Tap water quality pre-detection system

Abstract

The invention discloses a tap water quality pre-detection system, belongs to the technical field of water quality detection, and solves the problems that in the detection process, dead water is not fully discharged before sampling, so that a sample cannot represent daily water, and finally, a water quality safety evaluation result is inaccurate. Comprising a pipe network basic data acquisition module, a dynamic hydraulic characteristic analysis module, a dead water area quantitative correction module, a pollutant long-term enrichment sampling module, a detection data analysis module and a model construction and optimization module. And the pipe network basic data acquisition module is used for acquiring the topological structure of the pipe network through high-precision surveying and mapping. According to the method, the water quality detection accuracy, stability and traceability are improved, a pipe network digital twin model is constructed, a plurality of technologies are utilized to dynamically analyze the water flow state, quantify the volume ratio of a dead water area, correct model parameters and fuse a physical information neural network, and high-precision water quality evaluation and long-term stable monitoring are realized. And water quality detection of water discharged from a community water tank and the like is not needed.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality detection, and particularly to a pre-detection system for tap water quality. Background Art

[0002] The tap water pipe network is an important part of the urban water supply system, mainly responsible for delivering the treated tap water to thousands of households and various water-using units. It usually consists of a series of pipes of different sizes, pumping stations, valves, etc. The water coming out of the water plant first enters the large water conveyance main pipe, and then through the branch pipe networks at all levels, it gradually spreads to all corners of the city. To ensure stable water supply, the pipe network is equipped with booster pumping stations and water towers and other facilities. At the same time, the setting of valves facilitates the maintenance and emergency repair of the pipes. The normal operation of the tap water pipe network plays a crucial role in the economic development, social stability and people's quality of life of the city, and needs to be regularly detected and maintained to ensure its safe and stable water supply.

[0003] The water quality detection of the tap water pipe network is an important link to ensure water supply safety. The detection mainly targets key indicators such as microorganisms, residual chlorine, turbidity, heavy metals, etc. The detection usually samples at positions such as the outlet of the water plant, in the middle of the pipe network and at the end of the pipe network to reflect the water quality conditions of different links. The frequency of regular detection will be determined according to the water supply scale and risk assessment. Generally, large cities will detect daily or weekly. Modern detection mostly uses advanced instruments, such as portable water quality analyzers, etc., which can quickly and accurately obtain data. Once problems are found, relevant departments will quickly take measures, such as flushing the pipe network, adjusting the water plant process, etc., to ensure that the water supply quality meets the standards and guarantee the safety of residents' water use.

[0004] In some communities, due to the inability to achieve direct tap water supply, secondary water supply facilities for community tap water, that is, secondary water supply points such as community water tanks, are often set up in the community. Since the tap water is not directly supplied to households by the water plant in secondary water supply points such as community water tanks, the tap water will be temporarily stored in secondary water supply points such as community water tanks for a period of time. Therefore, it needs to be regularly detected. However, during the detection process, due to the failure to fully discharge the stagnant water before sampling, the sample cannot represent the daily water use, and finally the water quality safety assessment result is inaccurate.

[0005] Therefore, a pre-detection system for tap water quality is proposed to solve or alleviate the above problems. Summary of the Invention

