Groundwater circulation well remediation system and method based on reduction-oxidation coupling process

Abstract

The application discloses a groundwater circulation well repair system and method based on a reduction-oxidation coupling process and belongs to the technical field of in-situ groundwater repair. The application aims to solve the problems of incomplete single process treatment, easy passivation of zero-valent iron material, low mass transfer efficiency of the oxidation zone and insufficient precision of reagent addition in the repair of existing composite contaminated groundwater. The system is sequentially arranged from bottom to top along the vertical circulation path of groundwater in the well pipe, including a lower screen hole, a reduction zone, a water collecting and guiding structure, an oxidation zone drop swirl hydraulic structure and an upper water outlet screen section. The reduction zone is filled with porous carriers to inhibit passivation, and the oxidation zone relies on the natural vertical drop and the tangential guiding structure to form a swirl to strengthen mass transfer. The oxidation-reduction potential monitoring points at the water inlet and outlet realize closed-loop intelligent addition of oxidants. The application can realize in-situ ordered coupling of the reduction and oxidation processes, improve the repair efficiency and operation stability of the composite contaminated groundwater and reduce the energy consumption of the system.

Description

Technical Field

[0001] This invention relates to the field of in-situ groundwater remediation technology, and in particular to a groundwater circulation well remediation system and method based on a reduction-oxidation coupling process. Background Technology

[0002] Currently, in-situ remediation technology has become the mainstream treatment method in engineering applications for groundwater contaminated by a combination of chlorinated hydrocarbons, nitroaromatics, and high-valence heavy metals. Among them, zero-valent iron reduction technology and advanced oxidation technology based on persulfate and hydrogen peroxide have been widely promoted due to their mild reaction conditions and wide applicability to pollutants. However, in practical applications, single reduction processes can only achieve the reduction and transformation of pollutants and are difficult to completely mineralize recalcitrant organic pollutants. In addition, sulfided zero-valent iron monomer materials have technical defects such as easy surface passivation and rapid decline in reactivity over time. Single oxidation processes have limited degradation efficiency for reducing pollutants such as chlorinated hydrocarbons and nitroaromatics. At the same time, they generally suffer from poor mass transfer mixing between oxidant and water and insufficient precision in reagent dosing. Meanwhile, existing reduction-oxidation coupled remediation processes mostly adopt horizontal zoning or multi-well phased implementation, which not only involves large construction area and high engineering costs, but also makes it difficult to achieve orderly coupling and linkage of the reduction and oxidation processes under in-situ groundwater conditions. In some coupled remediation processes using circulating well structures, the oxidation reaction zone needs to rely on additional aeration devices or stirring equipment to enhance water flow turbulence and mass transfer, resulting in high system energy consumption, difficult maintenance of downhole equipment, and a lack of a closed-loop dosing control mechanism for reagents based on real-time monitoring of oxidation-reduction potential. This makes it difficult to effectively adapt to the actual working conditions of dynamic fluctuations in groundwater pollution load, ultimately resulting in low remediation efficiency and insufficient operational stability of compound polluted groundwater, making it difficult to meet the engineering requirements for long-term remediation. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a groundwater circulation well repair system and method based on a reduction-oxidation coupling process.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of the present invention discloses a groundwater circulation well repair system based on a reduction-oxidation coupling process, including a well pipe body buried in a groundwater aquifer, a water pump that provides power for groundwater circulation, and an oxidant injection device for delivering oxidant into the well. The well casing is arranged from bottom to top along the vertical circulation path of groundwater, with a lower screen, a reduction zone, a water collection and guiding structure, an oxidation zone drop vortex hydraulic structure, and an upper outlet screen section, so that the contaminated groundwater forms a forced vertical circulation path under the drive of the water pump. The reduction zone is filled with a porous carrier-supported sulfidated zero-valent iron composite material. Screens are provided at the upper and lower ends of the reduction zone. The porous carrier disperses and anchors the sulfidated zero-valent iron particles to inhibit surface passivation of zero-valent iron and prolong the reaction activity cycle. The water collection and guiding structure includes a water collection section and a vertical downpipe. The water collection section is located in the cavity space above the porous carrier-supported sulfidated zero-valent iron composite material filler layer and is used to receive the water from the reduction zone. A natural vertical drop of 1.0m to 3.0m is formed between the outlet of the vertical downpipe and the inlet of the drop vortex hydraulic structure in the oxidation zone, which is used to convert gravitational potential energy into water kinetic energy. The oxidation zone drop vortex hydraulic structure includes a vortex chamber and a tangential flow guide structure. The water from the reduction zone, transported through the vertical downpipe, is injected into the vortex chamber along the tangential flow guide structure, spontaneously forming a columnar vortex to enhance the turbulent mixing of water flow and the mass transfer efficiency of pollutants. The oxidant injection point of the oxidant injection device is located near the inlet of the vortex chamber; the oxidation zone vortex hydraulic structure is equipped with an inlet oxidation-reduction potential monitoring point and an outlet oxidation-reduction potential monitoring point. Both monitoring points are connected to the controller. The controller is used to adjust the dosing parameters of the oxidant injection device according to the oxidation-reduction potential data of the two monitoring points in order to maintain a suitable oxidation environment in the oxidation zone.

[0005] Furthermore, the porous carrier is selected from one or more combinations of boron-doped ordered mesoporous carbon, cerium-based rare earth porous MOF, hierarchical porous biochar, porous ceramsite, modified zeolite, and sepiolite; the sieve is used to support the porous carrier to immobilize the sulfided zero-valent iron composite material, while allowing groundwater to pass through uniformly.

[0006] Furthermore, the inlet drop edge of the oxidation zone drop vortex hydraulic structure is provided with a sawtooth-shaped structure. When the water flows through the sawtooth-shaped structure, it is segmented and refined, further enhancing the water turbulence and mass transfer effect.

[0007] Furthermore, the oxidation-reduction potential monitoring point at the water inlet is located on the water inlet side of the inner wall of the vortex chamber, and the oxidation-reduction potential monitoring point at the water outlet is located at the center water outlet of the vortex chamber. The controller is configured to adjust the oxidant dosage based on the deviation between the oxidation-reduction potential value at the effluent point and the preset target value, and to identify pollution load impacts by combining the rate of change of the difference between the oxidation-reduction potentials at the influent point and the effluent point, and adjust the oxidant dosing pulse interval accordingly.

[0008] Further, the water pump is a low-head submersible circulating water pump, which is arranged at the water outlet side of the oxidation zone and is configured to pump the oxidation-treated groundwater out through the upper water outlet screen section and drain it to the aquifer; the hydraulic mixing in the oxidation zone is completely driven by the gravitational potential energy converted from the vertical drop, and no additional stirring or aeration equipment is required.

[0009] In a second aspect of the present invention, there is disclosed a control method for a groundwater circulation well remediation system based on a reduction-oxidation coupling process, which is applied to the groundwater circulation well remediation system according to any one of the preceding claims, and comprises the following steps: S1, Circulation start-up and hydraulic regulation: start the water pump to establish a vertical circulation flow of groundwater, adjust the operating frequency of the water pump, so that after the effluent from the reduction zone converges into the vertical downcomer through the water collection and diversion structure, a natural vertical drop of 1.0m to 3.0m is formed between the effluent and the water inlet of the drop swirling hydraulic structure in the oxidation zone, and at the same time, the hydraulic residence time in the reduction zone is controlled to match the reduction kinetic characteristics of the target pollutants; S2, Reduction process control: allow groundwater to flow through the sulfidated zero-valent iron composite material layer immobilized on the porous carrier in the reduction zone, so that the pollutants undergo continuous reduction transformation; regularly detect the concentration of target pollutants and intermediate products in the inlet and outlet water of the reduction zone, and when the reduction efficiency drops to a preset threshold, trigger a prompt for activation, regeneration or replacement of the composite material; S3, Gravity swirling oxidation start-up: the effluent from the reduction zone is injected into the swirling chamber through the tangential diversion structure via the vertical downcomer, and a columnar swirling flow is spontaneously formed by virtue of the vertical drop; calculate the current swirling intensity and theoretical mixing time according to the signal of the flowmeter matched with the water pump, and maintain the swirling Reynolds number in the strong turbulent flow range by finely adjusting the water pump flow; S4, Closed-loop control for intelligent dosing of oxidant: inject oxidant into the water inlet of the swirling chamber through the oxidant injection device via the oxidant injection point, collect the oxidation-reduction potential data of the water inlet point and the water outlet point in real time, adjust the oxidant dosage and dosing pulse interval in closed-loop based on the target value of the oxidation-reduction potential at the water outlet point, so as to maintain a suitable oxidation environment in the oxidation zone; S5, Cyclic iteration and remediation end-point determination: dynamically adjust the circulation flow rate and the set value of oxidant dosing according to the feedback of water quality indicators from the well and surrounding monitoring wells; after being sequentially coupled treated by reduction-oxidation process, the groundwater flows back to the aquifer and undergoes continuous cyclic iteration until the concentration of all target pollutants in the target aquifer is stably lower than the remediation standard and the effluent meets the standard for multiple consecutive cycles, then the control system shuts down.

