Intelligent water-yellowing and double-odor purification treatment equipment for water supply

Abstract

The invention discloses intelligent water yellowing and odor purification treatment equipment for water supply, and relates to the technical field of water dispatching purification control. Comprising a real-time acquisition unit, an intelligent analysis unit, an execution control unit and a verification evaluation unit, water quality data is acquired in real time through a multi-area sensing terminal, and water quality changes are dynamically sensed by combining source tracing and flow direction modeling; the intelligent analysis unit adopts a fusion discrimination model to identify a water yellow phenomenon and'two-odor 'pollution; the execution control unit links a valve to switch a water path and start purification equipment according to the pollution source positioning and diffusion trend, so as to realize rapid blocking and treatment; and the verification and evaluation unit forms a closed-loop verification and continuous optimization mechanism through online mass spectrum monitoring and biosensing. Real-time response, self-adaptive scheduling and online standard reaching can be achieved when the water quality is abnormal, the energy saving performance, safety and user perception quality of a water supply system are improved, and the method has the advantages of being scientific in point distribution, accurate in recognition, rapid in response and efficient in processing.

Description

Technical Field

[0001] The present invention relates to the technical field of water supply scheduling and purification control, and in particular to intelligent water supply yellowing and two odor purification treatment and disposal equipment. Background Art

[0002] In urban water supply systems, water quality safety and energy-saving scheduling have always been core technical challenges. The traditional water supply model is mainly based on centralized treatment at the source water plant. It lacks a real-time perception and response mechanism for dynamic changes in water quality in the pipe network. As a result, once pollution or abnormal water quality occurs, the response is often delayed and coverage is insufficient. Especially under the influence of complex environmental factors, meteorological changes (such as heavy rain, precipitation runoff); abnormal municipal facilities (such as abnormal opening of manhole covers, construction disturbances); water transfer operations (inter-regional water transfer or water source switching) and other situations can easily cause local water quality problems. Typical phenomena include: yellow water phenomenon: turbidity and increased color of the water body, affecting sensory quality; two odor problems: the odor perception caused by geosmin (GSM) and 2-Methylisoborneol (2-MIB), although the impact on health is limited, it is very likely to cause user complaints and trust crises.

[0003] The "two odors" pollution is difficult to remove through traditional disinfection and can cause strong odor perception even at low concentrations, which puts higher demands on the real-time monitoring and emergency response capabilities of the water supply system.

[0004] The following technical deficiencies are common in existing water supply systems: insufficient water quality perception makes it difficult to detect local anomalies in a timely manner; valves and purification equipment deployment lack intelligent linkage and have slow response speeds; traditional compliance evaluation mainly relies on offline manual testing and cannot achieve an online compliance closed loop.

[0005] Therefore, there is an urgent need for a new type of intelligent control equipment that can deploy water quality monitoring and intelligent valve control equipment in sections in the pipeline network, combine with multi-source data perception such as meteorology and manhole covers, and automatically switch to purification equipment for real-time processing and compliance evaluation when "yellow water", "two odors" and water quality abnormalities caused by water diversion occur, so as to realize full-chain perception, analysis and adaptive scheduling of the water supply process. Summary of the Invention

[0006] The purpose of the present invention is to solve the problems of insufficient water quality perception, slow response speed and inability to achieve online closed-loop compliance, and to propose an intelligent water supply yellowing and two odor purification and treatment equipment.

[0007] The objectives of the present invention can be achieved through the following technical solutions: an intelligent water yellowing and two odor purification treatment and disposal equipment for water supply, comprising: a real-time collection unit, an intelligent analysis unit and an execution control unit. The real-time collection unit collects water quality data in real time through sensor terminals deployed at water source inlets, key nodes of branch pipe networks and building terminals; the intelligent analysis unit dynamically analyzes water quality data through a data middle platform and a cloud platform, identifies abnormal yellowing of water and pollution of two odorous substances, and locates the source of pollution; the execution control unit regulates the water supply path, valve status and emergency purification equipment based on the analysis results.