[0006] The purpose of the present invention is to solve the deficiencies in the prior art and propose a pre-detection system for tap water quality.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions: A tap water quality pre - detection system, comprising a pipe network basic data acquisition module, a dynamic hydraulic characteristic analysis module, a dead - water area quantification and correction module, a pollutant long - term enrichment sampling module, a detection data analysis module, and a model construction and optimization module; The pipe network basic data acquisition module is used to obtain the topological structure, node three - dimensional coordinates and pipe physical parameters of the pipe network through high - precision surveying, and construct a digital twin model including resistance characteristics based on pipe network modeling software; The dynamic hydraulic characteristic analysis module is connected to the pipe network basic data acquisition module and is used to monitor the flow data of key nodes in the pipe network in real time, calculate the water flow velocity and Reynolds number to determine the water flow state, and analyze its influence on the mixing ratio of dead water and live water; The dead - water area quantification and correction module is linked with the dynamic hydraulic characteristic analysis module. By injecting tracers into the pipe network and collecting water samples at multiple time periods, fitting the tracer decay curve to determine the proportion of the dead - water area volume, and correcting the parameters of the mixing model; The pollutant long - term enrichment sampling module is deployed at the end of the pipe network and is used to adsorb heavy metals and organic compounds by using a diffusive gradients in thin - films device and a solid - phase microextraction fiber respectively, and obtain the time - weighted average concentration of pollutants by analyzing the adsorption amount; The detection data analysis module receives and integrates the pipe network topology data, hydraulic characteristic data, tracer data and passive sampling data, performs normalization processing and then inputs them into a physics - informed neural network, and conducts model training in combination with the convection - diffusion equation constraint; The model construction and optimization module interacts with the detection data analysis module and uses the Bayesian optimization algorithm to dynamically adjust the dead - water mixing coefficient, minimize the error between the model prediction value and the laboratory test result, generate a three - dimensional pipe network pollutant distribution map, and intuitively display the concentration and migration law of each node.

[0008] Preferably, the pipe network basic data acquisition module is used to obtain the topological structure, node three - dimensional coordinates and pipe physical parameters of the pipe network through high - precision surveying, and construct a digital twin model including resistance characteristics based on pipe network modeling software, and specifically includes the following steps: Measure the three - dimensional coordinates of the pipe network nodes, record the spatial positions including the starting point, end point, elbow, tee and valve of the pipe, and collect the material type, nominal diameter and wall thickness parameters of the pipe; Based on the surveying data, use pipe network modeling software to construct a digital twin model, input the length, elevation difference and local resistance coefficient of each pipe section, generate a three - dimensional pipe network topology map including pipe layout and resistance characteristics, and dynamically adjust the local resistance coefficient of the pipe fittings according to the actual type to accurately simulate the head loss.

[0009] Preferably, the dynamic hydraulic characteristic analysis module is connected to the pipe network basic data acquisition module and is used to monitor the flow data of key nodes in the pipe network in real time, calculate the water flow velocity and Reynolds number to determine the water flow state, and analyze its impact on the mixing ratio of stagnant water and flowing water. The specific steps are as follows: Install non-invasive ultrasonic flowmeters at key nodes in the pipe network, continuously monitor the flow data for 72 hours and generate a minute-level dynamic flow change curve; Calculate the average flow velocity of each pipe section based on the flow data, combine with pipe diameter, fluid density and dynamic viscosity parameters, and distinguish laminar flow and turbulent flow states through Reynolds number calculation and mark high-risk stagnant water areas.

[0010] Preferably, the stagnant water area quantification and correction module is linked with the dynamic hydraulic characteristic analysis module. By injecting tracers into the pipe network and collecting water samples at multiple time periods, fitting the tracer decay curve to determine the proportion of the stagnant water area volume and correct the mixed model parameters. The specific steps are as follows: Inject fluorescent tracers into the pipe network water tank. At multiple preset time periods after injection, collect non-discharged water samples at the end and intermediate nodes of the pipe network, and detect the change of tracer concentration over time; Fit the proportion of the stagnant water area volume and the mixing rate constant through the tracer concentration decay curve, combine with the total volume parameter of the pipe network, correct the volume proportion of the stagnant water area in the dynamic mixing model, and optimize the inversion accuracy of the flowing water pollutant concentration.

[0011] Preferably, the long-term pollutant enrichment sampling module is deployed at the end of the pipe network and is used to adsorb heavy metals and organic substances by using diffusive gradients in thin films (DGT) devices and solid-phase microextraction fibers respectively, and obtain the time-weighted average concentration of pollutants by analyzing the adsorption amount. The specific steps are as follows: Fix DGT devices in the straight pipe section at the end of the pipe network, continuously expose for 7 days to adsorb heavy metal ions, analyze the resin membrane by nitric acid solution and detect the adsorption amount, and calculate the time-weighted average concentration of heavy metals in the water body; Deploy solid-phase microextraction fibers at the user-end faucet, make it contact the water flow at normal household flow rate for 30 minutes, adsorb organic substances and then analyze by thermal desorption and gas chromatography-mass spectrometry (GC-MS) to estimate the long-term average concentration of aqueous organic substances.

