Subway leakage water trace element tracing detection method and device

Abstract

The invention discloses a trace detection method and device for trace elements in metro leakage water, and relates to the technical field of metro leakage detection.The method comprises the steps that the stratum structure, the underground water burial depth and background water quality parameters of a target area are obtained, and a set number of leakage suspicious areas are delineated in combination with a tunnel design drawing; and selecting rare earth elements of which the background concentration is less than a set concentration threshold and a fluorescent compound to form a tracer agent. The rare earth elements are combined with the fluorescent compound to realize full-dimensional coverage of deep and shallow layers, and the problem of leak detection of hidden leakage in a traditional single method is solved. Through technology fusion of dual-mode tracing, three-dimensional monitoring and intelligent modeling, comprehensive breakthrough in detection depth, precision, efficiency and risk early warning capability is realized, and a systematic solution is provided for accurate prevention and control of subway deep leakage under complex geological conditions.

Description

Technical Field

[0001] This invention relates to the field of subway leakage detection technology, and in particular to a method and apparatus for detecting trace elements in subway leakage water. Background Technology

[0002] Existing subway water leakage detection technologies mainly include traditional physical detection techniques, manual detection techniques, and early tracing techniques, as detailed below:

[0003] Infrared thermography: This method identifies leaks based on the temperature difference between the leaking area and the surrounding environment, relying on thermal imaging equipment to capture changes in the surface temperature field.

[0004] Laser scanning technology: By scanning the surface of the tunnel with a laser, the location of the leak can be indirectly inferred by utilizing surface deformation.

[0005] Manual inspection: Leakage is judged by visual inspection, tool measurement, and experience.

[0006] Traditional tracing techniques: use radioactive isotopes (such as tritium) as tracers to identify leakage channels by tracing the migration path of the isotopes.

[0007] The problems with the aforementioned traditional detection technologies are:

[0008] Infrared thermography can only detect surface leaks and cannot penetrate structural layers to identify deep or hidden leak channels; it is significantly affected by environmental temperature and humidity, airflow in tunnels, etc., resulting in poor stability of detection results.

[0009] Laser scanning technology relies on surface deformation features (such as micro-expansion caused by wetting), and cannot directly trace the hidden seepage path of leakage; it is not sensitive enough to micro-leakage and is easily affected by the texture of the structure itself.

[0010] Manual inspections are extremely inefficient and difficult to cover long-distance, complex subway tunnels; the results are highly dependent on personnel experience and cannot quantify key parameters such as leakage rate and diffusion coefficient, making dynamic assessment difficult.

[0011] Traditional tracing techniques use radioactive isotopes (such as tritium), which pose environmental pollution risks and are subject to strict regulations on radioactive materials, limiting their application scenarios; they also have low sensitivity and are insufficient for identifying minute leaks. Summary of the Invention

[0012] To address the aforementioned technical problems, this invention provides a method and apparatus for detecting trace elements in subway water leakage. The technical solution is as follows:

[0013] A method for detecting trace elements in subway leakage water includes the following steps:

[0014] Step 1: Obtain the geological structure, groundwater depth and background water quality parameters of the target area, and delineate a set number of suspected leakage areas in conjunction with the tunnel design drawings;

[0015] Step 2: Select rare earth elements with a background concentration lower than a set concentration threshold and fluorescent compounds to form a tracer;

[0016] Step 3: Calculate the total amount of tracer to be added based on the estimated leakage volume and background concentration from the preliminary investigation;

[0017] Step 4: Distribute the injection volume according to the set concentration gradient. For multiple suspected leakage areas, inject dual-mode tracer from the edge to the center of the suspected area according to the allocated injection volume.

[0018] Step 5: Arrange a fixed sampling point at a set distance along the longitudinal direction of the tunnel, and add supplementary sampling points on the side walls and arch according to the grid to form a three-dimensional sampling matrix. Set up an automatic sampler at the sampling point. The automatic sampler collects the changes in tracer concentration in real time according to the set sampling frequency, and records the sampling time and spatial coordinates simultaneously.

[0019] Step 6: Using a mobile device equipped with a plasma mass spectrometer, a fluorescence spectrometer, and an infrared thermal imager, cruise along the tunnel at a set speed, and collect detection data from the plasma mass spectrometer, the fluorescence spectrometer, and the infrared thermal imager at set intervals.

[0020] Step 7: Analyze the collected plasma mass spectrometry data, fluorescence data, and infrared thermography data to obtain a three-dimensional temporal and spatial dataset of tracer concentration.

[0021] Step 8: Based on the seepage-diffusion coupling model, input the spatiotemporal distribution data of tracer concentration and formation parameters, and invert the three-dimensional topology of the seepage channel through numerical simulation; at the same time, calculate the seepage rate and diffusion coefficient.

[0022] Step 9: Input the leakage rate, diffusion coefficient and historical concentration data into the LSTM neural network model, and output the leakage development trend at a future set time, including the leakage growth rate and the direction of channel expansion.

[0023] Optionally, in step 9, the set risk level threshold is used to generate a visualization report of the risk level, scope of impact, and development trend by the LSTM neural network model.

[0024] Optionally, in step 2, a rare earth element with a background concentration of less than 0.1 ppm is selected to form a dual-mode tracer with a fluorescent compound.

[0025] Optionally, the salt tolerance of rare earth elements is ≥25×10⁻⁶. 4 The concentration of the fluorescent compound is mg / L, the excitation wavelength is 490 nm, and the detection limit is ≤0.001 mg / L.