[0010] Further, in step S3, calculating the current swirling intensity and theoretical mixing time according to the signal of the flowmeter matched with the water pump, and maintaining the swirling Reynolds number in the strong turbulent flow range by finely adjusting the water pump flow is specifically: The real-time flow time series output by the flow meter of the water pump is collected at a preset sampling frequency. The real-time flow time series is high-pass filtered by the built-in flow pulsation separation module to extract the high-frequency hydraulic pulsation component induced by the hydraulic instability of the columnar swirl in the swirl chamber. The phase space of the high-frequency hydraulic pulsation components is reconstructed, the delay time is determined by the mutual information method, and the embedding dimension is determined by the pseudo-nearest neighbor method. The attractor trajectory reflecting the evolution of the vortex structure of the swirling flow is reconstructed. The maximum Lyapunov exponent of the attractor trajectory is calculated and defined as the swirling flow chaos intensity used to quantify the degree of disorder in the complex motion of the swirling flow. The correlation dimension of the attractor trajectory is calculated synchronously, and the obtained correlation dimension is compared with the pre-stored correlation dimension benchmark value of the corresponding fully developed strong turbulent state. When the correlation dimension enters the neighborhood of the benchmark value, it is determined that the vortex has left the transition zone and reached the vortex series dissipation state required for strong turbulent mixing. A two-dimensional adjustment error function is constructed based on the difference between the correlation dimension and the benchmark value, and the deviation between the swirling chaos intensity and the preset chaos intensity target range. Based on the error function, a water pump frequency converter fine-tuning command is generated to slightly adjust the operating frequency to change the real-time flow rate until the correlation dimension converges to the benchmark value and the swirling chaos intensity stably falls within the target range. After the adjustment is stable, the average flow rate value of the current period is extracted, and the nominal turnover period is calculated in combination with the effective volume of the swirl chamber. The chaos-dissipation scaling law calibrated by the experiment and pre-stored is called to correct the nominal turnover period with the current swirl chaos intensity, so as to obtain the theoretical mixing time characterizing the time required to achieve the specified mixing uniformity in the whole chamber after the oxidant is injected.

[0011] Furthermore, in step S4, the specific logic of the closed-loop adjustment is as follows: When the oxidation-reduction potential value at the effluent point is lower than the preset target value, the dosage of oxidant should be increased. When the rate of change of the oxidation-reduction potential difference between the inlet and outlet points is greater than the preset rate of change threshold, and the difference between the oxidation-reduction potential value at the outlet point and the preset target value is greater than the preset deviation threshold, it is determined to be a high pollution load impact, and the dosage of oxidant is gradually increased and the dosing pulse interval is shortened. When the oxidation-reduction potential at the effluent point stabilizes within the preset target range and the concentration of pollutants in the effluent is lower than the preset limit, the current dosage is maintained to enter steady-state operation.

[0012] Furthermore, in step S2, the target pollutants include chlorinated hydrocarbons, nitroaromatic hydrocarbons, and high-valence heavy metals; the preset threshold is when the pollutant dechlorination efficiency is less than 80%, or the nitroaromatic hydrocarbon reduction conversion rate is less than a set limit.

[0013] Furthermore, the oxidant is persulfate or hydrogen peroxide, and the oxidant is activated in situ by ferrous iron released from zero-valent iron during the reduction reaction, generating strong oxidizing free radicals to deeply degrade pollutants.

[0014] This invention addresses the technical deficiencies in the prior art and has the following beneficial effects: By constructing vertically stratified reduction-oxidation two-stage reaction zones within groundwater circulation wells, the orderly coupling and linkage of reduction and oxidation processes under in-situ conditions can be achieved. This enables the synergistic treatment of groundwater contaminated with chlorinated hydrocarbons, nitroaromatics, and heavy metals, overcoming the shortcomings of single remediation technologies in achieving thorough treatment. The reduction zone utilizes a porous carrier to immobilize sulfidated zero-valent iron composite material, effectively inhibiting surface passivation of zero-valent iron and extending the material's reaction activity cycle. The oxidation zone leverages the natural vertical drop to create swirling hydraulic conditions that enhance mass transfer, eliminating the need for additional stirring or aeration equipment and reducing system operating energy consumption and downhole maintenance difficulty. Combined with a closed-loop oxidant dosing mechanism based on redox potential monitoring, precise and on-demand agent dosing can be achieved, adapting to dynamic fluctuations in pollution load and comprehensively improving the stability, treatment efficiency, and operational economy of in-situ groundwater remediation. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the overall system structure; Figure 2 This is a schematic diagram of the cross-sectional structure of the reduction zone; Figure 3 This is a schematic diagram of the plan and cross-section of the hydraulic structure of the drop-flow vortex in the oxidation zone; Figure 4 A magnified schematic diagram of the tangential flow guidance structure and drop tooth profile in the oxidation zone; The reference numerals in the attached drawings are explained as follows: 101, main body of the well casing; 102, water pump; 103, oxidant injection device; 104, lower sieve holes; 105, reduction zone; 107, vortex hydraulic structure of the oxidation zone; 109, porous carrier-supported sulfidated zero-valent iron composite material; 201, sieve; 202, vortex chamber; 203, tangential flow guide structure; 204, oxidation-reduction potential monitoring point at the water inlet; 205, oxidation-reduction potential monitoring point at the water outlet; 206, serrated tooth-shaped structure. Detailed Implementation

[0017] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0019] The instruction manual for the groundwater circulation well repair system 101 based on reduction-oxidation coupling technology described in this embodiment is attached. Figure 1 To be continued Figure 4 ,in Figure 1 This is a schematic diagram of the overall system structure. Figure 2 This is a schematic diagram of the cross-sectional structure of the reduction zone. Figure 3 This is a schematic diagram of the plan and cross-section of the vortex hydraulic structure in the oxidation zone. Figure 4 This is a magnified schematic diagram of the tangential flow-guiding structure and drop tooth profile in the oxidation zone.

[0020] like Figure 1 As shown, the remediation system includes a well casing 101 buried in the groundwater aquifer, a water pump 102 that provides power for groundwater circulation, and an oxidant injection device 103 for delivering oxidant into the well. The well casing 101 is a vertically arranged closed well casing structure, which can be made of high-density polyethylene (HDPE) or corrosion-resistant stainless steel. The diameter of the well casing can be set from 300mm to 800mm depending on the treatment scale of a single well. The depth of the well casing is determined according to the vertical thickness of the target pollution plume, usually from 10m to 30m. The top of the well casing is flush with the ground and is equipped with a sealed well cover to prevent external pollutants from entering the well.

[0021] The well body 101 is provided with a lower screen 104, a reduction zone 105, a water collection and guiding structure, an oxidation zone drop vortex hydraulic structure 107, and an upper effluent screen section arranged sequentially from bottom to top along the vertical circulation path of groundwater. Under the suction drive of the water pump 102, the contaminated groundwater forms a forced vertical circulation path between the well and the aquifer: "aquifer → lower screen → reduction zone → water collection and guiding structure → oxidation zone drop vortex hydraulic structure → upper effluent screen section → aquifer". A three-dimensional circulating flow field is formed around the well to continuously capture groundwater in the contamination plume for circulation treatment.

[0022] The lower screen holes 104 are located on the lower sidewall of the well casing 101, corresponding to the lower depth of the aquifer. The screen holes have a diameter of 1mm to 5mm and are evenly and staggered, with an opening rate controlled between 15% and 30%. This allows contaminated groundwater from the aquifer to enter the well evenly while preventing sand and suspended solids from entering the well casing and causing blockage of the packing material. The upper effluent screen section is located on the upper sidewall of the well casing 101, corresponding to the upper depth of the aquifer. Its pore size and opening rate are consistent with the lower screen holes 104. This section is used to return the treated groundwater to the aquifer, forming a complete cycle.