[0012] Preferably, the detection data analysis module receives and integrates pipe network topology data, hydraulic characteristic data, tracer data and passive sampling data, performs normalization processing and then inputs them into a physics-informed neural network, and combines the convection-diffusion equation constraints for model training. The specific steps are as follows: Integrate pipe network topology data, dynamic flow data, tracer data and passive sampling data, and perform normalization processing to unify the data format; Input the processed data into the physics-informed neural network. Combining the convection, diffusion, and adsorption laws of pollutants in the pipe network, train the model through physical equation constraints to predict the pollutant concentration distribution at each node in the pipe network.

[0013] Preferably, the model construction and optimization module interacts with the detection data analysis module and uses the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, minimize the error between the model prediction value and the laboratory test result, and generate a three-dimensional pipe network pollutant distribution map to intuitively display the concentration and migration law of each node, specifically including the following steps: Use the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, and optimize the parameter configuration of the mixing model by iteratively minimizing the error between the model prediction value and the laboratory standard test result; Based on the optimized model, combine the pipe network hydraulic conditions and pollutant migration characteristics to generate a three-dimensional pollutant distribution map to intuitively display the concentration gradient of each node and the pollutant diffusion path in the pipe network.

[0014] The present invention has the following beneficial effects: The present invention can significantly improve the accuracy of water quality detection, long-term monitoring stability, and pollution source tracing ability, construct a digital twin model of the pipe network, accurately restore the pipeline layout, resistance characteristics, and three-dimensional spatial relationship, provide a reliable structural basis for dynamic analysis, real-time monitor the flow data through an ultrasonic flowmeter and calculate the Reynolds number, dynamically identify the water flow state, combine the tracer decay experiment to quantify the proportion of the dead water area volume, correct the parameters of the mixing model, and eliminate the underestimation or overestimation error of pollutant concentration caused by water flow stagnation. The DGT and SPME passive sampling technologies capture the time-weighted average concentration of heavy metals and organic compounds through long-term adsorption, avoid the influence of instantaneous sampling by water flow fluctuations and uneven pollutant distribution, and truly reflect the long-term trend of water quality. The physics-informed neural network integrates the pipe network topology, hydraulic parameters, and pollutant migration laws, uses Bayesian optimization to dynamically adjust the dead water mixing coefficient, minimizes the deviation between the model prediction value and the laboratory test, and the generated three-dimensional pollutant heat map can accurately locate the pollution hot spots and predict the pollution development trend, realizing the full-chain linkage from pipe network structure analysis to pollutant spatio-temporal distribution inversion, and finally achieving high-precision water quality assessment and long-term stable monitoring, so as to realize the detection of tap water quality without draining water from secondary water supply points such as community water tanks. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 This is a schematic structural diagram of the present invention.

[0017] In the figure: 1 is the pipeline network basic data acquisition module; 2 is the dynamic hydraulic characteristic analysis module; 3 is the dead water area quantification and correction module; 4 is the long-term pollutant enrichment sampling module; 5 is the detection data analysis module; 6 is the model construction and optimization module. Specific embodiments