[0026] Optionally, in step 3, the formula for calculating the total amount of tracer administered, W, is:

[0027] ;

[0028] in It is an estimated leakage amount calculated based on Darcy's law; It is the geological adsorption correction coefficient; This is the background concentration.

[0029] Optionally, in step 4, 3-5 injection points are set up from the edge to the center for each suspected leakage area. The tracer is injected through a pre-set 50-80mm borehole, and the injection volume is distributed according to the concentration gradient of 50ppm to 100ppm and then to 200ppm. During injection, a micro pump is used to control the pressure at 0.2-0.5MPa.

[0030] Optionally, in step 7, the collected plasma mass spectrometry data, fluorescence data, and infrared thermographic data are spatiotemporally synchronized, environmental interference noise is removed by wavelet transform, and a three-dimensional temporal and spatial dataset of tracer concentration is generated by combining the laser point cloud intensity correction algorithm.

[0031] Optionally, in step 8, the seepage-diffusion coupling model is constructed based on the bidirectional coupling logic of seepage field driving diffusion field and diffusion field inverting seepage field;

[0032] Based on Darcy's law for heterogeneous environments, and considering formation porosity, permeability tensor, and fluid viscosity, a three-dimensional flow control equation is established:

[0033]

[0034] Where P is the fluid pressure. Let t be the fluid density and t be the time.

[0035] Based on Fick's law and the adsorption-desorption effect of tracers in porous media, a tracer concentration migration equation is established:

[0036] ;

[0037] Where D is the diffusion coefficient. It is the seepage velocity. It is the formation adsorption attenuation coefficient.

[0038] A trace element tracer detection device for subway leakage water is provided to implement a method for trace element tracer detection of subway leakage water. The device includes a tracer injection module, a multimodal sampling unit, and a data analysis terminal. The tracer injection module includes a micro pump, a multi-channel grouting pipe, a concentration monitoring sensor, and a pressure sensor. The power output end of the micro pump is connected to the main pipe of the multi-channel grouting pipe through a high-pressure pipeline. The branch pipe openings of the grouting pipe correspond to preset injection points in suspected leakage areas. The concentration monitoring sensor and the pressure sensor are respectively installed at the outlet end of the multi-channel grouting pipe to monitor the concentration of the injected tracer in real time.

[0039] The multimodal sampling unit includes an automatic sampler, a mobile detection vehicle, an inductively coupled plasma mass spectrometer, a fluorescence spectrometer, an infrared thermal imager, an automatic cruise guide rail, and a positioning module. The automatic sampler is installed at a fixed sampling point. The mobile detection vehicle moves along a track inside the tunnel. The inductively coupled plasma mass spectrometer, fluorescence spectrometer, infrared thermal imager, automatic cruise guide rail, and positioning module are installed on the mobile detection vehicle. The data analysis terminal is communicatively connected to the automatic sampler, inductively coupled plasma mass spectrometer, fluorescence spectrometer, infrared thermal imager, automatic cruise guide rail, and positioning module.

[0040] Optionally, the automatic sampler includes a water sample storage chamber and a temperature control unit. The water sample collection port of the water sample storage chamber is connected to the tunnel seepage water collection point via a hose, and the temperature control unit is connected to a data analysis terminal to receive sampling trigger commands.

[0041] In summary, the present invention has at least one of the following beneficial technical effects:

[0042] This invention provides a method and device for detecting trace elements in subway leakage water. The combination of rare earth elements and fluorescent compounds achieves full-dimensional coverage of both deep and shallow layers, solving the problem of missed detection of hidden leakage by traditional single methods.

[0043] The inductively coupled plasma mass spectrometer achieves a detection resolution of 0.1 ppb for rare earth elements, while the fluorescence spectrometer has a dynamic range of 10. 6 By combining dense sampling points with a three-dimensional sampling matrix, it can capture trace leaks as low as 0.05 L / min, improving detection sensitivity by three orders of magnitude compared to traditional techniques.

[0044] Based on the seepage-diffusion coupling model and the spatiotemporal distribution data of tracer concentration, the three-dimensional morphology of leakage channels can be inverted with small reconstruction error and support millimeter-level crack identification, solving the problem that traditional manual inspection is difficult to quantify the structure of leakage channels.

[0045] By coupling the concentration gradient with Darcy's law, leakage rate and diffusion coefficient can be accurately calculated, providing a quantitative basis for leakage risk assessment and overcoming the limitations of traditional methods that rely on experience and have vague parameters.

[0046] From multi-source data fusion to 3D modeling and trend prediction, everything is done automatically through algorithms, reducing human intervention, avoiding subjective errors in traditional manual analysis, and significantly improving data processing efficiency.

[0047] LSTM neural network models can predict future leakage trends based on historical data with small prediction errors, enabling a shift from passive detection to proactive prevention and control.

[0048] By integrating dual-mode tracing, three-dimensional monitoring, and intelligent modeling technologies, a comprehensive breakthrough has been achieved in detection depth, accuracy, efficiency, and risk warning capabilities, providing a systematic solution for the precise prevention and control of deep seepage in subways under complex geological conditions. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the process for detecting trace elements in subway water leakage according to the present invention. Detailed Implementation

[0050] The present invention will be further described in detail below with reference to the accompanying drawings.

[0051] This invention discloses a method and apparatus for detecting trace elements in subway water leakage.

[0052] Reference Figure 1 Example 1: A method for detecting trace elements in subway water leakage, comprising the following steps:

[0053] Step 1: Obtain the geological structure, groundwater depth and background water quality parameters of the target area, and delineate a set number of suspected leakage areas in conjunction with the tunnel design drawings;

[0054] Step 2: Select rare earth elements with a background concentration lower than a set concentration threshold and fluorescent compounds to form a tracer;

[0055] Step 3: Calculate the total amount of tracer to be added based on the estimated leakage volume and background concentration from the preliminary investigation;

[0056] Step 4: Distribute the injection volume according to the set concentration gradient. For multiple suspected leakage areas, inject dual-mode tracer from the edge to the center of the suspected area according to the allocated injection volume.