[0023] The water pump 102 is a low-lift submersible circulating water pump, installed on the outlet side of the oxidation zone's drop vortex hydraulic structure 107, i.e., at the upper part of the well pipe. The inlet end of the water pump 102 is connected to the outlet end of the oxidation zone's vortex chamber 202, and the outlet end is connected to the inner cavity of the upper outlet screen section. When the water pump 102 is running, it draws in the treated water in the oxidation zone and transports it upwards, discharging it through the upper outlet screen section to the aquifer. Simultaneously, it forms a hydraulic gradient from bottom to top within the well, driving the groundwater from the aquifer to continuously replenish the well through the lower screen holes 104. In an optional embodiment, the water pump 102 can also be installed on the ground, extending into the well's oxidation zone outlet side via a pumping pipeline, and its hydraulic drive function is the same as that of the submersible pump.

[0024] like Figure 2 As shown, the reduction zone 105 is located above the lower sieve holes 104 and is a columnar packing cavity inside the well pipe. The cavity height, i.e., the packing layer thickness, can be set from 1m to 5m, adjusted according to the concentration of the target pollutant and the kinetic requirements of the reduction reaction. Screens 201 are respectively installed at the upper and lower ends of the reduction zone 105. The screens 201 are made of 304 or 316L stainless steel, with sieve apertures of 0.5mm to 1mm, smaller than the particle size of the composite material particles filling the interior. The lower screen 201 supports the packing layer and allows the groundwater entering through the lower sieve holes 104 to be evenly distributed and flow upwards through the packing layer. The upper screen 201 prevents the packing particles from floating upwards with the water flow into the water collection and guiding structure, while allowing the treated effluent to pass through evenly.

[0025] The reduction zone 105 is filled with a porous carrier-supported zero-valent iron sulfide composite material 109. The composite material 109 uses a porous carrier as a skeleton to highly disperse and anchor the zero-valent iron sulfide particles in the pores and surface of the carrier. The porous carrier can be selected from one or more combinations of boron-doped ordered mesoporous carbon (B-OMC), cerium-based rare earth porous MOF, hierarchical porous biochar, porous ceramsite, modified zeolite, and sepiolite. The particle size of the carrier particles is usually 2 mm to 10 mm, with a well-developed hierarchical pore structure and a high specific surface area.

[0026] In a specific embodiment, when using hierarchical porous biochar as a carrier, agricultural waste straw can be selected and prepared by oxygen-limited pyrolysis at 500℃ to 700℃, followed by acid-base activation to create pores, resulting in a hierarchical porous biochar carrier with a specific surface area of ​​500m² / g to 1200m² / g. Subsequently, nano-zero valent iron precursors are loaded into the pores of the carrier by liquid-phase impregnation, followed by reduction with sodium borohydride and modification with sodium sulfide to obtain a fixed-load zero valent iron sulfide composite material 109, wherein the loading of zero valent iron sulfide is 10wt% to 30wt%, and the degree of sulfidation, i.e., the molar ratio of FeSx to FeO, is controlled at 0.05 to 0.2. The porous carrier's framework structure can separate and fix the nano-zero-valent iron particles, preventing particle agglomeration and deactivation, while providing adsorption and enrichment sites for pollutants. The sulfide shell formed by sulfidation modification can regulate the electron transfer path of zero-valent iron and inhibit the formation of a dense iron oxide passivation layer on the surface, thereby effectively extending the reaction activity cycle of the material and solving the defects of easy passivation and rapid activity decay of sulfide zero-valent iron monomers.

[0027] When boron-doped ordered mesoporous carbon is used as a support, its ordered mesoporous channels can achieve high dispersion and confined growth of zero-valent iron, making the material more resistant to passivation and suitable for high-concentration, long-term operation repair scenarios. When cerium-based rare earth porous MOF is used as a support, the valence state cycle of rare earth cerium can synergistically enhance electron transfer efficiency with zero-valent iron, further improving the reduction reaction rate.

[0028] When groundwater flows through the 105 packing layer in the reduction zone, chlorinated hydrocarbon pollutants in the water undergo reduction and dechlorination reactions, nitroaromatic pollutants undergo nitro reduction to transform into amine intermediates, and high-valence heavy metals such as Cr(VI) and As(V) are reduced to low-valence states and precipitate for removal. Simultaneously, the molecular structures of some recalcitrant organic pollutants are reduced and destroyed, transforming them into less toxic and more easily oxidized intermediates, creating conditions for subsequent deep oxidation treatment. During the reaction, zero-valent iron is oxidized, continuously releasing divalent iron ions, which enter the subsequent oxidation zone with the effluent and act as a natural activator for the oxidant, eliminating the need for additional activating agents.

[0029] The water collection and guiding structure is located above the reduction zone 105 and includes a water collection section and a vertical downpipe. It is used to collect the water effluent from the reduction zone 105 and guide it to the drop vortex hydraulic structure 107 in the oxidation zone, while forming a natural vertical drop to provide hydraulic mixing energy.

[0030] The water collection section is either a water collection trough arranged in a ring along the inner wall of the well pipe, or a water collection funnel located at the center of the well pipe; in this embodiment, a ring-shaped water collection trough is used, with the opening of the trough facing downwards toward the reduction zone 105. Water flowing upwards from the reduction zone 105 enters the ring-shaped water collection trough and is collected. The bottom of the trough is connected to the upper end of the vertical downpipe. The cross-sectional area of ​​the ring-shaped water collection trough is determined according to the designed circulation flow rate to ensure smooth water collection without backlog.

[0031] The vertical downpipe is a vertically installed closed pipe extending downwards along the central axis of the well pipe. Its lower outlet is connected to the inlet of the tangential flow guide structure 203 of the oxidation zone's drop vortex hydraulic structure 107. A natural vertical drop of 1.0m to 3.0m is formed between the stable water level at the outlet of the reduction zone 105 and the inlet of the oxidation zone's vortex chamber 202. This drop can be set by adjusting the length of the vertical downpipe and the installation height of the vortex chamber 202. In actual engineering, a drop of 2.0m is preferred, which ensures sufficient gravitational potential energy conversion efficiency while avoiding excessive vertical dimensions of the well pipe that would increase construction costs.

[0032] After the water from the reduction zone 105 is collected by the water collection section, it enters the vertical downpipe. Under the action of gravity, it accelerates downward along the pipe, converting the gravitational potential energy into the water kinetic energy, and is injected into the tangential flow guide structure 203 of the oxidation zone at a high flow rate. This process does not require additional power equipment and can provide energy for the hydraulic mixing of the oxidation zone by relying solely on the natural water level difference, effectively reducing the system's operating energy consumption.

[0033] like Figure 3 , Figure 4 As shown, the oxidation zone drop vortex hydraulic structure 107 includes a vortex chamber 202 and a tangential flow guide structure 203, which are set in the space below the water collection and flow guide structure and above the reduction zone 105, and realize the oxidation reaction function by utilizing the vertical space inside the well pipe.

[0034] The vortex chamber 202 is a vertically arranged cylindrical sealed cavity with a diameter smaller than the inner diameter of the well pipe body 101. An annular water flow channel is formed between the side wall of the cavity and the inner wall of the well pipe. The effective volume of the vortex chamber 202 is determined according to the design circulation flow rate and the hydraulic residence time required for the oxidation reaction. Typically, the hydraulic residence time in the oxidation zone is controlled between 2 minutes and 10 minutes.

[0035] The tangential flow-guiding structure 203 is located on the upper side wall of the vortex chamber 202. Its inlet end is connected to the outlet of the vertical downpipe, and the outlet end extends into the cavity along the tangential direction of the vortex chamber 202. The high-speed water flow accelerated by the vertical downpipe is injected tangentially into the cavity of the vortex chamber 202, flows along the inner wall of the cavity, and spontaneously forms a stable columnar spiral flow, generating a strong centrifugal force field and shear force field within the cavity. The intense turbulence of the spiral flow can break the boundary layer and concentration stratification of the water body, significantly enhance the mass transfer rate and contact probability between the oxidant and pollutants, and improve the oxidation reaction efficiency.

[0036] Furthermore, the inlet drop edge of the vortex chamber 202 is provided with a serrated tooth structure 206. This serrated tooth structure 206 is located below the outlet end of the tangential flow guide structure 203, at the edge of the weir where the water falls into the vortex chamber 202, and is arranged in a continuous triangular serrated pattern. The height of a single tooth is 5–20 mm, the tooth spacing (distance between adjacent tooth tips) is 10–30 mm, and the tooth angle is 60°–90°, which can be selected according to the design flow rate and drop height. When the water falls through the weir, it is divided into multiple streams by the serrated tooth structure 206, increasing the shearing effect and turbulence intensity of the water flow, further breaking up water clumps, and enhancing the initial mixing effect of the oxidant and water.