[0008] As a preferred embodiment of the present invention, it also includes a verification and evaluation unit, which compares the compliance of several water quality indicators with national standards in real time; provides toxicity warnings through zebrafish or cell metabolism reactors; and links execution results with monitoring data to optimize analysis models and equipment layout.

[0009] As a preferred embodiment of the present invention, the real-time acquisition unit includes:

[0010] Water source tracing, based on the pipe network topology modeling as a directed graph, calculates the water source probability vector of each node through flow weight and generates labels;

[0011] Sensor deployment optimization calculates deployment priority based on node response capability, number of covered paths, and blind spot risk, and dynamically adjusts sensor locations.

[0012] As a preferred embodiment of the present invention, the water source tracing method calculates the path weight by the following formula:

[0013]

[0014] in is the instantaneous flow on the i-th edge in the path, Out(u i ) is the node v i all downstream connection edges; and construct a "source-terminal" impact matrix for correlation analysis to generate a water source tracing log.

[0015] As a preferred embodiment of the present invention, the intelligent analysis unit includes:

[0016] To identify the yellow characteristics of water, a random forest classification model is constructed based on chromaticity, turbidity, and environmental factors, and the alarm threshold is dynamically adjusted;

[0017] For the two odor detection, the SVM or CNN model is used to perform pattern recognition on the calibrated odor substance concentrations, and the convection-diffusion equation is combined to predict the pollution spread.

[0018] As a preferred embodiment of the present invention, the specific process of identifying the yellowing characteristics of water is as follows:

[0019] S11: Use random forest classifier to identify yellow water anomalies; the training data consists of normal and abnormal water quality samples, and the output category label y∈{0,1}:

[0020] When y=0, it is normal water quality; when y=1, it is abnormal yellow water quality;

[0021] Random forest outputs the discrimination results

[0022] S21: Introduce the environmental impact correction factor λ to dynamically adjust the alarm threshold:

[0023] λ=f(R,G)

[0024] Where R is the current rainfall; G is the manhole cover opening and closing abnormality index; f(·) is the empirical regression function;

[0025] Final alarm threshold:

[0026] T′=T0×(1+λ)

[0027] When the water yellowness score S satisfies S>T′, it is judged as abnormal and triggers subsequent tracing processing.

[0028] As a preferred embodiment of the present invention, the specific process of the two odor detection modules is as follows:

[0029] S21: The sensor outputs the original concentration data C raw , perform background noise calibration and obtain the calibration concentration C calC :

[0030] C cal =C raw -C baseline

[0031] Where: C baseline is the environmental baseline noise concentration for system self-learning;

[0032] S22: Use support vector machine (SVM) to build a low-concentration odor recognition model, and the decision function is:

[0033] f(X')=sign(w·X'+b)

[0034] Where: X′ is the normalized feature vector, w is the weight vector; b is the bias term;

[0035] When the model output is f(X′)=+1, there is odor pollution; when the model output is f(X′)=-1, there is no obvious odor;

[0036] S23: Combined with the water velocity v1, diffusion coefficient D1, and time t1, a prediction model for the spread of two odorous substances is established, using a one-dimensional simplified convection-diffusion equation:

[0037]

[0038] Where: C1 is the concentration of odorous substances, x1 is the spatial position, t1 is the time, v1 is the water flow rate, and D1 is the diffusion coefficient; the pollutant diffusion range and arrival time are predicted, and the pollution source is located in combination with the pipeline network GIS data.

[0039] As a preferred embodiment of the present invention, the execution control unit includes:

[0040] Waterway switching: closing valves or switching to backup water sources based on estimated pollution diffusion time;

[0041] Purification control: when exceeding the standard is detected, emergency treatment will be initiated until the effluent meets the chromaticity <15 degrees and GSM / 2-MIB <5ng / L.