[0057] Step 5: Arrange a fixed sampling point at a set distance along the longitudinal direction of the tunnel, and add supplementary sampling points on the side walls and arch according to the grid to form a three-dimensional sampling matrix. Set up an automatic sampler at the sampling point. The automatic sampler collects the changes in tracer concentration in real time according to the set sampling frequency, and records the sampling time and spatial coordinates simultaneously.

[0058] Step 6: Using a mobile device equipped with a plasma mass spectrometer, a fluorescence spectrometer, and an infrared thermal imager, cruise along the tunnel at a set speed, and collect detection data from the plasma mass spectrometer, the fluorescence spectrometer, and the infrared thermal imager at set intervals.

[0059] Step 7: Analyze the collected plasma mass spectrometry data, fluorescence data, and infrared thermography data to obtain a three-dimensional temporal and spatial dataset of tracer concentration.

[0060] Step 8: Based on the seepage-diffusion coupling model, input the spatiotemporal distribution data of tracer concentration and formation parameters, and invert the three-dimensional topology of the seepage channel through numerical simulation; at the same time, calculate the seepage rate and diffusion coefficient.

[0061] Step 9: Input the leakage rate, diffusion coefficient and historical concentration data into the LSTM neural network model, and output the leakage development trend at a future set time, including the leakage growth rate and the direction of channel expansion.

[0062] By adopting the above technical solution, rare earth elements, such as Ce and La, can trace deep and hidden seepage paths, and fluorescent compounds, such as sodium fluorescein, can quickly respond to near-surface micro-leakage. The combination of the two achieves full-dimensional coverage of deep and shallow layers, solving the problem of missed detection of hidden seepage by traditional single methods.

[0063] The inductively coupled plasma mass spectrometer (ICP-MS) achieves a detection resolution of 0.1 ppb for rare earth elements, while the fluorescence spectrometer has a dynamic range of 10. 6 By combining dense sampling points with a three-dimensional sampling matrix, it can capture trace leaks as low as 0.05 L / min, improving detection sensitivity by three orders of magnitude compared to traditional techniques.

[0064] Based on the seepage-diffusion coupling model and the spatiotemporal distribution data of tracer concentration, the three-dimensional morphology (length, width, and direction) of the leakage channel can be inverted with a reconstruction error of <5%, supporting millimeter-level crack identification, and solving the problem that traditional manual inspection is difficult to quantify the structure of the leakage channel.

[0065] By coupling the concentration gradient with Darcy's law, leakage rate and diffusion coefficient can be accurately calculated, providing a quantitative basis for leakage risk assessment and overcoming the limitations of traditional methods that rely on experience and have vague parameters.

[0066] Fixed automatic samplers (1 hour / time) and mobile inspection vehicles (2km / h cruise, 30 seconds / time data collection) form a static and dynamic sampling network, realizing full tunnel monitoring without blind spots. The detection efficiency is 5-10 times higher than manual inspection, and the detection time for a single tunnel section is less than 24 hours.

[0067] From multi-source data fusion (ICP-MS, fluorescence, infrared) to 3D modeling and trend prediction, all processes are completed automatically through algorithms, reducing human intervention, avoiding subjective errors in traditional manual analysis, and improving data processing efficiency by more than 80%.

[0068] Accurate prediction of dynamic trends: The LSTM neural network model can predict the leakage development trend in the next 72 hours based on historical data, such as the growth rate of leakage and the direction of channel expansion. The prediction error is small, realizing the transformation from passive detection to active prevention and control.

[0069] By integrating dual-mode tracing, three-dimensional monitoring, and intelligent modeling technologies, a comprehensive breakthrough has been achieved in detection depth, accuracy, efficiency, and risk warning capabilities, providing a systematic solution for the precise prevention and control of deep seepage in subways under complex geological conditions.

[0070] In Example 2, step 9, the set risk level threshold is used to generate a visual report of the risk level, scope of impact, and development trend by the LSTM neural network model.

[0071] By adopting the above technical solutions, risk classification and repair guidance are provided. By setting risk thresholds (high risk >1L / min, medium risk 0.1-1L / min, low risk <0.1L / min), a visual risk report and repair suggestions are generated, such as the selection of grouting materials and pressure parameters. This provides a direct basis for engineering decisions, reduces structural damage caused by leakage, such as segment corrosion and track bed frost heave, and indirectly reduces maintenance costs.

[0072] In Example 3, step 2, rare earth elements with a background concentration of less than 0.1 ppm and fluorescent compounds were selected to form a dual-mode tracer.

[0073] Example 4: Salt tolerance of rare earth elements ≥ 25 × 10⁻⁶ 4 The concentration of the fluorescent compound is mg / L, the excitation wavelength is 490 nm, and the detection limit is ≤0.001 mg / L.

[0074] By adopting the above technical solution, the background concentration of rare earth elements, such as Ce and La, is <0.1ppm, which avoids interference from natural background values ​​in the formation, and the salt tolerance is ≥25×10 4 mg / L, it can adapt to high-salt and complex strata, such as the saline-alkali geology of coastal subways, and accurately track deep and hidden seepage paths; fluorescent compounds, such as sodium fluorescein with an excitation wavelength of 490nm and a detection limit of ≤0.001mg / L, can quickly respond to near-surface micro-leakage, such as dripping from the joints of pipe segments. The combination of the two achieves full-dimensional coverage of seepage paths in both deep and shallow layers, solving the problem of insufficient tracking ability of traditional single tracers for deep seepage.