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention described and illustrated in the drawings here can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0020] It should be noted that like reference numerals and letters denote like items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0021] In the description of the present invention, it should be understood that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the present invention is usually placed during use, or the orientation or positional relationship commonly understood by those skilled in the art. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0022] In addition, the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0023] In the description of the present invention, it should also be noted that, unless otherwise clearly specified and defined, the terms "set", "install", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0024] A tap water quality pre-detection system, as Figure 1 shown, includes a pipe network basic data acquisition module 1, a dynamic hydraulic characteristic analysis module 2, a dead water area quantification and correction module 3, a long-term pollutant enrichment sampling module 4, a detection data analysis module 5, and a model construction and optimization module 6; Among them, the pipe network basic data acquisition module 1 is used to obtain the topological structure of the pipe network, the three-dimensional coordinates of the nodes, and the physical parameters of the pipes through high-precision surveying and mapping, and build a digital twin model including resistance characteristics based on pipe network modeling software; Specifically: Measure the three-dimensional coordinates of the pipe network nodes, record the spatial positions including the starting point, ending point, elbow, tee, and valve of the pipe, and collect the material type, nominal diameter, and wall thickness parameters of the pipe; Based on the surveying and mapping data, use pipe network modeling software to build a digital twin model, input the length, elevation difference, and local resistance coefficient of each pipe section, and generate a three-dimensional pipe network topology map including the pipe layout and resistance characteristics, where the local resistance coefficient of the pipe fittings is dynamically adjusted according to the actual type to accurately simulate the head loss; Among them, the dynamic hydraulic characteristic analysis module 2 is connected to the pipe network basic data acquisition module 1 and is used to monitor the flow data of the key nodes of the pipe network in real time, calculate the water flow velocity and Reynolds number to determine the water flow state, and analyze its influence on the mixing ratio of dead water and live water; Specifically: Install a non-intrusive ultrasonic flowmeter at the key nodes of the pipe network, continuously monitor the flow data for 72 hours and generate a minute-level dynamic flow change curve; Calculate the average flow velocity of each pipe section according to the flow data, combine the pipe diameter, fluid density, and dynamic viscosity parameters, and distinguish the laminar flow and turbulent flow states through Reynolds number calculation and mark the high-risk areas of dead water.

[0025] Among them, the dead water area quantification and correction module 3 is linked with the dynamic hydraulic characteristic analysis module 2. By injecting a tracer into the pipe network and collecting water samples at multiple time intervals, fit the tracer decay curve to determine the proportion of the dead water area volume and correct the parameters of the mixing model; Specifically: Inject a fluorescent tracer into the pipe network water tank. During multiple preset time periods after injection, collect non-discharged water samples at the end and intermediate nodes of the pipe network, and detect the change of tracer concentration over time; Fit the proportion of the dead water area and the mixing rate constant through the tracer concentration decay curve, and combine with the total volume parameter of the pipe network to correct the volume proportion of the dead water area in the dynamic mixing model and optimize the inversion accuracy of the live water pollutant concentration.

[0026] Among them, the long-term pollutant enrichment sampling module 4 is deployed at the end of the pipe network and is used to adsorb heavy metals and organic matter by using a diffusive gradients in thin films device and a solid-phase microextraction fiber respectively, and obtain the time-weighted average concentration of pollutants by analyzing the adsorption amount; Specifically: Fix the diffusive gradients in thin films device at the straight pipe section at the end of the pipe network, continuously expose it for 7 days to adsorb heavy metal ions, analyze the resin membrane with nitric acid solution and detect the adsorption amount, and calculate the time-weighted average concentration of heavy metals in the water body; Deploy a solid-phase microextraction fiber at the user-end faucet, make it contact the water flow at the normal household flow rate for 30 minutes, adsorb organic matter, and then analyze it by thermal desorption and gas chromatography-mass spectrometry to estimate the long-term average concentration of aqueous organic matter.

[0027] Among them, the detection data analysis module 5 receives and integrates the pipe network topology data, hydraulic characteristic data, tracer data and passive sampling data, performs normalization processing and then inputs it into the physics-informed neural network, and combines the convective-diffusion equation constraint for model training; Specifically: Integrate the pipe network topology data, dynamic flow data, tracer data and passive sampling data, and perform normalization processing to unify the data format; Input the processed data into the physics-informed neural network, combine the convection, diffusion and adsorption laws of pollutants in the pipe network, and train the model through physical equation constraints to predict the pollutant concentration distribution at each node of the pipe network.

[0028] Among them, the model construction and optimization module 6 interacts with the detection data analysis module 5 and uses the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, minimize the error between the model prediction value and the laboratory test result, generate a three-dimensional pipe network pollutant distribution map, and intuitively display the concentration and migration law of each node.

[0029] Specifically: Use the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, and minimize the error between the model prediction value and the laboratory standard test result through iteration to optimize the parameter configuration of the mixing model; Based on the optimized model, combine the pipe network hydraulic conditions and pollutant migration characteristics to generate a three-dimensional pollutant distribution map, and intuitively display the concentration gradient and pollutant diffusion path of each node in the pipe network.