[0042] As a preferred embodiment of the present invention, the backup water source scheduling is selected by feasibility scoring:

[0043] S31: Rapidly locate pollution sources by inputting node water quality sampling concentration and GIS pipe network flow data, and calling the pollution diffusion model to determine whether the change in pollutant concentration gradient conforms to the diffusion characteristics of the upstream source;

[0044] If it matches, mark the suspected pollution source node;

[0045] If it does not meet the requirements, expand the detection range and reposition until the suspected pollution source node is marked;

[0046] S32: Scheduling resource optimization and response prediction, based on the location of the pollution source and the backup water source list R k , importance weight W m , conduct backup water supply feasibility scoring U k :

[0047]

[0048] Among them, Q k is the water supply capacity of the backup water source k.

[0049] As a preferred embodiment of the present invention, the closed-loop optimization evaluates water quality compliance through the following formula:

[0050] Deploy mass spectrometers at city outlets or factory outlets to conduct comprehensive water quality testing: 106 indicators are compared:

[0051]

[0052] Where: X j1 is the actual value of the j1th detection index; X j1,GB5749 This is the national standard limit; if all indicators meet the standards, normal water supply will be restored; if there are any items exceeding the limit, the emergency treatment status will continue to be maintained.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. By deploying high-frequency real-time sensing terminals at multiple levels, including the source, branch pipe network, and building terminals, and combining them with a water source tracing algorithm, this invention can achieve global perception of dynamic changes in water quality and source-related determination, effectively improving the speed and accuracy of detecting local anomalies.

[0055] 2. After the intelligent analysis unit identifies key water quality anomalies such as yellow water and two odors, the present invention automatically links the execution control unit to achieve water supply path switching, rapid valve sealing and emergency purification treatment, significantly shortening the response time, inhibiting the spread of pollution, and ensuring the continuity and safety of water supply.

[0056] 3. This invention introduces a chromaticity-turbidity-environmental factor fusion recognition model and a low-concentration odor detection model (SVM / CNN), combined with hydraulic diffusion prediction, to achieve early warning of abnormal events and rapid location of pollution sources. It is particularly suitable for dealing with low-concentration odor problems that are difficult to identify with traditional methods.

[0057] 4. Based on the GIS pipe network model and pollution diffusion simulation, the present invention intelligently optimizes the scheduling of backup water sources, automatically switches the water supply path, and activates the local purification unit to ensure that the polluted area can quickly restore qualified water supply, thereby improving the system's adaptive scheduling and emergency response capabilities.

[0058] 5. The present invention uses terminal verification devices such as mass spectrometers and toxic biosensors to combine monitoring data with response actions to form a data closed loop, support self-correction and threshold optimization of the analysis model, promote continuous iterative improvement of system performance in actual operation, and truly realize online closed-loop control of water supply compliance. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0060] Figure 1 This is a block diagram of the device principle of the present invention. DETAILED DESCRIPTION

[0061] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0062] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0063] It should also be understood that the terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" as used in this disclosure and the claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0064] See also Figure 1 As shown, an intelligent water yellowing and two odor purification and treatment equipment for water supply includes: a real-time collection unit, an intelligent analysis unit, an execution control unit and a verification and evaluation unit;

[0065] The real-time collection unit achieves global perception and high-frequency sampling of water quality data through sensor terminals deployed in different areas. The main deployment locations include:

[0066] Source monitoring area: highly sensitive multi-parameter sensing equipment is deployed at water intake points and water source entrances to collect raw water quality indicators;

[0067] Branch network area: Deploy sensor terminals at key nodes and branch nodes in the urban water supply network to achieve dynamic monitoring of regional water quality changes;

[0068] Building terminal area: For high-rise buildings or multi-story residential areas, micro-sensor modules are deployed at the end water outlet to understand the water quality conditions on the user side;

[0069] In order to make the sensor terminal deployment location more efficient and accurate, water source tracing is used to obtain the collected water source data of the corresponding sensor;