[0075] The natural low background of rare earth elements (<0.1ppm), combined with the 0.1ppb resolution of inductively coupled plasma mass spectrometry (ICP-MS), can capture trace amounts of tracer migration signals, improving the detection sensitivity by three orders of magnitude compared to traditional radioisotope tracers.

[0076] The low detection limit of fluorescent compounds (≤0.001 mg / L) and specific excitation wavelength (490 nm) allow for precise differentiation of tracer signals from ambient stray light using a fluorescence spectrometer, significantly improving anti-interference capabilities compared to infrared thermography.

[0077] The tracer migration path is clear and traceable: the chemical stability of rare earth elements and the rapid response characteristics of fluorescent compounds can clearly distinguish deep seepage channels, such as rock fissures, from shallow surface seepage through the spatiotemporal distribution of concentration. Combined with the seepage-diffusion coupling model, the three-dimensional topological structure inversion error of seepage channels is less than 5%, which is an order of magnitude higher than the positioning accuracy of laser scanning technology that relies on surface deformation.

[0078] The concentration data of the model tracer can be cross-validated, reducing the error caused by the formation adsorption and chemical reaction of a single tracer, making the leakage rate calculation error ≤0.05L / min and the diffusion coefficient fitting accuracy ≥95%, thus providing a reliable basis for the quantitative assessment of leakage.

[0079] Non-radioactive rare earth elements and biodegradable fluorescent compounds are selected to replace traditional radioactive isotopes (such as tritium), avoiding radiation pollution and regulatory restrictions, and meeting the green and environmentally friendly requirements of subway engineering.

[0080] Example 5, in step 3, the formula for calculating the total amount W of tracer administered is:

[0081] ;

[0082] in It is an estimated leakage amount calculated based on Darcy's law; It is the geological adsorption correction coefficient; This is the background concentration.

[0083] By adopting the above technical solution, the three core parameters of estimated leakage volume, geological adsorption correction coefficient, and background concentration are directly linked to achieve scientific quantification of the dosage: the estimated leakage volume calculated by Darcy's law can be dynamically adjusted according to the scale of leakage, such as increasing the dosage in large leakage areas and reducing the dosage in micro leakage areas, avoiding the problems of excessive waste or insufficient tracking caused by traditional experience-based dosage.

[0084] A geological adsorption correction coefficient is introduced, and a higher value is taken for different strata, such as clay layer with strong adsorption and sand layer with weak adsorption.

[0085] Adjusting the dosage ensures that the tracer maintains a detectable concentration during migration, significantly improving the effective utilization rate compared to traditional methods that do not consider adsorption (such as fixed-ratio dosing).

[0086] To improve tracer signal identification and reduce background interference, the formula incorporates a background concentration parameter, allowing for dynamic adjustment of the dosage based on the natural content of the tracer in the formation. For areas with high background concentrations, such as a formation with a natural Ce content close to 0.1 ppm, the dosage is increased through formula calculation to ensure that the tracer concentration is significantly higher than the background value (usually more than 10 times the background value), preventing the signal from being masked by the natural background. For areas with low background concentrations, such as deep rock masses with La content < 0.05 ppm, the dosage is reduced, which meets the detection requirements while avoiding the environmental accumulation risk caused by excessive tracer dosage, thus improving signal identification compared to a fixed concentration.

[0087] This provides a reliable foundation for subsequent concentration inversion modeling. Accurately calculated dosage ensures that the tracer forms an identifiable concentration gradient in the leakage channel. Combined with subsequent sampling concentration data, the migration rate and diffusion path of the tracer can be more accurately derived, providing reliable initial parameters for the inversion of the three-dimensional topology of the leakage channel by the seepage-diffusion coupling model. It avoids model inversion errors caused by unreasonable dosage, such as excessive dosage leading to concentration saturation or insufficient dosage leading to signal loss. This reduces the error in leakage rate calculation and improves the diffusion coefficient fitting accuracy to over 90%, significantly enhancing the modeling reliability compared to traditional empirical dosage.

[0088] In Example 6, in step 4, 3-5 injection points are set up from the edge to the center for each suspected leakage area. The tracer is injected through a pre-drilled hole of 50-80mm. The injection volume is distributed according to the concentration gradient of 50ppm to 100ppm and then to 200ppm. During injection, a micro pump is used to control the pressure at 0.2-0.5MPa.

[0089] By adopting the above technical solution, injecting from the edge to the center of the suspected leakage area at a concentration gradient of 50ppm to 100ppm and then to 200ppm, a gradient distribution of low concentration at the edge and high concentration at the center can be formed in the suspected area. This allows the tracer to show a clear concentration change pattern as it diffuses with the seepage, such as decreasing from the center to the edge. This provides a clear signal marker for subsequent inversion of the leakage channel through the concentration gradient, improving the identification of the migration path compared to injection of a single concentration.

[0090] Three to five injection points are set up in each suspected area. Combined with the precise positioning of pre-drilled holes (50-80mm), the hidden cracks inside the suspected area can be covered, such as the sides of construction joints and around the joints of pipe segments. This avoids the problem of uneven distribution of tracer caused by single-point injection and ensures that both deep and shallow seepage paths can be effectively marked.

[0091] By using a micro-pump to control the injection pressure at 0.2-0.5MPa, the tracer can effectively penetrate deep into the fractures to overcome the pore resistance of the formation, while avoiding formation disturbances caused by high pressure, such as fracture expansion and segment deformation. This solves the secondary engineering risks caused by uncontrolled pressure in traditional grouting, such as damage to the tunnel structure.