[0037] During operation, the hydraulic mixing in the oxidation zone is driven entirely by the gravitational potential energy converted from the vertical drop, eliminating the need for additional mechanical stirring or aeration equipment. This reduces the number of moving parts in the well, lowers the difficulty of equipment maintenance and system energy consumption, and avoids the risk of volatile pollutants escaping from aeration.

[0038] The oxidant injection device 103 includes an oxidant storage tank, a metering pump, a dosing pipeline, and an oxidant injection point. The oxidant injection point is located near the inlet of the vortex chamber 202, specifically inside the outlet of the tangential flow guide structure 203, allowing the oxidant to enter the vortex chamber 202 synchronously with the high-speed incoming water, and to mix rapidly and uniformly under the action of swirling turbulence. The oxidant can be a persulfate such as sodium persulfate, potassium persulfate, hydrogen peroxide, etc. In this embodiment, sodium persulfate is selected as the oxidant. Under the activation of ferrous ions carried by the effluent in the reduction zone 105, it generates sulfate radicals and hydroxyl radicals, which can deeply degrade organic pollutants.

[0039] The oxidation zone vortex hydraulic structure 107 is equipped with an inlet oxidation-reduction potential (ORP) monitoring point 204 and an outlet ORP monitoring point 205. Both monitoring points are equipped with ORP probes, which are connected to a ground controller via signal lines. The inlet ORP monitoring point 204 is located on the inlet side of the inner wall of the vortex chamber 202, near the outlet of the tangential flow guide structure 203, and is used to monitor the ORP after the initial mixing of the inlet water and oxidant. The outlet ORP monitoring point 205 is located at the center outlet of the vortex chamber 202, specifically at the outlet at the top center of the vortex chamber 202, and is used to monitor the ORP of the outlet water after complete reaction.

[0040] The controller is simultaneously connected to the frequency converter signals of the oxidant metering pump and the water pump 102. It can collect real-time data from two ORP monitoring points and adjust the oxidant dosage, dosage pulse interval, and water pump 102 operating frequency according to preset control logic to maintain a suitable oxidation environment in the oxidation zone, achieve precise on-demand oxidant dosing, avoid waste and secondary pollution caused by excessive dosing of reagents, and adapt to the working conditions of fluctuating pollution load.

[0041] The groundwater circulation well remediation method based on reduction-oxidation coupling process described in this invention, applied to the aforementioned remediation system, specifically includes the following steps: S1, Circulation Start-up and Hydraulic Regulation When the system starts, the water pump is started first, and the pump operating frequency is gradually increased to establish a stable vertical circulation flow rate of groundwater in the well. During the adjustment process, the operating frequency of the water pump is adjusted according to the permeability coefficient of the target aquifer and the planar distribution range of the pollution plume to maintain an appropriate circulation radius and circulation cycle; for example, when the aquifer permeability coefficient is 1 m / d to 10 m / d, the circulation radius is controlled to be 5 m to 15 m, and the circulation cycle of a single well is 1 day to 3 days to ensure that the pollution plume around the well can be effectively captured and enter the circulation system.

[0042] Simultaneously, the hydraulic residence time in the reduction zone is controlled by adjusting the water pump flow rate to match the reduction kinetics of the target pollutant. For chlorinated hydrocarbons and nitroaromatics, the hydraulic residence time in the reduction zone is typically controlled between 5 and 30 minutes to ensure that the pollutants undergo sufficient reduction and transformation within the zone. After hydraulic adjustment, the effluent from the reduction zone flows through a water collection and guiding structure into a vertical downpipe, creating a stable natural vertical drop of 1.0 to 3.0 meters between the effluent and the inlet of the vortex chamber in the oxidation zone, providing stable dynamic conditions for subsequent gravity vortex oxidation.

[0043] S2, Reduction Process Control After the system is running stably, the contaminated groundwater continues to enter the well through the lower screen holes and flows upward through the porous carrier-supported zero-valent iron sulfide composite material layer in the reduction zone. The pollutants are in full contact with the material surface and a reduction and transformation reaction continues to occur.

[0044] During operation, the system regularly samples and detects the inlet and outlet water of the reduction zone, the detection period can be set from 3d to 7d, and the detection indicators include the concentration of target reduction pollutants such as chlorinated hydrocarbons, p-chloronitrobenzene, Cr(VI), the concentration of characteristic intermediate products and the concentration of ferrous ions. The reduction efficiency of pollutants, such as dechlorination efficiency and nitro reduction conversion rate, is calculated. A preset reduction efficiency threshold is used as the criterion for material activity. For example, for chlorinated hydrocarbon pollutants, the preset dechlorination efficiency threshold is 80%; when the dechlorination efficiency of two consecutive detections is lower than 80%, it is determined that the material activity has decreased, and the system triggers a prompt signal to remind operation and maintenance personnel to activate, regenerate or replace the porous carrier-supported sulfidated zero-valent iron composite material. Among them, in-situ dilute hydrochloric acid flushing can be used for activation regeneration to remove the passivation layer on the material surface and restore the reaction activity; when the material activity cannot be restored through regeneration, the filler is replaced.

[0045] S3. Start-up of gravity swirling oxidation After being collected by the water collecting part, the effluent from the reduction zone enters the vertical downcomer, accelerates flowing relying on the vertical drop of 1.0m to 3.0m, is injected into the swirling chamber at a high speed along the tangential guide structure, spontaneously forms a columnar spiral flow in the chamber, and enters the gravity swirling oxidation state.

[0046] During operation, the controller collects real-time circulating flow signals through the electromagnetic flowmeter matched with the water pump, and automatically calculates the current swirling intensity and theoretical mixing time according to the flow data and the structural parameters of the swirling chamber. Specifically, the swirling Reynolds number is calculated according to the hydraulic diameter of the swirling chamber and the water flow velocity, and the swirling Reynolds number is compared with the judgment threshold of the strong turbulent zone, wherein the Reynolds number threshold of the strong turbulent zone can be set to above 4000; when the flow fluctuation causes the swirling Reynolds number to be lower than the threshold and the mixing intensity is insufficient, the controller sends a fine adjustment instruction to the water pump frequency converter, slightly increases or decreases the operating frequency of the water pump, adjusts the circulating flow, ensures that the swirling Reynolds number is always maintained in the strong turbulent interval, and guarantees the mixing and mass transfer effect in the oxidation zone.

[0047] In an optimized embodiment, the present invention adopts an intelligent diagnosis method based on flow fluctuation chaotic feature analysis to replace the traditional single Reynolds number criterion, and realizes more accurate and reliable online judgment and control of the swirling mixing state. The principle is that the swirling flow in the strong turbulent state is not completely random, but presents deterministic chaotic characteristics, which are hidden in the high-frequency and weak fluctuations of the flow. Through chaotic time series analysis technology, a "fingerprint" that can uniquely characterize the cascade dissipation state of turbulent vortices can be extracted, so as to realize accurate identification and feedback control of strong turbulent mixing. The specific implementation steps are as follows: The real-time flow time series Q(t) output by the electromagnetic flowmeter matched with the water pump is collected at a preset sampling frequency (for example, 10 Hz to 100 Hz, which needs to be higher than 5 times the fundamental frequency of the swirling flow). Due to the stable operation of the water pump and the relative stability of the main swirling flow, the original flow signal appears as a superposition of a slowly varying DC component and a weak fluctuation component. The information concerned by the present invention is contained in the fluctuation component. Therefore, through the built-in flow pulsation separation module, high-pass filtering is performed on Q(t) to filter out the low-frequency components representing the change of average flow, and accurately extract the high-frequency hydraulic pulsation component induced by the hydraulic instability of the columnar swirling flow in the swirling chamber (such as vortex core precession and secondary flow oscillation), which is recorded as q'(t). The cut-off frequency of the high-pass filtering is dynamically set according to the fundamental frequency of the swirling flow, and the fundamental frequency of the swirling flow f base can be obtained by the formula f base ≈ 4·Q avg / (π· D 2 · h) for estimation, where Q avg is the short-term average flow, D is the diameter of the swirling chamber, and h is the effective water depth of the swirling chamber. The cut-off frequency is set as 0.8·f base to ensure that all pulsation information related to the swirling flow structure is retained, while low-frequency interferences such as pump speed fluctuation are eliminated.