[0030] When the system is working, first, a RTK surveying instrument is used to measure the three-dimensional coordinates of the nodes in the pipe network with centimeter-level accuracy, including the spatial coordinates of the key positions such as the starting point, ending point, elbow, tee, and valve of the pipeline. At the same time, the material type, nominal diameter, and wall thickness parameters of the pipeline are recorded. These data are input through pipe network modeling software, such as EPANET or WaterGEMS, and combined with the pipe section length, elevation difference, and local resistance coefficient. For example, the valve resistance coefficient is set to 16 - 20, and the elbow resistance coefficient is set to 0.75, to construct a digital twin model that includes the pipeline layout, resistance characteristics, and hydraulic parameters, providing a basic framework for subsequent analysis.

[0031] Subsequently, the dynamic hydraulic characteristics analysis module 2 deploys non-intrusive ultrasonic flowmeters at the key nodes of the pipe network, specifically at the positions of the main water supply pipe and the user inlet pipe. The flow velocity is monitored in real time through the time difference method principle, and the instantaneous flow rate data within 72 hours are continuously recorded to generate a water consumption change curve. Based on the flow rate data, the average flow velocity and Reynolds number are calculated to determine the flow state as laminar flow, turbulent flow, or transitional flow, and the mixing ratio of dead water and live water under different flow states is analyzed. For example, in the turbulent flow state, the pollutants in the dead water area diffuse faster due to the intense mixing of the water flow, while in the laminar flow state, the pollutants in the dead water area stay longer. These analysis results provide support for the dynamic hydraulic conditions of the subsequent model.

[0032] Next, the dead water area quantification and calibration module 3 precisely calibrates the dead water area in the pipe network through a tracer experiment. The system injects rhodamine WT fluorescent tracer into the secondary water supply tank, and the initial concentration is set to 1 ppb. Subsequently, at multiple preset time periods after injection, such as 0 minutes, 10 minutes, 30 minutes, 1 hour, 2 hours, 4 hours, 8 hours, 12 hours, 24 hours, etc., water samples that have not been discharged are collected at the end and intermediate nodes of the pipe network, and the change in tracer concentration is detected using a fluorescence spectrophotometer. By fitting the concentration decay curve and combining the exponential decay model, the volume ratio of the dead water area and the mixing rate constant are calculated. If the tracer concentration drops to 30% of the initial value within 24 hours, it indicates that the volume of the dead water area accounts for 40% of the total volume of the pipe network. These parameters are absorbed by the dynamic mixing model to correct the back-calculation algorithm for the concentration of live water pollutants, ensuring that the subsequent detection results are closer to the true values.

[0033] Meanwhile, the long-term pollutant enrichment sampling module 4 installs a diffusive gradients in thin films device at the straight pipe section at the end of the pipe network. The Chelex-100 resin membrane inside adsorbs heavy metal ions in the water body through Fick's diffusion law. After the diffusive gradients in thin films device is continuously exposed for 7 days, the Chelex-100 resin membrane is taken out, and the adsorbed heavy metals are desorbed with nitric acid solution. The metal concentration in the desorbed solution is detected by an inductively coupled plasma mass spectrometer, and the time-weighted average concentration of heavy metals in the water body is calculated by combining the exposure time and the diffusion coefficient.

[0034] At the user - end faucet, the solid - phase microextraction fiber adsorbs organic substances with a polydimethylsiloxane coating. The solid - phase microextraction fiber contacts the water flow for 30 minutes at a domestic water flow rate of 0.5 - 1 L / min. Subsequently, the adsorbed organic substances are released into a gas chromatography - mass spectrometry instrument for analysis by thermal desorption. The long - term average concentration of aqueous organic substances is deduced based on the distribution coefficient, effectively avoiding the contingency of instantaneous sampling.

[0035] The detection data analysis module 5 receives the original data from each module, including the pipe network topology, dynamic flow monitoring results, tracer decay curves, and detection values of DGT and SPME. It normalizes this multi - source data. Specifically, it converts the flow data into a unified time - series format and standardizes the concentration data to milligrams per liter. The processed data is input into a physics - informed neural network. The network architecture includes 3 - 4 hidden layers, with 64 - 128 neurons in each layer. The input parameters cover pipe network node coordinates, time, flow velocity, and pollutant types. The output is the predicted pollutant concentration at each node. During the training process, the neural network not only minimizes the error between the predicted value and the measured data but also embeds physical constraints of the convection - diffusion equation, such as the relationship between the rate of pollutant migration with the water flow and the diffusion intensity, ensuring that the model conforms to both data patterns and physical laws.