[0092] With a borehole diameter of 50-80mm, it can be matched with multi-channel grouting pipes and is compatible with tracer solutions of different particle sizes, such as suspensions containing rare earth elements. At the same time, it reduces the damage to the tunnel structure caused by drilling. Compared with traditional large-diameter boreholes, it reduces structural damage and is especially suitable for safety inspection of existing operating subways.

[0093] In Example 7, step 7 involves spatiotemporal synchronization of the collected plasma mass spectrometry data, fluorescence data, and infrared thermographic data. Environmental interference noise is removed through wavelet transform, and a laser point cloud intensity correction algorithm is used to generate a three-dimensional spatiotemporal dataset of tracer concentration.

[0094] In Example 8, step 8, the seepage-diffusion coupling model is constructed based on the bidirectional coupling logic of seepage field driving diffusion field and diffusion field inverting seepage field;

[0095] Based on Darcy's law for heterogeneous environments, and considering formation porosity, permeability tensor, and fluid viscosity, a three-dimensional flow control equation is established:

[0096]

[0097] Where P is the fluid pressure. Let t be the fluid density and t be the time.

[0098] Based on Fick's law and the adsorption-desorption effect of tracers in porous media, a tracer concentration migration equation is established:

[0099] ;

[0100] Where D is the diffusion coefficient. It is the seepage velocity. It is the formation adsorption attenuation coefficient.

[0101] By employing the above technical solutions, spatiotemporal synchronization of plasma mass spectrometry data, fluorescence data, and infrared thermography data is achieved, ensuring precise temporal and spatial correspondence between data from different devices and sampling points. This resolves the matching error problem caused by time differences or location deviations in multi-source data. Combining wavelet transform to remove environmental interference noise, such as spurious concentration fluctuations caused by equipment vibration and temperature / humidity fluctuations, improves the data signal-to-noise ratio, providing high-purity tracer concentration signals for subsequent modeling.

[0102] By combining the laser point cloud intensity correction algorithm, the influence of tunnel structural cracks and real leakage channels on tracer migration can be accurately distinguished, avoiding the interference of tracer concentration anomalies caused by structural cracks in the analysis. This makes the generated tracer concentration-time-space three-dimensional dataset more focused on the real leakage path, and the inversion error is significantly reduced when using the uncorrected dataset for modeling.

[0103] The three-dimensional seepage control equations, constructed based on Darcy's law for heterogeneous formations, incorporate formation porosity and permeability tensors to reflect geological anisotropy and fluid viscosity. This overcomes the simplistic assumptions of traditional homogeneous models regarding complex formations, such as rock masses with uneven fracture development and layered soil layers. It can more accurately describe the seepage characteristics of different regions, such as the difference in seepage velocity between densely fractured areas and intact rock masses, significantly reducing the calculation error of the seepage field. A bidirectional coupling logic enables dynamic correlation between seepage diffusion, improving the reliability of the inversion.

[0104] The model is based on a two-way coupling logic: the seepage field drives the diffusion field, and the diffusion field inverts the seepage field.

[0105] The seepage velocity calculated in the seepage field is used as the input to the diffusion field equation to clarify the dominant direction of tracer migration with the fluid;

[0106] The tracer concentration distribution monitored in the diffusion field is fed back into the seepage field. The permeability tensor is corrected by the least squares method, and the accuracy of the seepage field is iteratively optimized until the error between the calculated concentration and the measured concentration is less than 5%.

[0107] Compared with traditional single seepage models or diffusion models, this dynamic correlation mechanism can more realistically reflect the actual physical process of seepage driving tracer migration and tracer migration inferring seepage characteristics.

[0108] Based on Fick's law and combined with the tracer concentration migration equation of adsorption-desorption effect, the formation adsorption attenuation coefficient is introduced to accurately quantify the adsorption loss of tracers in porous media, such as the adsorption of rare earth elements in clay layers, and avoid the deviation in diffusion coefficient calculation caused by ignoring the adsorption effect.

[0109] Example 9: A trace element tracer detection device for subway leakage water, used to implement a trace element tracer detection method for subway leakage water. The device includes a tracer injection module, a multimodal sampling unit, and a data analysis terminal. The tracer injection module includes a micro pump, a multi-channel grouting pipe, a concentration monitoring sensor, and a pressure sensor. The power output end of the micro pump is connected to the main pipe of the multi-channel grouting pipe through a high-pressure pipeline. The branch pipe openings of the grouting pipe correspond to preset injection points in suspected leakage areas. The concentration monitoring sensor and the pressure sensor are respectively installed at the outlet end of the multi-channel grouting pipe to monitor the concentration of the injected tracer in real time.

[0110] The multimodal sampling unit includes an automatic sampler, a mobile detection vehicle, an inductively coupled plasma mass spectrometer, a fluorescence spectrometer, an infrared thermal imager, an automatic cruise guide rail, and a positioning module. The automatic sampler is installed at a fixed sampling point, the mobile detection vehicle moves along the track inside the tunnel, and the inductively coupled plasma mass spectrometer, fluorescence spectrometer, infrared thermal imager, automatic cruise guide rail, and positioning module are installed on the mobile detection vehicle. The data analysis terminal is communicatively connected to the automatic sampler, inductively coupled plasma mass spectrometer, fluorescence spectrometer, infrared thermal imager, automatic cruise guide rail, and positioning module, respectively.