[0070] After data collection, it is first pre-processed by the local microprocessor (such as filtering, outlier removal, standardization, etc.) and transmitted to the analysis center through the edge gateway;

[0071] The intelligent analysis unit is a joint analysis center formed by the data center and the cloud platform. The analysis center formed by the data center and the cloud platform has set up special identification and processing mechanisms specifically for the yellowing of water and two odor pollution (geosmin GSM and 2-methylisoborneol 2-MIB), including:

[0072] Water yellow feature recognition and analysis: By building a fusion model, it comprehensively determines whether the water body has abnormal color or an increasing trend of turbidity, and issues real-time alarms and records the occurrence points and diffusion trends of abnormal events;

[0073] Two-odor (GSM and 2-MIB) detection and analysis uses a highly sensitive sensing module to obtain changes in the concentration of odorous substances in water samples; uses a support vector machine (SVM) or convolutional neural network (CNN) model to perform pattern recognition on low-concentration odorous substances to improve detection sensitivity; analyzes the dynamic change curve of substance concentration and combines information such as temperature, water flow rate, and changes in water diversion paths to establish an odorous substance propagation model;

[0074] After discovering an anomaly, the pollution source is quickly located based on spatial distribution information and flow direction models. Combining the network's Geographic Information System (GIS) with floor layout information, it optimizes the allocation of dispatch resources and predicts responses for different areas (e.g., regional zones and building floors).

[0075] The execution control unit is deployed at key physical control nodes and implements regulatory responses through execution components; it automatically adjusts the water supply path and flow rate according to water quality abnormalities to control the spread of pollutants; in areas with high pollution risk or overflow risk, the control valve automatically closes or directs it to the emergency treatment unit; smart homes with backup household water treatment modules can perform local water purification or switch to backup water sources.

[0076] The verification and assessment unit implements a closed-loop verification mechanism for water supply quality:

[0077] Install online mass spectrometers at major water outlets (including plant outlets and major urban discharge outlets) to monitor in real time the 106 indicators specified in the "Sanitary Standard for Drinking Water" (GB5749-2022);

[0078] Integrate bio-responsive sensing modules (such as zebrafish acute toxicity detection and cell metabolism reactors) to provide biological early warning of toxic abnormalities;

[0079] Compare monitoring data with execution response results to form a closed data loop, thereby continuously optimizing analysis models, threshold strategies, and equipment layout;

[0080] The water source data collected by the corresponding sensor is obtained as follows:

[0081] First, using the pipe network topology of the urban water supply network, the urban water supply system is modeled as a directed graph G = (V, E), where V represents each node (such as water intake points, branch nodes, pumping stations, and building terminals), and E represents the pipeline edges connected by water flow. Each edge is accompanied by weight information, such as pipe diameter d, flow rate Q, and water pressure P.

[0082] Combined with the connections between each water intake point and the end user, from any sensor terminal node v end Trace upstream to all possible water sources src , and record each path The path weight (representing the probability of water supply) is calculated by the formula:

[0083]

[0084] in is the instantaneous flow on the i-th edge in the path, Out(u i ) is the node v i All downstream connecting edges of . The water source probability vector is obtained by normalization:

[0085]

[0086] Indicates the proportion of water received by the terminal point that comes from different sources;

[0087] Using information such as flow meters, valve status, and scheduling records, the flow Q of each edge is updated in real time to correct the traceability path weight. Scheduling events (such as pump station start and stop) will trigger real-time graph structure adjustments and automatically update path calculation results.

[0088] Furthermore, through the correlation analysis between the source water quality and the downstream node water quality: for the tracing results of multiple terminal points, the "source-terminal" impact matrix M is constructed, and its elements M ij Indicates the impact of source i on terminal j; combined with the change in water quality index ΔC j , perform Pearson correlation analysis or Granger causality test to identify the source of pollution changes:

[0089]

[0090] Each piece of sensor data is labeled with its corresponding water source label (e.g., "water source A - branch 3 - terminal B") and a water source traceability log is generated. This structured data is fed into the intelligent analysis unit.

