates a remediation mode control command. This command drives the flow field modulation injection subsystem to open the first electric feed valve and close the second electric feed valve, resuming the injection of the remediation agent. In this state, the intelligent control terminal uses the micropore connectivity efficiency index as a feedback adjustment variable. The output power of the variable frequency constant pressure pump or the duty cycle of the pulse modulation is dynamically adjusted in real time according to the magnitude of the micropore connectivity efficiency index. By maintaining the micropore connectivity efficiency index at a high level, the system ensures that the mass transfer efficiency of the remediation agent within the micropores is maximized, thereby achieving precise removal of deep-seated contaminants.

[0034] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An artificial intelligence-based contaminated site remediation system, characterized in that, include: The flow field modulation injection subsystem is configured to inject a fluid medium into an underground formation and, in response to an input modulation control signal, drive an actuator to cause periodic fluctuations in the pressure or flow velocity of the fluid medium, thereby establishing a physical carrier flow field in the formation. A synchronous response monitoring subsystem is deployed in the downstream influence area of ​​the flow field modulation injection subsystem and is configured to continuously acquire the formation response signal after the physical carrier flow field is transmitted through the formation medium. The intelligent control terminal is communicatively connected to the flow field modulation injection subsystem and the synchronous response monitoring subsystem, respectively. The intelligent control terminal is configured to perform the following operations: A reference signal is generated, the frequency of which is synchronized with the fluctuation frequency of the physical carrier flow field; Phase-locked demodulation is performed on the formation response signal and the reference signal to extract the transmission time delay characteristic parameters of the formation response signal relative to the physical carrier flow field; The flow channel type of the current formation medium is identified based on the transmission time delay characteristic parameters, and feedback control commands are generated based on the identification results. The feedback control command is used to adjust the type of fluid medium injected by the flow field modulation injection subsystem.

2. The system according to claim 1, characterized in that, The intelligent control terminal is also configured to run a spectrum optimization module: Before starting the flow field modulation injection subsystem, background noise data of the synchronization response monitoring subsystem is collected, and the spectral entropy of the background noise data is calculated. The center frequency of the frequency band with the lowest spectral entropy is selected as the carrier frequency of the modulation control signal, and this carrier frequency is sent to the flow field modulation injection subsystem.

3. The system according to claim 1, characterized in that, The actuator of the flow field modulation injection subsystem is a hydraulic pulse generator; The hydraulic pulse generator is configured to generate a physical carrier flow field with a square wave or trapezoidal wave form having a steep rising edge at the injection end by periodically switching the on / off state of the fluid passage or periodically adjusting the output pressure of the fluid source.

4. The system according to claim 3, characterized in that, The hydraulic pulse generator specifically includes: A constant-pressure fluid source is used to provide stable fluid pressure. The three-way switching valve has an inlet, an injection port, and a return port; the inlet is connected to the constant pressure fluid source, the injection port is connected to the formation injection well, and the return port is connected to the return container. The intelligent control terminal is configured to drive the three-way switching valve to quickly switch between the injection port being open and the return port being open using the carrier frequency.

5. The system according to claim 1, characterized in that, The intelligent control terminal is configured to use the micropore connectivity efficiency index as the basis for drug injection control. The micropore connectivity efficiency index is constructed based on the two-dimensional joint distribution of the amplitude of the formation response signal and the transmission time delay characteristic parameters, and its mathematical model includes: An amplitude saturation factor is configured to exhibit a non-linear saturation trend as the amplitude increases, in order to suppress signal artifacts caused by high concentration accumulation. The phase matching factor is configured to exhibit bandpass filtering characteristics centered on a preset target permeation time delay value, in order to filter signals within the effective micropore mass transfer range; The micropore connectivity efficiency index is determined by the product of the amplitude saturation factor and the phase matching factor.

6. The system according to claim 5, characterized in that, The micropore connectivity efficiency index The specific calculation formula is as follows: ， in, For the amplitude, The phase angle is used as a characteristic parameter of transmission time delay. The target phase angle corresponding to the target penetration time delay value. This is the global gain coefficient. This is the sensitivity adjustment coefficient. This is the resonant bandwidth parameter.

7. The system according to claim 1, characterized in that, It also includes source-side noise reduction circuitry; The input terminal of the source-end noise reduction circuit is coupled to the drive motor current loop of the flow field modulation injection subsystem for acquiring the pump drive current waveform. The output of the source-end noise reduction circuit is coupled to the signal conditioning front end of the synchronous response monitoring subsystem; The source-end noise reduction circuit is configured to invert the pump drive current waveform and superimpose it onto the analog front end of the formation response signal to cancel out electromagnetic interference from the same source.

8. A method for remediating contaminated sites based on artificial intelligence, characterized in that, Includes the following steps: The flow field modulation injection subsystem is used to drive the injection equipment to generate physical pressure pulse waves with a specific carrier frequency, thereby actively modulating the underground flow field. The downstream stratum response signal is collected using the synchronous response monitoring subsystem, and a reference signal synchronized with the carrier frequency is generated using the intelligent control terminal. The formation response signal and the reference signal are subjected to phase-locked demodulation to extract transmission time delay characteristic parameters; The transmission time delay characteristic parameter is used as a physical fingerprint to distinguish the flow channel type: when the parameter indicates that the time delay is close to zero, it is determined to be a short-circuit flow; when the parameter indicates that the time delay is within a preset range, it is determined to be a seepage flow. Closed-loop control is implemented based on the determined flow channel type: if it is determined to be a short-circuit flow, the injection medium is switched to a sealing material; if it is determined to be a permeable flow, the injection medium is switched to a repair agent.

Images