[0048] The chaotic characteristic of turbulence is reflected in the geometric structure in the phase space of the attractor trajectory formed by the time evolution of its state variable (referring to the high-frequency hydraulic pulsation component q'(t)). To reveal this structure, it is necessary to reconstruct a high-dimensional phase space from the single time series q'(t). The recognized delayed coordinate embedding method is used for phase space reconstruction: first, the mutual information method is used to determine the optimal delay time τ, that is, the mutual information function between q'(t) and q'(t+τ) is calculated, and τ corresponding to the first minimum point when the function curve drops to the minimum value is taken; second, the false nearest neighbor method is used to determine the minimum embedding dimension m, that is, starting from a smaller dimension, the dimension is continuously increased. When increasing the dimension no longer causes a significant reduction in the number of "false nearest neighbor points" (that is, points that are close in projection in low-dimensional space but separated in high-dimensional space), this dimension is the minimum m required for reconstruction. According to the determined τ and m, the one-dimensional pulsation component time series q'(t) can be reconstructed into a series of state points in m-dimensional space, and the connection of these points constitutes the attractor trajectory reflecting the evolution of the turbulent vortex structure of the swirling flow.

[0049] On the reconstructed attractor trajectory, two key invariants are calculated to quantify the chaotic characteristics of the current swirling flow: 1. Swirling Chaos Intensity (Maximum Lyapunov Exponent λ1): λ1 > 0 is a decisive indicator of the chaotic characteristics of a dynamical system. Its magnitude characterizes the rate of divergence from nearby orbits, i.e., the strength of the system's sensitivity to initial conditions. This invention uses the Rosenstein small data method to calculate λ1 and defines it as the "swirling chaos intensity" to quantitatively evaluate the complexity and disorder level of swirling motion.

[0050] 2. Attractor Correlation Dimension (D²): D² is another important indicator of the geometric complexity of the attractor. This invention uses the Grassberger-Procaccia algorithm to calculate D². When the flow transitions from laminar to fully developed strong turbulence, the attractor correlation dimension increases from a relatively low value and eventually tends towards a saturation value. This saturation value corresponds to the "vortex train dissipation state" required for strong turbulent mixing, and is called the baseline value D of the correlation dimension. 2base .

[0051] The correlation dimension benchmark value D 2base The method for obtaining the flow rate is crucial, as it forms the cornerstone of this method. It needs to be pre-stored in the controller through laboratory calibration. The calibration process is as follows: Construct a transparent physical model that perfectly matches the geometry and structure of the actual circulating well's oxidation zone cyclone chamber. Configure a high-precision variable frequency pump and flow meter upstream of its inlet. Start the system and adjust the circulation flow rate to an extreme value far exceeding the maximum design flow rate, ensuring the cyclone Reynolds number (Re) is clearly greater than 10000. At this point, it can be physically confirmed that the flow state has reached fully developed strong turbulence. Run the system stably under this condition for at least 10 minutes. Collect flow pulsation signals and reconstruct the attractor using the same method as in steps one and two above. Calculate its correlation dimension using the Grassberger-Procaccia algorithm. Repeat the sampling calculation at least five times and take the statistical average. This average value is used as the baseline value D for the correlation dimension of this specific cyclone chamber structure. 2base The data is then embedded into the controller program. Similarly, by changing the flow conditions, a table of correlation dimension benchmark values ​​or fitted curves under different flow rates can be pre-stored in the controller.

[0052] During online operation, the controller compares the real-time calculated D2 with the pre-stored D2_base. When the correlation dimension D2 enters the baseline value D... 2base The neighborhood (e.g., in D) 2base When the vortex flow is within ±5% of the transition zone, it can be determined that the vortex has left the transition zone and reached the vortex series dissipation state required for strong turbulent mixing.

[0053] To accurately adjust and maintain the swirling state within the optimal mixing range, this invention constructs a two-dimensional adjustment error function E: E(ΔD,Δλ) = α·|ΔD| + β·|Δλ| Where ΔD is the real-time correlation dimension D2 and the baseline value D 2base The difference is Δλ, which is the deviation between the real-time swirling chaos intensity λ1 and the preset chaos intensity target range (the λ1 interval corresponding to the optimal mixing found through experiments); α and β are weighting coefficients. Based on this error function E, the controller generates fine-tuning instructions for the water pump frequency converter through PID or fuzzy logic algorithms, slightly adjusting the operating frequency to change the real-time flow rate, realizing a closed-loop optimization process until D2 converges to the reference value and λ1 stably falls within the target range. At this point, the swirling mixing state is reached and maintained at its optimum.

[0054] After the flow rate stabilizes, extract the average flow rate Q for the current period. stable Combined with the known effective volume V of the vortex chamber eff Calculate the nominal turnover period T nom = V eff / Q stable To accurately predict the time required for the oxidant to achieve a specified level of homogeneity (e.g., 95%) throughout the chamber, T cannot be used directly. nom However, it is necessary to introduce a correction to the mixing efficiency based on the current chaotic state. This invention corrects this through a pre-stored "chaos-dissipation scaling law," which is pre-stored through the following tracer impulse response experiment: In the calibration experiment described in step three, under a series of stable operating conditions with different chaos intensities λ1, a certain concentration of inert tracer (such as NaCl solution) is instantaneously injected through the inlet. At the outlet, an online conductivity meter is used to record the concentration change curve over time using high-frequency sampling. The actual mixing time T required for the tracer concentration to reach 95% of the equilibrium concentration is recorded. actual Calculate the mixing time correction factor C(λ1) = T actual / T nom By plotting scatter plots of C(λ1) under different λ1 values ​​and performing curve fitting, the empirical formula for the scaling law of the vortex chamber, C = f(λ1), can be obtained. During online operation, the controller calls this formula to calculate the correction factor based on the current stable vortex chaos intensity λ1, ultimately obtaining the theoretical mixing time T, which characterizes the time required to achieve a specified mixing uniformity throughout the chamber after oxidant injection. theory = C(λ1) · T nom This time value can provide a direct basis for the intelligent control of the subsequent oxidant dosing pulse timing.

[0055] S4, Intelligent closed-loop control of oxidant dosing The oxidant is injected into the cyclone chamber inlet via a metering pump, entering synchronously with the incoming water for mixing and reaction. The controller uses real-time oxidation-reduction potential data from the inlet and outlet points as input to execute closed-loop dosing control. The specific logic is as follows: Based on the type of pollutant and the remediation target, preset target values ​​and preset target ranges for the oxidation-reduction potential (ORP) at the effluent point are set. For example, for chlorinated hydrocarbon complex pollution, the preset target value for effluent ORP is 400mV, and the target range is 350mV to 450mV. At the same time, preset change rate thresholds and preset deviation thresholds are set. For example, the difference change rate threshold is set to 50mV / min, and the deviation threshold is set to 100mV.

[0056] When the oxidation-reduction potential value at the water outlet is lower than the preset target value, the controller outputs a signal to increase the amount of oxidant added and enhance the oxidation reaction intensity.

[0057] When the rate of change of the oxidation-reduction potential difference between the inlet and outlet points exceeds the preset rate of change threshold, and the difference between the oxidation-reduction potential value at the outlet point and the preset target value exceeds the preset deviation threshold, the system is determined to be experiencing a high pollution load impact. At this time, the controller gradually increases the oxidant dosage and shortens the dosage pulse interval, for example, shortening the normal dosage pulse interval from 10 to 30 minutes to 2 to 5 minutes, to quickly replenish the oxidation capacity to cope with the load impact.

[0058] When the oxidation-reduction potential at the effluent point stabilizes within the preset target range and the concentration of pollutants in the effluent remains below the preset limit, the oxidation reaction is determined to have entered a steady state. The controller maintains the current oxidant dosage and dosing interval, and enters a steady-state operation mode, thus saving reagent usage while ensuring treatment effectiveness.

[0059] Through the above closed-loop control, the oxidation reaction zone can always maintain a suitable oxidation environment, realize intelligent and precise dosing of oxidant, and adapt to the dynamic fluctuations of groundwater pollution load.

[0060] S5. Iteration and Endpoint Determination for Repair During system operation, water quality indicators of the effluent from the well and groundwater from surrounding monitoring wells are collected periodically. Based on the feedback from these indicators, operating parameters such as circulation flow rate and target oxidant dosage are dynamically adjusted to optimize the remediation effect. Groundwater treated by a combination of reduction and oxidation is returned to the aquifer through the upper effluent screen, forming a continuous circulating flow field around the well. This continuously captures groundwater from the pollution plume and recirculates it into the well for further treatment. In the complex pollution, reduction-resistant components are deeply degraded in the oxidation zone, intermediate products generated during the reduction process are further thoroughly mineralized, and heavy metals are stabilized and removed, achieving synergistic treatment of complex pollution.