[0091] The tracing results and node sensitivity (such as pollution response time, impact range, etc.) are used to calculate the priority of each area, which is in the form of a layout optimization function:

[0092] U(v)=α×R(v)+β×C(v)+γ×D(v)

[0093] Where R(v) is the response capability, which is used to measure the response timeliness and recognition sensitivity of node v to water quality fluctuation events;

[0094]

[0095] T delay (v) is the delay from the occurrence of the upstream water quality event to the observation of the fluctuation at this node;

[0096] SNR(v) is the signal-to-noise ratio of the signal collected by the node, which is used to reflect the sampling quality;

[0097] Where C(v) is the number of covered paths, which is used to quantify the “network influence breadth” of the node’s location, that is, how many different water source paths the node is located at the intersection of or downstream of;

[0098] C(v)=|{P i |v∈P i}|

[0099] P i represents the i-th water source path;

[0100] If a node is at the intersection or branch point of multiple paths, its value will be higher, indicating that the node layout is globally representative;

[0101] Where D(v) is the current blind spot risk, which is used to evaluate the potential safety hazard if there is no sensor monitoring at this location;

[0102]

[0103] L(v) is the shortest distance to the node where sensors are deployed (the larger the distance is, the greater the distance is without sensor);

[0104] Var(C hist (v)) is the fluctuation variance of historical water quality data (the larger the value, the higher the instability);

[0105] Calculate the preferred values ​​U(v) of all candidate nodes based on the above indicators, sort them from high to low by score, and select the top K as recommended deployment points for staff to deploy;

[0106] The specific process of water yellow feature recognition analysis is as follows: (Building a fusion discrimination model based on color, turbidity, suspended solids concentration (TSS) and environmental factors (such as rainfall, manhole cover switch status))

[0107] S11: Use the Random Forest classifier to identify abnormal yellow water quality. The training data consists of normal and abnormal water quality samples. The model outputs the category label y∈{0,1}:

[0108] When y=0, it is normal water quality; when y=1, it is abnormal yellow water quality;

[0109] The formula for random forest output discrimination results is as follows:

[0110]

[0111] where h i2 (·) is the i2th decision tree classifier; mode represents the mode of the prediction results of each tree.

[0112] S21: Introduce the environmental impact correction factor λ to dynamically adjust the alarm threshold:

[0113] λ=f(R,G)

[0114] Where R is the current rainfall; G is the manhole cover opening abnormality index (abnormal opening count / normal period); f(·) is the empirical regression function or the weight adjustment function obtained by fitting based on historical data;

[0115] Final alarm threshold:

[0116] T′=T0×(1+λ)

[0117] When the water yellowness score S satisfies S>T′, it is judged as abnormal and triggers subsequent tracing process;

[0118] The two odors (GSM and 2-MIB) detection and analysis are specifically as follows:

[0119] S21: The sensor outputs the original concentration data C raw , perform background noise calibration and obtain the calibration concentration C calC :

[0120] C cal =C raw -C baseline

[0121] Where: C baseline is the environmental baseline noise concentration for system self-learning;

[0122] S22: Use support vector machine (SVM) to build a low-concentration odor recognition model, and the decision function is:

[0123] f(X')=sign(w·X'+b)

[0124] Where: X′ is the normalized feature vector, w is the weight vector; b is the bias term;

[0125] When the model output is f(X′)=+1, there is odor pollution; when the model output is f(X′)=-1, there is no obvious odor;

[0126] S23: Combined with the water velocity v1, diffusion coefficient D1, and time t1, a prediction model for the spread of two odorous substances is established, using a one-dimensional simplified convection-diffusion equation:

[0127]

[0128] Where: C1 is the concentration of odorous substances, x1 is the spatial position, t1 is the time, v1 is the water flow rate, and D1 is the diffusion coefficient;