[0111] Example 10: The automatic sampler includes a water sample storage chamber and a temperature control unit. The water sample collection port of the water sample storage chamber is connected to the tunnel seepage water collection point through a hose. The temperature control unit is connected to the data analysis terminal and receives sampling trigger commands.

[0112] By adopting the above technical solution, the data analysis terminal calculates the total dosage and gradient concentration distribution scheme of the dual-mode tracer, rare earth elements and fluorescent compounds, such as 50ppm to 100ppm to 200ppm, based on the location of suspected leakage areas, estimated leakage volume and background concentration from the previous investigation. The terminal then sends the injection volume, pressure threshold, 0.2-0.5MPa and other instructions to the tracer injection module via the industrial bus.

[0113] After receiving the command, the micro pump delivers the tracer to the main pipe of the multi-channel grouting pipe through the high-pressure pipeline; the branch pipe openings of the grouting pipe correspond to the preset injection points in the suspected leakage area, 3-5 per area, distributed from the edge to the center, and injected sequentially according to the gradient concentration.

[0114] A concentration monitoring sensor installed at the outlet of the grouting pipe detects the concentration of the injected tracer in real time, while a pressure sensor monitors the injection pressure simultaneously. If the concentration deviates from the preset gradient or the pressure exceeds the threshold, the sensor feeds the data back to the data analysis terminal, which then adjusts the flow rate of the micro pump in real time to ensure that the injection parameters are accurate and controllable.

[0115] Fixed sampling point monitoring: The automatic sampler is arranged according to the preset three-dimensional sampling matrix, with one fixed point every 10m in the longitudinal direction, and supplementary points in a 5×5m grid on the side walls and arch. Its water sample collection port is connected to the tunnel seepage water collection point, such as drainage ditch or seepage drip point, through a hose.

[0116] The temperature control unit receives sampling trigger commands from the data analysis terminal, such as timed sampling every hour or triggered by a tracer concentration threshold, and maintains the temperature of the water sample storage chamber at 4°C to prevent tracer degradation.

[0117] The automatic sampler synchronously records the sampling time and spatial coordinates, and feeds back the sampling status to the terminal.

[0118] Mobile dynamic detection: The mobile detection vehicle cruises along the tunnel at a preset speed, such as 2 km / h, via an automatic cruise guide rail or a wheel drive system. The positioning module and GPS mileage marker calibration obtain location information in real time and synchronize it to the data analysis terminal.

[0119] The inductively coupled plasma mass spectrometer (ICP-MS) is connected to the water sample storage chamber of the autosampler via a sample introduction line. It collects water samples every 30 seconds to detect the concentration of rare earth elements with a resolution of 0.1 ppb.

[0120] The fluorescence spectrometer simultaneously acquires fluorescence signals from water samples at an excitation wavelength of 490 nm, capturing changes in the concentration of fluorescent compounds.

[0121] Infrared thermal imagers capture real-time temperature field images of the tunnel arch and sidewalls to help eliminate surface temperature interference. The above detection data, including ICP-MS data, fluorescence data, and infrared thermal image data, are bound to the location information of the positioning module and transmitted wirelessly to the data analysis terminal via 5G.

[0122] After receiving the raw data from the multimodal sampling unit, the terminal first uses a spatiotemporal synchronization algorithm to align the fixed sampling point data with the mobile detection data based on mileage positioning, forming a preliminary time-space-concentration dataset. Then, it uses wavelet transform to filter environmental noise, such as concentration fluctuations caused by equipment vibration, and combines a laser point cloud intensity correction algorithm to eliminate interference from structural cracks, ultimately generating a high-purity tracer concentration spatiotemporal three-dimensional dataset.

[0123] The terminal calls the built-in seepage-diffusion coupling model algorithm library, inputting the spatiotemporal distribution data of tracer concentration and formation parameters:

[0124] The seepage velocity is calculated based on the three-dimensional seepage equation of the heterogeneous Darcy's law.

[0125] By combining the tracer migration equation with adsorption-desorption effect, the three-dimensional topology (length, direction, and branches) of the leakage channel is inverted, and the leakage rate and diffusion coefficient are calculated.

[0126] Generate a 3D model and heat map of the leakage channels to visually display the leakage distribution.

[0127] The terminal inputs leakage rate, diffusion coefficient and historical concentration data into the LSTM neural network model to predict the leakage development trend in the next 72 hours, such as the growth rate of leakage volume and the direction of channel expansion, and classifies the risk level (high risk / medium risk / low risk) according to the threshold.

[0128] The terminal generates a visual report containing risk level and remediation suggestions (such as acrylate grouting), which is then pushed to the operation and maintenance system via an early warning signal transmitter;

[0129] Simultaneously, the terminal sends subsequent instructions to the tracer injection module and the multimodal sampling unit, such as whether to replenish the tracer injection or adjust the sampling frequency, thus forming a closed-loop control.

[0130] The following examples illustrate the implementation principle of the trace element tracing detection method and device for subway leakage water according to the present invention:

[0131] A section of Metro Line 3 in a certain city has a tunnel length of 1.2km, traversing strata of alternating clay and gravel layers. During operation, dripping and flowing water were repeatedly observed at the construction joints of the arch and side walls. Traditional infrared thermography can only locate surface wet areas and cannot trace deeper leakage channels. The detection method and device of this invention are used for precise detection. The specific implementation process is as follows:

[0132] Step 1: Delineate the suspected leakage area:

[0133] Collect stratigraphic survey report for the target area: clay layer porosity 25%, permeability 1.2×10⁻ 5 cm / s, porosity of the sand and gravel layer is 35%, and permeability is 5.8×10⁻ 4 cm / s; groundwater depth 8-12m, background concentration of Ce element 0.03ppm, La element 0.02ppm in background water quality.