[0061] When the concentration of all target pollutants in the target aquifer is consistently below the remediation standard limit, and the pollutant concentration meets the standard without rebound in three or more consecutive circulating water tests, the remediation is deemed to have reached its endpoint. The control system then gradually reduces the pump frequency and oxidant dosage, and finally shuts down to end the remediation.

[0062] In a preferred embodiment, it further includes: When the concentration of the target pollutant in the effluent rises abnormally, the multivariate time series of pollutant concentration and circulation flow rate of the influent and effluent of the reduction zone and the influent and effluent of the oxidation zone are extracted based on the time span of the first decay of the average mutual information between each variable to a local minimum value, and the causal lag time window of each variable is determined. Within each window, phase space reconstruction is performed on each variable sequence to obtain time-delay embedding vectors, which are then arranged in parallel to form a multivariate joint phase space. Each time point in the phase space is used as a node, and directional connections are defined by the neighborhood cross relationship between nodes to construct a multivariate nonlinear recursive network. For any two variable nodes in a nonlinear recursive network, the values ​​of the driving variables are estimated by the cross-mapping of the neighborhood mapping manifold of the driven variables. The predicted convergence coefficient of the convergent cross-mapping is calculated. If the predicted convergence coefficient shows an increasing convergence trend in the time domain and is higher than the causal convergence threshold, it is determined that there is a directional causal driving force. Using the judgment result as the basis for weighting directed edges, retaining the edges that pass the threshold test, and constructing an instantaneous causal propagation topology graph; The effluent concentration node in the oxidation zone is set as the causal sink. All complete causal paths are searched in reverse. The logarithmic ratio of the causal convergence coefficients of the edges traversed by each path is calculated as the cumulative amount of causal flow spectrum gradient. Based on the functional attributes of the nodes along the path, the cumulative amount is clustered and decomposed into reduction phase causal flow spectrum, oxidation phase causal flow spectrum and hydraulic phase causal flow spectrum. The proportion of each phase causal flow spectrum is compared with the pre-stored perturbation propagation counterfactual causal state feature code. The feature code includes the following: under the reduction hysteresis state, the causal flow spectrum of the reduction phase is concentrated at a low gradient while the causal flow spectrum of the oxidation phase is discrete; under the oxidation shock state, the causal flow spectrum of the oxidation phase is a single-point peak while the causal flow spectrum of the hydraulic phase is stable; under the hydraulic short-circuit state, the causal flow spectrum of the hydraulic phase is dominant and the causal flow spectra of the other two phases are both attenuated and divergent. Based on the signature matching results, specific diagnostic information is generated pointing to the control of reduction zone, oxidant addition, or hydraulic circulation.

[0063] It should be noted that the controller continuously collects target pollutant concentration data in the effluent of the oxidation zone online, and the sampling interval is synchronized with the analysis cycle of the online monitoring instrument. This embodiment uses a composite judgment condition to define when an abnormal rise warrants intervention, balancing detection sensitivity and anti-interference capability. The first judgment condition is a combined trend and outlier criterion: within the most recent consecutive sampling periods, the concentration of the target pollutant in the effluent shows a monotonically increasing trend, and the latest measured value exceeds a specified multiple standard deviation of the historical average value during the steady-state period. Here, the historical average value during the steady-state period is the moving average of this indicator during the stable operation period of the system before the anomaly occurred, and the standard deviation is also based on statistics from that period. The multiple can be set to 3 times to constitute a statistically significant outlier. The second judgment condition is a combined criterion of exceeding the standard and oxidation-reduction potential: the latest measured value exceeds the remediation target limit by a certain proportion, and the continuous increasing trend of the difference between the oxidation-reduction potential of the influent and effluent points exceeds a preset rate of change threshold within a specified period before and after this point.

[0064] Once any of the above conditions are met, it is determined that the concentration of the target pollutant in the effluent has abnormally increased, and the system automatically triggers the subsequent causal diagnosis procedure. Through this composite logic, transient fluctuations caused by random noise in the analytical instrument will not falsely trigger the diagnostic process, while the actual water quality deterioration caused by process factors can be captured in a timely manner.

[0065] Once the abnormal rebound is confirmed, high-frequency records within a time window containing sufficient data points are automatically extracted from the time of the anomaly to form a multivariate time series dataset. To determine the causal lag of disturbance propagation between variables, the average mutual information function is calculated for every two variables. During the calculation, one variable is fixed at the current time, while different time delays are gradually introduced into the other variable. The delay time corresponding to the first decay of the mutual information value to a local minimum as the delay increases is taken as the causal lag time window between the two. Taking the effluent concentration in the reduction zone and the influent concentration in the oxidation zone as an example, the local minimum delay between the two is the typical lag of hydraulic transmission in the reduction zone.

[0066] After determining the causal lag windows between variables, phase space reconstruction is performed on each univariate time series: that is, the optimal delay time is obtained independently using the mutual information method for each variable, and the minimum embedding dimension is determined using the pseudo-nearest neighbor method, thus expanding the original one-dimensional time series into a high-dimensional time-delay embedding vector. After the reconstruction of all variables is completed, these time-delay embedding vectors are concatenated to form a unified multivariate joint phase space. In this space, each time sampling point is mapped to a high-dimensional state point, and these state points naturally become nodes of the recursive network. The establishment of directed edges in the network depends on the neighborhood cross-relationship between nodes. Specifically, considering the nearest neighbor sets of any two nodes in the space, if the majority neighbors of one node are systematically lagging behind the corresponding neighbors of another node in terms of time stamp, it is inferred that there is a directed influence channel from the latter to the former, thus establishing a directed edge. By traversing all node pairs according to this rule, a complete multivariate nonlinear recursive network is constructed. This network topologically encodes the dynamic coupling and information flow between various process links within the repair system.

[0067] While recursive networks reveal the potential causal framework between variables, the strength of connections and the confidence in the driving direction still require further quantification. Therefore, this embodiment introduces a convergent cross-mapping technique: if one variable has a genuine causal drive on another, the dynamic information of the driving variable must be recorded in the time-series evolution of the driven variable. Thus, the value of the driving variable can be reconstructed from the phase space manifold of the driven variable. In practice, for a pair of variables X and Y to be analyzed, attractor manifolds are constructed using their time-delayed embedding vectors in the joint phase space. Several nearest neighbors of the current state point are located on the manifold, and the cross-mapping estimate is obtained by taking a distance-weighted average of the states of the other variable corresponding to these neighbors. The calculation of the predicted convergence coefficient is performed incrementally, starting with a shorter sequence length and repeatedly executing the above cross-mapping estimation, gradually increasing the length of the data used to obtain a series of correlation coefficient values ​​related to the length, i.e., the predicted convergence coefficient. If the coefficient shows a significant increase and eventually converges to a relatively stable value as the sequence length increases, it indicates that the predictive power of the cross-mapping comes from deterministic causal driving rather than accidental correlation. When the convergence stability value exceeds the preset causal convergence threshold, the directional causal driving in the corresponding direction is determined to be valid.

[0068] After obtaining all pairwise causal relationships that pass the threshold test, an instantaneous causal propagation topology graph is constructed using the natural logarithm of the predicted convergence stable value corresponding to each relationship as the edge weight. In the graph, the effluent concentration node in the oxidation zone is designated as the unique causal sink. Based on this, a reverse path search is performed starting from the causal sink to identify all complete directed paths extending from the upstream root node to the effluent concentration sink node. For each path, the cumulative value of the weights of each connecting edge on the path is the cumulative causal flow spectrum gradient of that path, which reflects the overall driving force of disturbance propagation on that causal chain. According to the functional affiliation of the nodes passed through by the path, the cumulative values ​​of all paths are classified and summarized: those passing through the influent or effluent nodes in the reduction zone are classified into the reduction phase causal flow spectrum, those passing through the influent or effluent nodes in the oxidation zone are classified into the oxidation phase causal flow spectrum, and those passing through the flow rate nodes are classified into the hydraulic phase causal flow spectrum. Thus, the causal flow spectrum of the system is decomposed into three-phase components, and the proportion and distribution of each phase component constitute a unique causal fingerprint under the current fault state. To accurately interpret this fingerprint, the controller pre-stores disturbance propagation counterfactual causal state feature codes calibrated through historical fault simulation or simulation experiments. These feature codes describe the causal flow spectrum patterns of several typical process failures in a patterned language: the reduction hysteresis state is characterized by a low gradient concentration in the causal flow spectrum of the reduction phase and a relatively discrete causal flow spectrum of the oxidation phase; the oxidation shock state is characterized by a single-point spike in the causal flow spectrum of the oxidation phase while the causal flow spectrum of the hydraulic phase remains stable; and the hydraulic short-circuit state is characterized by the hydraulic phase causal flow spectrum dominating the overall spectrum, and both the reduction and oxidation phase causal flow spectra exhibiting decaying divergence.