[0129] Based on the numerical solution of boundary conditions (such as initial pollution source concentration and node flow velocity changes), the diffusion range and arrival time of odorous substances in the pipe network are predicted;

[0130] After the abnormality is discovered, the following are specifically done:

[0131] S31: Rapidly locate the pollution source by inputting the node water quality sampling concentration and GIS pipe network flow direction data, calling the pollution diffusion model, and requesting the GIS module to load the local pipe network topology; determine whether the change in pollutant concentration gradient conforms to the diffusion characteristics of the upstream source;

[0132] If it matches, mark the suspected pollution source node;

[0133] If it does not meet the requirements, expand the detection range and reposition until the suspected pollution source node is marked;

[0134] S32: Scheduling resource optimization and response prediction, based on the location of the pollution source and the backup water source list R k , importance weight W m , conduct backup water supply feasibility scoring U k :

[0135]

[0136] Among them, Q k is the water supply capacity of the backup water source k; select U k The largest backup source is switched;

[0137] The execution control unit realizes the control response through the execution component, specifically:

[0138] S41: Perform waterway switching and valve control, simulate pollution flow based on GIS pipe network information and estimate pollutant diffusion time:

[0139]

[0140] Where: d is the pipe distance from node i to j, vij is the water flow velocity; according to T ij Quickly adjust the valve switch status to prioritize blocking the transmission channel;

[0141] S42: When the pollution exceeds the limit, the purification unit will be automatically started:

[0142] Treated water quality Cout:

[0143] Chroma detection:

[0144]

[0145] GSM / 2-MIB concentration after treatment:

[0146]

[0147] When the requirement is met Color out <15 degrees, GSM out <5 ng / L, 2-MIB out <5ng / L, stop treatment, otherwise continue deep treatment;

[0148] The verification and evaluation unit implements a closed-loop verification mechanism for water supply quality. The specific calculation process is as follows:

[0149] Deploy mass spectrometers at city outlets or factory outlets to conduct comprehensive water quality testing: 106 indicators are compared:

[0150]

[0151] Where: X j1 is the actual value of the j1th detection index; X j1,GB5749 This is the national standard limit; if all indicators meet the standards, normal water supply will be restored; if there are any items exceeding the limit, the emergency treatment status will continue to be maintained.

[0152] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent water supply yellowing and two odor purification treatment and disposal equipment, comprising: The real-time acquisition unit, intelligent analysis unit and execution control unit are characterized by: The real-time collection unit collects water quality data in real time through sensor terminals deployed at water source inlets, key nodes of branch pipe networks, and building terminals; The intelligent analysis unit dynamically analyzes water quality data through the data center and cloud platform, identifies abnormal yellow water and pollution of two odorous substances, and locates the pollution source; The execution control unit regulates the water supply path, valve status and emergency purification equipment based on the analysis results.

2. The intelligent water yellowing and two odor purification treatment equipment according to claim 1 is characterized in that: It also includes a verification and evaluation unit, which compares the compliance of several water quality indicators with national standards in real time; provides toxicity warnings through zebrafish or cell metabolism reactors; and links execution results with monitoring data to optimize analysis models and equipment layout.

3. The intelligent water yellowing and two odor purification treatment equipment according to claim 1 is characterized in that: The real-time acquisition unit includes: Water source tracing, based on the pipe network topology modeling as a directed graph, calculates the water source probability vector of each node through flow weight and generates labels; Sensor deployment optimization calculates deployment priority based on node response capability, number of covered paths, and blind spot risk, and dynamically adjusts sensor locations.

4. The intelligent water yellowing and two odor purification treatment equipment according to claim 3 is characterized in that: The water source tracing method calculates the path weight using the following formula: in is the instantaneous flow on the i-th edge in the path, Out(u i ) is the node v i all downstream connection edges; and construct a "source-terminal" impact matrix for correlation analysis to generate a water source tracing log.