[0134] Based on the tunnel design drawings, we focused on construction joints (12 in total), expansion joints (3 in total), and areas with dense segment joints, and identified 5 suspected leakage areas (numbered S1-S5, each area with a longitudinal length of about 50m).

[0135] Step 2: Dual-mode tracer selection:

[0136] Rare earth element: Ce(NO3)3 was selected (background concentration 0.03ppm < 0.1ppm, salt tolerance 28×10). 4 mg / L, adapted to the high-salt environment of sand and gravel layers).

[0137] Fluorescent compound: Sodium fluorescein (excitation wavelength 490nm, detection limit 0.0008mg / L≤0.001mg / L, suitable for near-surface trace leakage detection).

[0138] Step 3: Calculation of total tracer dosage:

[0139] Based on Darcy's law, the estimated leakage rate Q is calculated as follows: the average leakage rates of zones S1-S5 are 0.3 L / min, 0.5 L / min, 0.2 L / min, 0.6 L / min, and 0.4 L / min, respectively, with a total Q = 2.0 L / min.

[0140] Geological adsorption correction coefficient K: 1.1 for clay layer area, 0.7 for sand and gravel layer, weighted average K=0.9.

[0141] Background concentration C = 0.03 ppm (calculated as elemental Ce).

[0142] Total deployment amount W = K × Q × ×10⁻ 9 =0.9×2.0×0.03×10⁻ 9 =5.4×10⁻¹¹kg converted to solution volume: based on a Ce(NO3)3 concentration of 100ppm, the total injection volume is approximately 540L.

[0143] Step 4: Gradient concentration injection:

[0144] Four injection points (60mm borehole diameter) are set up from the edge to the center of each suspected area, with injection volumes distributed in gradients from 50ppm to 100ppm to 200ppm:

[0145] S1 zone: Inject 30L of 50ppm solution at each of the two edge points, and inject 20L of 100ppm solution and 15L of 200ppm solution at each of the two center points;

[0146] For other areas, the injection pressure is adjusted according to the leakage rate, and the injection pressure is controlled by a micro pump at 0.3-0.4 MPa (higher value for clay layers and lower value for sand and gravel layers).

[0147] Step 5: Layout of the 3D sampling matrix:

[0148] Fixed sampling points: 1 point is set every 10m along the longitudinal direction of the tunnel, for a total of 120 points. 80 additional points are set on the side walls and arches according to a 5×5m grid, for a total of 200 sampling points.

[0149] The automatic sampler is set to a sampling frequency of 1 hour / time, and synchronously records the sampling time (accurate to the second) and spatial coordinates (such as the arch at K3+120.5m).

[0150] Step 6: Motion Detection

[0151] The mobile testing vehicle cruises at a speed of 2 km / h, collecting one set of data every 30 seconds.

[0152] ICP-MS was used to determine the Ce elemental concentration (resolution 0.1 ppb).

[0153] The intensity of sodium fluorescein was detected by a fluorescence spectrometer (excitation wavelength 490 nm).

[0154] Temperature field images captured by an infrared thermal imager (resolution 640×512).

[0155] Step 7: Data Processing

[0156] Spatiotemporal synchronization: Based on mileage marker positioning, the fixed sampling points are aligned with the mobile detection data, with an error of ≤0.5m.

[0157] Noise filtering: Concentration fluctuations caused by equipment vibration are removed through wavelet transform (signal-to-noise ratio improved by 45%).

[0158] Laser point cloud correction: Eliminate concentration anomalies caused by 3 structural cracks (non-leakage channels) and generate a concentration-time-space 3D dataset.

[0159] Step 8: Leakage Channel Inversion

[0160] Input parameters for the flow-diffusion coupling model: formation porosity, permeability tensor, and tracer concentration data.

[0161] Inversion results:

[0162] There is a deep seepage channel in the construction joint of the arch in S2 area, which is about 15m long, runs at an angle of 30° to the tunnel axis, and has a seepage rate of 0.8L / min.

[0163] Two shallow channels exist on the sidewall of zone S4, with a leakage rate of 0.3 L / min and a diffusion coefficient of 1.2 × 10⁻⁻⁻⁶. 6 m² / s;

[0164] The 3D topology reconstruction error is 3.2% (<5%).

[0165] Step 9: Intelligent Early Warning and Risk Classification:

[0166] An LSTM neural network is input into historical concentration data to predict trends over the next 72 hours.

[0167] The leakage in zone S2 will increase by 20%, and the channel will extend towards K3+250m.

[0168] Risk levels: Zone S2 is high risk (0.8L / min is close to the high risk threshold, and the expansion trend indicates it is high risk), and Zone S4 is medium risk.

[0169] A visualization report was generated, recommending that acrylate grouting (pressure 0.4MPa) be used in zone S2 and polyurethane grouting (pressure 0.2MPa) be used in zone S4.

[0170] The inspection took 20 hours, which is 8 times more efficient than manual inspection (about 160 hours);

[0171] A total of 3 deep leakage channels and 5 shallow leakage channels were identified, with a minimum leakage rate of 0.06 L / min (the missed detection rate of traditional infrared thermography is 60%).

[0172] A follow-up inspection one month after the repair showed that the leakage was completely stopped, verifying the accuracy of the detection and the effectiveness of the solution.