[0069] During diagnosis, the real-time decomposed three-phase causal flow spectrum proportion characteristics are compared with each feature code for similarity. Pattern matching methods such as normalized cross-correlation or eigenvector distance can be used. Once a match is found with a feature code, the controller outputs the corresponding diagnostic information and control instructions. Reduction hysteresis corresponds to activity decay in the reduction zone, suggesting triggering material activation or replacement; oxidation shock corresponds to insufficient instantaneous oxidant dosage or uneven mixing, suggesting increasing the oxidant dosage or shortening the dosage pulse interval; hydraulic short circuit corresponds to the presence of a priority flow channel or packing blockage in the circulation well, suggesting adjusting the pump frequency or checking the integrity of the packing layer. The above diagnostic information can be simultaneously pushed to the host computer human-machine interface to guide maintenance personnel in precise handling. It can also be linked with the existing functions of the controller to automatically adjust the pump frequency and oxidant dosage parameters, completing a complete intelligent operation and maintenance closed loop from anomaly detection to cause location and autonomous control.

[0070] In a preferred embodiment, it further includes: Time-series monitoring data of the active decay process of porous carriers under various site conditions were collected, including multi-channel sequences of reduction efficiency, inlet and outlet water pressure difference and iron ion concentration dissolved from the packing material, and the corresponding decay stage labels were labeled to form a source domain dataset. A one-dimensional deep convolutional autoencoder is constructed, and unsupervised pre-training is performed on the source domain dataset with the goal of minimizing the reconstruction error of multi-channel sequences. After training, the output of the encoder bottleneck layer is extracted as the source domain degradation sensitive feature vector. The system acquires the online monitoring sequence of the current site since the system was put into operation. When the time span of the sequence is less than the time span of a single complete decay cycle sample in the source domain dataset, the current system is determined to be in a data incomplete state. The multi-channel monitoring sequence of the current site is truncated according to the source domain time window length and input into the trained encoder. The initial degradation sensitive feature vector of the target domain with domain offset is obtained in the bottleneck layer. The source domain degradation-sensitive feature vector and the target domain initial degradation-sensitive feature vector are mapped to the same regenerating kernel Hilbert space, respectively. The maximum mean difference between the two is calculated in the regenerating kernel Hilbert space to quantify the degree of offset between the source domain and the target domain in the feature distribution. The weights of all convolutional layers in the encoder are frozen. A domain adaptive normalization layer is introduced after the bottleneck layer. The optimization direction is to minimize the maximum mean difference in the Hilbert space of the regeneration kernel. The scaling and translation parameters of the domain adaptive normalization layer are updated only through gradient backpropagation. This makes the target domain features aligned with the source domain features after transformation by this layer, and outputs a domain-invariant degradation-sensitive feature vector. By inputting the domain-invariant degradation-sensitive feature vector into the activity decay trend prediction network jointly trained by the source domain dataset, the remaining activity lifetime of the current filler is predicted, and the activity decay calibration curve of the current site is constructed accordingly, so as to complete the rapid assessment of the activity decay trend under the condition of incomplete data.

[0071] It should be noted that, firstly, multi-channel time-series monitoring data of porous carrier-supported sulfidated zero-valent iron composite materials from commissioning to failure were collected from multiple long-running similar remediation sites. This data included at least three channels: reduction efficiency, pressure difference between inlet and outlet water in the reduction zone, and concentration of iron ions dissolved from the packing material. The complete sequences from each site were divided into equal-length samples according to a fixed time window length, with each sample corresponding to a decay stage label. The decay stage was defined based on the decrease in reduction efficiency relative to the initial value after commissioning; for example, the initial high-activity period corresponded to a decrease of no more than 20%, the stable reaction period to a decrease between 20% and 50%, the slow decay period to a decrease between 50% and 80%, and the rapid deactivation period to a decrease exceeding 80%. Labels were assigned by professionals after comprehensive verification, combining the inflection point of a sharp increase in pressure difference and the abrupt change in the iron ion dissolution rate. The collection of all labeled samples constitutes the source domain dataset.

[0072] A one-dimensional deep convolutional autoencoder consists of an encoder and a decoder connected in series. The encoder contains several one-dimensional convolutional layers and pooling layers. Each convolutional layer slides a one-dimensional convolutional kernel along the time axis to extract local degradation patterns at different time scales in the multi-channel sequence. The pooling layers downsample the convolutional output in the time dimension, compressing the time span layer by layer. The last layer of the encoder, the bottleneck layer, outputs a low-dimensional feature vector. This vector has a dimension much smaller than that of the input sequence, retaining only the degradation-sensitive information most critical to reconstructing the input. The decoder consists of a corresponding number of upsampling layers and deconvolutional layers, gradually restoring the time dimension from the bottleneck features, and finally reconstructing a multi-channel output with the same dimension as the input sequence. The autoencoder is unsupervised pre-trained on the source domain dataset with the goal of minimizing the mean squared error between the input sequence and the reconstructed sequence. After training, only the encoder part is retained, and its bottleneck layer output is defined as the source domain degradation-sensitive feature vector.

[0073] Differences in water quality, hydrogeological conditions, and operational status across different sites can lead to systematic shifts in the distribution of degradation-sensitive features between the source and target domains. Directly using a pre-trained model from the source domain to predict the target site will produce significant biases; therefore, the degree of shift needs to be quantified first. This embodiment uses the maximum mean difference (MMD) in the regenerating kernel Hilbert space as a metric. The regenerating kernel Hilbert space is a high-dimensional feature space implicitly defined by a kernel function, where the difference between two distributions can be rigorously measured by the distance between their respective embedding means. A Gaussian radial basis kernel is used for calculation, with the kernel bandwidth being the median of pairwise distances among all samples in the source domain. After mapping the feature vectors of the source and target domains to this space, the squared distance between their embedding means is calculated, yielding the MMD value. A larger MMD value indicates a more severe domain shift.

[0074] To eliminate domain offset, a domain adaptive normalization layer is inserted after the bottleneck layer of the encoder. This layer is essentially a learnable channel-wise linear transformation, containing the same scaling and translation parameters as the bottleneck feature dimension. During training, the weights of all convolutional layers in the encoder are frozen, and only the parameters of this layer are used as optimization variables. The objective function is to minimize the MMD value of the source and target domain features after transformation in the Hilbert space of the regenerating kernel. The two parameters are updated through gradient backpropagation until the MMD converges to below a preset threshold. Through this process, the target domain features are distributed and aligned to the source domain feature space while preserving their own degradation patterns, outputting a domain-invariant degradation-sensitive feature vector.

[0075] The activity decay trend prediction network is a multilayer perceptron trained on the source domain dataset. Its input is the source domain degradation-sensitive feature vector, and its output is the corresponding predicted value of the remaining active lifetime of the filler. The supervision signal of the remaining active lifetime is generated by converting the degradation stage labels of each sample into standardized values. Since the domain-invariant degradation-sensitive feature vector has been aligned to the source domain feature space, the prediction network can be directly applied without retraining. For the current target site, the domain-invariant degradation-sensitive feature vector is input into the prediction network to obtain the current predicted value of the remaining active lifetime. As the operating data of the target site gradually accumulates, the above steps are repeated at fixed intervals to obtain a series of remaining lifetime prediction points arranged in time. After smoothing and fitting, these points form the activity decay calibration curve of the current site. The obtained activity decay calibration curve can provide a quantitative basis for determining the timing of composite material activation, regeneration, or replacement in step S2, enabling reliable filler lifetime predictions to be obtained even when there is insufficient data in the early stages of newly built sites.