5. The intelligent water yellowing and two odor purification treatment equipment according to claim 1 is characterized in that: The intelligent analysis unit includes: To identify the yellow characteristics of water, a random forest classification model is constructed based on chromaticity, turbidity, and environmental factors, and the alarm threshold is dynamically adjusted; For the two odor detection, the SVM or CNN model is used to perform pattern recognition on the calibrated odor substance concentrations, and the convection-diffusion equation is combined to predict the pollution spread.

6. The intelligent water yellowing and two odor purification treatment equipment according to claim 5 is characterized in that: The specific process of identifying the yellow feature of water is as follows: S11: Use random forest classifier to identify yellow water anomalies; the training data consists of normal and abnormal water quality samples, and the output category label y∈{0,1}: When y=0, it is normal water quality; when y=1, it is abnormal yellow water quality; Random forest outputs the discrimination results S21: Introduce the environmental impact correction factor λ to dynamically adjust the alarm threshold: λ=f(R,G) Where R is the current rainfall; G is the manhole cover opening and closing abnormality index; f(·) is the empirical regression function; Final alarm threshold: T′=T0×(1+λ) When the water yellowness score S satisfies S>T′, it is judged as abnormal and triggers subsequent tracing processing.

7. The intelligent water yellowing and two odor purification treatment equipment according to claim 5 is characterized in that: The specific process of the two odor detection modules is as follows: S21: The sensor outputs the original concentration data C raw , perform background noise calibration and obtain the calibration concentration C calC : C cal =C raw -C baseline Where: C baseline is the environmental baseline noise concentration for system self-learning; S22: Use support vector machine (SVM) to build a low-concentration odor recognition model, and the decision function is: f(X')=sign(w·X'+b) Where: X′ is the normalized feature vector, w is the weight vector; b is the bias term; When the model output is f(X′)=+1, there is odor pollution; when the model output is f(X′)=-1, there is no obvious odor; S23: Combined with the water velocity v1, diffusion coefficient D1, and time t1, a prediction model for the spread of two odorous substances is established, using a one-dimensional simplified convection-diffusion equation: Where: C1 is the concentration of odorous substances, x1 is the spatial position, t1 is the time, v1 is the water flow rate, and D1 is the diffusion coefficient; the pollutant diffusion range and arrival time are predicted, and the pollution source is located in combination with the pipeline network GIS data.

8. The intelligent water yellowing and two odor purification treatment equipment according to claim 1 is characterized in that: The execution control unit includes: Waterway switching: closing valves or switching to backup water sources based on estimated pollution diffusion time; Purification control: when exceeding the standard is detected, emergency treatment will be initiated until the effluent meets the chromaticity <15 degrees and GSM / 2-MIB <5ng / L.

9. The intelligent water yellowing and two odor purification treatment equipment according to claim 8 is characterized in that: The backup water source scheduling is selected based on feasibility scoring: S31: Rapidly locate pollution sources by inputting node water quality sampling concentration and GIS pipe network flow data, and calling the pollution diffusion model to determine whether the change in pollutant concentration gradient conforms to the diffusion characteristics of the upstream source; If it matches, mark the suspected pollution source node; If it does not meet the requirements, expand the detection range and reposition until the suspected pollution source node is marked; S32: Scheduling resource optimization and response prediction, based on the location of the pollution source and the backup water source list R k , importance weight W m , conduct backup water supply feasibility scoring U k : Among them, Q k is the water supply capacity of the backup water source k.

10. The intelligent water yellowing and two odor purification treatment equipment according to claim 2 is characterized in that: The closed-loop optimization evaluates water quality compliance using the following formula: Deploy mass spectrometers at city outlets or factory outlets to conduct comprehensive water quality testing: 106 indicators are compared: X j1 ≤X j1,GB5749 , Where: X j1 is the actual value of the j1th detection index; X j1,GB5749 This is the national standard limit; if all indicators meet the standards, normal water supply will be restored; if there are any items exceeding the limit, the emergency treatment status will continue to be maintained.

Images