[0173] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A subway leakage water microelement tracer detection method, characterized in that, The method comprises the following steps: Step 1, obtain the stratum structure, groundwater depth and background water quality parameters of the target area, and combine the tunnel design drawings to circle a plurality of leakage suspicious areas with a set number; Step 2, select a rare earth element with a background concentration less than a set concentration threshold and a fluorescent compound to form a tracer; Step 3, calculate the total injection amount of the tracer based on the estimated leakage amount and the background concentration in the previous investigation; Step 4, distribute the injection amount according to a set concentration gradient, and inject the double-mode tracer in the suspicious areas from the edge to the center according to the distributed injection amount; Step 5, arrange a fixed sampling point every set distance along the longitudinal direction of the tunnel, additionally arrange supplementary sampling points on the sidewall and the vault in a grid manner to form a three-dimensional sampling matrix, and set an automatic sampler at the sampling point, which collects the concentration change of the tracer in real time according to a set sampling frequency, and synchronously records the sampling time and the spatial coordinates; Step 6, use a mobile device to carry an ion mass spectrometer, a fluorescence spectrometer and an infrared thermal imager, cruise along the tunnel at a set speed, and collect the detection data of the ion mass spectrometer, the fluorescence spectrometer and the infrared thermal imager every set time; Step 7, analyze the collected ion mass spectrometer data, fluorescence data and infrared thermal image data to obtain a three-dimensional data set of the concentration of the tracer in time and space; Step 8, input the concentration of the tracer in time and space and the stratum parameters based on a seepage diffusion coupling model, and inversely calculate the three-dimensional topological structure of the leakage channel through numerical simulation; meanwhile, calculate the leakage rate and the diffusion coefficient; Step 9, input the leakage rate, the diffusion coefficient and the historical concentration data into an LSTM neural network model, and output the leakage development trend of a future set time, including the leakage amount growth amplitude and the channel expansion direction.

2. The subway water leakage microelement tracer detection method according to claim 1, characterized in that, In step 9, a risk level threshold is set, and the LSTM neural network model generates a visual report of the risk level, the influence range and the development trend.

3. The method according to claim 2, wherein the method is characterized by, In step 2, a rare earth element with a background concentration less than 0.1 ppm and a fluorescent compound are selected to form a double-mode tracer.

4. The subway water leakage microelement tracer detection method according to claim 3, characterized in that, Salt tolerance of rare earth elements The excitation wavelength of the fluorescent compound is 490 nm, and the detection limit is ≤0.001 mg / L.

5. The subway water leakage microelement tracer detection method according to claim 4, characterized in that, In step 3, the calculation formula of the total injection amount W of the tracer is: ； wherein is the estimated leakage based on Darcy's law; is the geological adsorption correction factor; is the background concentration.

6. The subway water leakage microelement tracer detection method according to claim 5, characterized in that, In step 4, 3-5 injection points are arranged from the edge to the center of each leakage suspicious area, the tracer is injected through a pre-set 50-80 mm borehole, the injection amount is distributed according to a concentration gradient of 50 ppm to 100 ppm to 200 ppm, and a micro-pump is used to control the pressure at 0.2-0.5 MPa during injection.

7. The subway water leakage microelement tracer detection method according to claim 6, characterized in that, In step 7, the collected ion mass spectrometer data, fluorescence data and infrared thermal image data are synchronized in time and space, environmental interference noise is removed through wavelet transform, and a three-dimensional data set of the concentration of the tracer in time and space is generated by combining a laser point cloud intensity correction algorithm.

8. The subway water leakage microelement tracer detection method according to claim 7, characterized in that, In step 8, the seepage diffusion coupling model is constructed based on the logic of bidirectional coupling of the seepage field driving the diffusion field and the diffusion field inversely calculating the seepage field; Based on the heterogeneous Darcy law, the three-dimensional seepage control equation is established by considering the stratum porosity, the permeability tensor and the fluid viscosity; ； where P is fluid pressure, is fluid density, t is time; Based on Fick's law, the tracer concentration migration equation is established by combining the adsorption and desorption effect of the tracer in the pore medium: ； where D is the diffusion coefficient, is the seepage velocity, is the formation adsorption attenuation coefficient.

9. A subway leakage water microelement tracer detection device, characterized in that: The device for realizing the subway leakage water microelement tracer detection method of claim 8 comprises a tracer injection module, a multimodal sampling unit and a data analysis terminal, the tracer injection module comprises a micro pump, a multi-channel grouting pipe, a concentration monitoring sensor and a pressure sensor, the power output end of the micro pump is connected with the main pipe of the multi-channel grouting pipe through a high-pressure pipeline, the branch pipe openings of the grouting pipe correspond to the injection points of the preset suspicious leakage areas, the concentration monitoring sensor and the pressure sensor are respectively installed at the outlet end of the multi-channel grouting pipe to monitor the concentration of the injected tracer in real time; The multimodal sampling unit comprises an automatic sampler, a mobile detection vehicle, a plasma mass spectrometer, a fluorescence spectrometer, an infrared thermal imager, an automatic cruise guide rail and a positioning module, the automatic sampler is installed at a fixed sampling point, the mobile detection vehicle moves along the track in the tunnel, the plasma mass spectrometer, the fluorescence spectrometer, the infrared thermal imager, the automatic cruise guide rail and the positioning module are installed on the mobile detection vehicle, and the data analysis terminal is in communication connection with the automatic sampler, the plasma mass spectrometer, the fluorescence spectrometer, the infrared thermal imager, the automatic cruise guide rail and the positioning module.

10. The subway water leakage microelement tracer detection device according to claim 9, characterized in that: The automatic sampler comprises a water sample storage cabin and a temperature control unit, the water sample collection port of the water sample storage cabin is connected with the tunnel leakage water collection point through a hose, and the temperature control unit is connected with the data analysis terminal to receive a sampling trigger instruction.

Images