[0036] The model construction and optimization module 6 further dynamically adjusts the dead - water mixing coefficient through the Bayesian optimization algorithm. For example, the mixing coefficient is initially set to 0.5, and by iteratively calculating the root - mean - square error between the model predicted value and the laboratory standard test results, the coefficient is gradually optimized to 0.35, making the model more suitable for the actual pipe network conditions.

[0037] Finally, the system generates a three - dimensional map of the pollutant distribution in the pipe network, visually showing the concentration gradients of heavy metals and organic substances at each node in the form of a heat map, and marking the main diffusion paths of pollutants, such as accumulating from the water supply main artery to the end. At the same time, it combines historical data to predict the pollution trend.

[0038] The pipe network model of the basic data acquisition module provides structural support for hydraulic analysis. The flow data of the hydraulic characteristics analysis module drives the calibration process of the tracer experiment. The calibrated dead - water parameters optimize the interpretation accuracy of passive sampling data. The detection data feeds back to optimize the model after neural network training, ensuring that the detection results truly reflect the daily water quality and enhancing the reliability of the secondary water supply safety assessment from the root.

[0039] The above - mentioned are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A tap water quality pre - detection system, characterized in that, It includes a pipeline network basic data acquisition module (1), a dynamic hydraulic characteristic analysis module (2), a dead water area quantification and correction module (3), a long-term pollutant enrichment sampling module (4), a detection data analysis module (5), and a model construction and optimization module (6); The pipeline network basic data acquisition module (1) is used to obtain the topological structure, node three-dimensional coordinates and pipeline physical parameters of the pipeline network through high-precision surveying, and construct a digital twin model including resistance characteristics based on pipeline network modeling software; The dynamic hydraulic characteristic analysis module (2) is connected to the pipeline network basic data acquisition module (1) and is used to monitor the flow data of key nodes of the pipeline network in real time, calculate the water flow velocity and Reynolds number to determine the water flow state, and analyze its influence on the mixing ratio of dead water and live water; The dead water area quantification and correction module (3) is linked with the dynamic hydraulic characteristic analysis module (2). By injecting a tracer into the pipeline network and collecting water samples at multiple time intervals, it fits the tracer decay curve to determine the proportion of the dead water area volume and corrects the mixing model parameters; The long-term pollutant enrichment sampling module (4) is deployed at the end of the pipeline network and is used to adsorb heavy metals and organic substances by using a diffusive gradient in thin films device and a solid-phase microextraction fiber respectively, and obtain the time-weighted average concentration of pollutants by analyzing the adsorption amount; The detection data analysis module (5) receives and integrates the pipeline network topology data, hydraulic characteristic data, tracer data and passive sampling data, performs normalization processing and then inputs it into a physics-informed neural network, and conducts model training in combination with the convection-diffusion equation constraint; The model construction and optimization module (6) interacts with the detection data analysis module (5) and uses the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, minimize the error between the model prediction value and the laboratory test result, generate a three-dimensional pipeline network pollutant distribution map, and intuitively display the concentration and migration law of each node; 2. The tap water quality pre - detection system according to claim 1, characterized in that, The pipeline network basic data acquisition module (1) is used to obtain the topological structure, node three-dimensional coordinates and pipeline physical parameters of the pipeline network through high-precision surveying, and construct a digital twin model including resistance characteristics based on pipeline network modeling software. The specific steps are as follows: Measure the three-dimensional coordinates of the pipeline network nodes, record the spatial positions including the starting point, ending point, elbow, tee and valve of the pipeline, and collect the material type, nominal diameter and wall thickness parameters of the pipeline; Based on the surveying data, use pipeline network modeling software to construct a digital twin model, input the length, elevation difference and local resistance coefficient of each pipe section, and generate a three-dimensional pipeline network topology map including pipeline layout and resistance characteristics. The local resistance coefficient of the pipe fittings is dynamically adjusted according to the actual type to accurately simulate the head loss; 3. The tap water quality pre - detection system according to claim 1, characterized in that, The dynamic hydraulic characteristic analysis module (2) is connected to the pipeline network basic data acquisition module (1) and is used to monitor the flow data of key nodes of the pipeline network in real time, calculate the water flow velocity and Reynolds number to determine the water flow state, and analyze its influence on the mixing ratio of dead water and live water. The specific steps are as follows: Install a non-intrusive ultrasonic flowmeter at the key nodes of the pipeline network, continuously monitor the flow data for 72 hours and generate a minute-level dynamic flow change curve; Calculate the average flow velocity of each pipe section according to the flow data, combine the pipe diameter, fluid density and dynamic viscosity parameters, and distinguish the laminar flow and turbulent flow states through the Reynolds number calculation and mark the high-risk dead water areas.