[0076] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A groundwater circulation well remediation system based on a reduction-oxidation coupled process, comprising a well casing body buried in a groundwater aquifer, a water pump providing power for groundwater circulation, and an oxidant injection device for delivering oxidant into the well, characterized in that: The well casing is arranged from bottom to top along the vertical circulation path of groundwater, with a lower screen, a reduction zone, a water collection and guiding structure, an oxidation zone drop vortex hydraulic structure, and an upper outlet screen section, so that the contaminated groundwater forms a forced vertical circulation path under the drive of the water pump. The reduction zone is filled with a porous carrier-supported sulfidated zero-valent iron composite material. Screens are provided at the upper and lower ends of the reduction zone. The porous carrier disperses and anchors the sulfidated zero-valent iron particles to inhibit surface passivation of zero-valent iron and prolong the reaction activity cycle. The water collection and guiding structure includes a water collection section and a vertical downpipe. The water collection section is located in the cavity space above the porous carrier-supported sulfidated zero-valent iron composite material filler layer and is used to receive the water from the reduction zone. A natural vertical drop of 1.0m to 3.0m is formed between the outlet of the vertical downpipe and the inlet of the drop vortex hydraulic structure in the oxidation zone, which is used to convert gravitational potential energy into water kinetic energy. The oxidation zone drop vortex hydraulic structure includes a vortex chamber and a tangential flow guide structure. The water from the reduction zone, transported through the vertical downpipe, is injected into the vortex chamber along the tangential flow guide structure, spontaneously forming a columnar vortex to enhance the turbulent mixing of water flow and the mass transfer efficiency of pollutants. The oxidant injection point of the oxidant injection device is located near the inlet of the vortex chamber; the oxidation zone vortex hydraulic structure is equipped with an inlet oxidation-reduction potential monitoring point and an outlet oxidation-reduction potential monitoring point. Both monitoring points are connected to the controller. The controller is used to adjust the dosing parameters of the oxidant injection device according to the oxidation-reduction potential data of the two monitoring points in order to maintain a suitable oxidation environment in the oxidation zone.

2. The groundwater circulation well remediation system based on reduction-oxidation coupling process according to claim 1, characterized in that, The porous carrier is selected from one or more combinations of boron-doped ordered mesoporous carbon, cerium-based rare earth porous MOF, hierarchical porous biochar, porous ceramsite, modified zeolite, and sepiolite; the sieve is used to support the porous carrier to fix the sulfided zero-valent iron composite material, while allowing groundwater to pass through uniformly.

3. The groundwater circulation well remediation system based on reduction-oxidation coupling process according to claim 1, characterized in that, The inlet drop edge of the oxidation zone drop vortex hydraulic structure is provided with a sawtooth-shaped structure. When the water flows through the sawtooth-shaped structure, it is divided and refined, which further enhances the water turbulence and mass transfer effect.

4. The groundwater circulation well remediation system based on reduction-oxidation coupling process according to claim 1, characterized in that, The oxidation-reduction potential monitoring point at the water inlet is located on the water inlet side of the inner wall of the vortex chamber, and the oxidation-reduction potential monitoring point at the water outlet is located at the center water outlet of the vortex chamber. The controller is configured to adjust the oxidant dosage based on the deviation between the oxidation-reduction potential value at the effluent point and the preset target value, and to identify pollution load impacts by combining the rate of change of the difference between the oxidation-reduction potentials at the influent point and the effluent point, and adjust the oxidant dosing pulse interval accordingly.

5. The groundwater circulation well remediation system based on reduction-oxidation coupling process according to claim 1, characterized in that, The water pump is a low-lift submersible circulating water pump, which is located on the outlet side of the oxidation zone. It is used to pump the groundwater after oxidation treatment through the upper outlet screen and discharge it to the aquifer. The hydraulic mixing in the oxidation zone is driven entirely by the gravitational potential energy converted from the vertical drop, without the need for additional stirring or aeration equipment.

6. A groundwater circulation well remediation method based on a reduction-oxidation coupling process, applied to the groundwater circulation well remediation system according to any one of claims 1 to 5, characterized in that, Includes the following steps: S1. Circulation Start-up and Hydraulic Regulation: Start the water pump to establish the vertical circulation flow of groundwater, adjust the operating frequency of the water pump, so that the water from the reduction zone flows into the vertical downpipe through the water collection and guiding structure, and forms a natural vertical drop of 1.0m to 3.0m between the water and the inlet of the vortex hydraulic structure in the oxidation zone. At the same time, control the hydraulic residence time in the reduction zone to match the reduction kinetic characteristics of the target pollutants. S2. Reduction process control: Groundwater flows through the porous carrier of the reduction zone and is fixed in the sulfidated zero-valent iron composite material layer, so that pollutants continue to undergo reduction and transformation. Regularly monitor the concentration of target pollutants and intermediate products in the influent and effluent of the reduction zone. When the reduction efficiency drops to a preset threshold, trigger a prompt for the activation, regeneration, or replacement of the composite material. S3. Gravity vortex oxidation start-up: Water from the reduction zone is injected into the vortex chamber through a vertical downpipe along the tangential flow guide structure, and a columnar vortex is spontaneously formed by the vertical drop. The current vortex intensity and theoretical mixing time are calculated based on the signal from the flow meter of the water pump. The vortex Reynolds number is maintained in the strong turbulence range by finely adjusting the water pump flow rate. S4. Intelligent closed-loop control of oxidant dosing: Oxidant is injected into the inlet of the cyclone chamber through the oxidant injection device and the redox potential data of the inlet and outlet are collected in real time. Based on the target value of the redox potential of the outlet, the oxidant dosing amount and the dosing pulse interval are adjusted in a closed loop to maintain a suitable oxidation environment in the oxidation zone. S5. Cyclic Iteration and Remediation Endpoint Determination: Based on the water quality feedback from monitoring wells in and around the well, dynamically adjust the circulation flow rate and oxidant dosage setting value; after the groundwater is treated by a reduction-oxidation sequence coupling process, it is returned to the aquifer and continuously circulated until the concentration of all target pollutants in the target aquifer is stably lower than the remediation standard and the effluent meets the standard after multiple consecutive cycles, at which point the control system stops.

7. The method according to claim 6, characterized in that, In step S3, the current swirl intensity and theoretical mixing time are calculated based on the signal from the flow meter connected to the water pump. The swirl Reynolds number is maintained in the strong turbulence range by fine-tuning the water pump flow rate. Specifically: The real-time flow time series output by the flow meter of the water pump is collected at a preset sampling frequency. The real-time flow time series is high-pass filtered by the built-in flow pulsation separation module to extract the high-frequency hydraulic pulsation component induced by the hydraulic instability of the columnar swirl in the swirl chamber. The phase space of the high-frequency hydraulic pulsation components is reconstructed, the delay time is determined by the mutual information method, and the embedding dimension is determined by the pseudo-nearest neighbor method. The attractor trajectory reflecting the evolution of the vortex structure of the swirling flow is reconstructed. The maximum Lyapunov exponent of the attractor trajectory is calculated and defined as the swirling flow chaos intensity used to quantify the degree of disorder in the complex motion of the swirling flow. The correlation dimension of the attractor trajectory is calculated synchronously, and the obtained correlation dimension is compared with the pre-stored correlation dimension benchmark value of the corresponding fully developed strong turbulent state. When the correlation dimension enters the neighborhood of the benchmark value, it is determined that the vortex has left the transition zone and reached the vortex series dissipation state required for strong turbulent mixing. A two-dimensional adjustment error function is constructed based on the difference between the correlation dimension and the benchmark value, and the deviation between the swirling chaos intensity and the preset chaos intensity target range. Based on the error function, a water pump frequency converter fine-tuning command is generated to slightly adjust the operating frequency to change the real-time flow rate until the correlation dimension converges to the benchmark value and the swirling chaos intensity stably falls within the target range. After the adjustment is stable, the average flow rate value of the current period is extracted, and the nominal turnover period is calculated in combination with the effective volume of the swirl chamber. The chaos-dissipation scaling law calibrated by the experiment and pre-stored is called to correct the nominal turnover period with the current swirl chaos intensity, so as to obtain the theoretical mixing time characterizing the time required to achieve the specified mixing uniformity in the whole chamber after the oxidant is injected.

8. The method according to claim 6, characterized in that, In step S4, the specific logic of the closed-loop adjustment is as follows: When the oxidation-reduction potential value at the effluent point is lower than the preset target value, the dosage of oxidant should be increased. When the rate of change of the oxidation-reduction potential difference between the inlet and outlet points is greater than the preset rate of change threshold, and the difference between the oxidation-reduction potential value at the outlet point and the preset target value is greater than the preset deviation threshold, it is determined to be a high pollution load impact, and the dosage of oxidant is gradually increased and the dosing pulse interval is shortened. When the oxidation-reduction potential at the effluent point stabilizes within the preset target range and the concentration of pollutants in the effluent is lower than the preset limit, the current dosage is maintained to enter steady-state operation.

9. The method according to claim 6, characterized in that, In step S2, the target pollutants include chlorinated hydrocarbons, nitroaromatic hydrocarbons, and high-valence heavy metals.

10. The method according to claim 6, characterized in that, The oxidant is persulfate or hydrogen peroxide. The oxidant is activated in situ by ferrous iron released from zero-valent iron during the reduction reaction, generating strong oxidizing free radicals to deeply degrade pollutants.

Images