4. The tap water quality pre - detection system according to claim 1, characterized in that, The dead water area quantification and correction module (3) is linked with the dynamic hydraulic characteristics analysis module (2). By injecting tracers into the pipe network and collecting water samples at multiple time periods, fitting the tracer decay curve to determine the proportion of the dead water area volume, and correcting the parameters of the mixing model. The specific steps are as follows: Inject fluorescent tracers into the pipe network water tank. At multiple preset time periods after injection, collect the undischarged water samples at the end and intermediate nodes of the pipe network, and detect the change of tracer concentration over time. Fit the proportion of the dead water area volume and the mixing rate constant through the tracer concentration decay curve. Combine the total volume parameter of the pipe network to correct the volume proportion of the dead water area in the dynamic mixing model and optimize the inversion accuracy of the concentration of the living water pollutants.

5. The tap water quality pre - detection system according to claim 1, characterized in that, The long-term pollutant enrichment sampling module (4) is deployed at the end of the pipe network and is used to adsorb heavy metals and organic matter by using diffusive gradients in thin-films devices and solid-phase microextraction fibers respectively, and obtain the time-weighted average concentration of pollutants by analyzing the adsorption amount. The specific steps are as follows: Fix the diffusive gradients in thin-films device at the straight pipe section at the end of the pipe network, continuously expose it for 7 days to adsorb heavy metal ions, analyze the resin membrane with nitric acid solution and detect the adsorption amount, and calculate the time-weighted average concentration of heavy metals in the water body. Deploy solid-phase microextraction fibers at the user-end faucet, make it contact the water flow at the normal household flow rate for 30 minutes, adsorb organic matter, and then analyze it by thermal desorption and gas chromatography-mass spectrometry to estimate the long-term average concentration of aqueous organic matter.

6. The tap water quality pre - detection system according to claim 1, characterized in that, The detection data analysis module (5) receives and integrates the pipe network topology data, hydraulic characteristics data, tracer data and passive sampling data, performs normalization processing and then inputs it into the physics-informed neural network, and combines the convection-diffusion equation constraint for model training. The specific steps are as follows: Integrate the pipe network topology data, dynamic flow data, tracer data and passive sampling data, and perform normalization processing to unify the data format. Input the processed data into the physics-informed neural network, combine the convection, diffusion and adsorption laws of pollutants in the pipe network, and train the model through physical equation constraints to predict the pollutant concentration distribution at each node of the pipe network.

7. The tap water quality pre - detection system according to claim 1, characterized in that, The model construction and optimization module (6) interacts with the detection data analysis module (5) and uses the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, minimize the error between the model prediction value and the laboratory test result, and generate a three-dimensional pipe network pollutant distribution map to visually display the concentration and migration law of each node. The specific steps are as follows: Use the Bayesian optimization algorithm to dynamically adjust the dead water mixing coefficient, and minimize the error between the model prediction value and the laboratory standard test result through iteration to optimize the parameter configuration of the mixing model. Based on the optimized model, combine the pipe network hydraulic conditions and pollutant migration characteristics to generate a three-dimensional pollutant distribution map to visually display the concentration gradient and pollutant diffusion path of each node in the pipe network.

Images