Method for diagnosing inflow and infiltration of low-pollution water in sewage pipeline based on fluorescence spectrum image

Abstract

The invention relates to the technical field of municipal engineering, and provides a sewage pipeline low-pollution water inflow and infiltration diagnosis method based on a fluorescence spectrum image, and the method comprises the steps: collecting different types of sewage and different types of low-pollution water in a sewage pipeline troubleshooting region, and according to the mixing condition of the sewage and the low-pollution water in the sewage pipeline troubleshooting region, determining whether the sewage and the low-pollution water are mixed or not; the collected different types of sewage and different types of low-pollution water are mixed according to different proportions, a three-dimensional fluorescence spectrophotometer is used for detecting a mixed water sample, a fluorescence spectrum image corresponding to the mixed water sample is obtained, and according to the fluorescence spectrum image corresponding to the mixed water sample, the sewage type and proportion and the low-pollution water type and proportion, the low-pollution water is detected. And training the water sample type diagnosis model and the water sample type proportion diagnosis model, and performing low-pollution water inflow and infiltration diagnosis on the sewage pipeline to be diagnosed based on the models. In this way, low-pollution water inflow and infiltration diagnosis can be achieved when multiple types of sewage coexist in a sewage pipeline.

Description

Technical Field

[0001] The invention relates to the technical field of municipal engineering, and in particular to a method for diagnosing the inflow and infiltration of low-polluted water in a sewage pipe based on fluorescence spectrum images. Background Art

[0002] During the long-term operation of urban sewage pipe systems, they are often affected by the inflow and infiltration of low-pollutant water (such as groundwater, mountain spring water, seawater, river water, etc.). After these low-pollutant waters enter the sewage pipes, the concentration of organic matter in the sewage plant inlet water will decrease, affecting the efficiency of biochemical treatment and possibly increasing the operating costs of the sewage treatment plant. In addition, overloaded operation of the pipes may cause overflow pollution and affect the safety of the water environment. Therefore, accurately diagnosing the inflow and infiltration of low-pollutant water in the sewage pipes is of great significance to ensure the normal operation of the urban sewage system.

[0003] At present, the diagnostic methods for low-pollutant water inflow and infiltration in sewage pipes mainly include flow monitoring, water quality analysis and tracer detection. Although these methods can reflect the leakage of pipelines to a certain extent, they often have limitations such as long monitoring cycle, low spatial resolution and high cost. For example, traditional water quality analysis methods mainly rely on indicators such as chemical oxygen demand and total organic carbon, which are difficult to accurately distinguish between sewage and low-pollutant water sources. Therefore, it is urgent to develop more efficient, rapid and economical diagnostic methods to improve diagnostic efficiency and accuracy. Summary of the invention

[0004] In a first aspect, an embodiment of the present invention provides a method for diagnosing low-polluted water inflow and infiltration in a sewage pipe based on fluorescence spectrum images, the method comprising: Collect different types of sewage and different types of low-pollutant water in the sewage pipe inspection area; According to the mixing conditions of sewage and low-pollutant water in the sewage pipeline inspection area, the collected different types of sewage and different types of low-pollutant water are mixed in different proportions, and the mixed water samples are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum image corresponding to the mixed water samples; The fluorescence spectrum image corresponding to the mixed water sample is used as sample data, and the sewage type and low-pollution water type corresponding to the mixed water sample are used as sample labels to construct a water sample type diagnosis training data set. The water sample type diagnosis training data set is used for model training to obtain a water sample type diagnosis model. The mixed water samples are divided into groups according to the sewage type and low-pollution water type corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water samples is used as sample data, and the sewage type ratio and low-pollution water type ratio corresponding to the mixed water samples are used as sample labels to construct a water sample type ratio diagnosis training data set. The water sample type ratio diagnosis training data set is used for model training to obtain a water sample type ratio diagnosis model corresponding to the group. For the sewage pipe to be diagnosed, the inspection wells are opened one by one during the peak water consumption on sunny days and water samples in the inspection wells are collected. The water samples in the inspection wells are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum images corresponding to the water samples in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type diagnosis model to perform water sample type diagnosis. The water samples in the inspection wells are divided into groups according to the diagnosed sewage types and low-pollution water types of the water samples in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis. According to the sewage type ratio and low-pollution water type ratio of the diagnosed water samples in the inspection wells, low-pollution water inflow and infiltration diagnosis is performed on the sewage pipe to be diagnosed.

[0005] In some implementable embodiments of the first aspect, different types of sewage are determined based on the types of water users in the sewage pipe inspection area, including domestic sewage, catering wastewater, commercial wastewater and / or papermaking wastewater; different types of low-pollution water are determined based on the geographical characteristics of the sewage pipe inspection area, including groundwater, mountain spring water, sea water and / or river water.

[0006] In some implementable methods of the first aspect, the mixing of sewage and low-pollution water in the sewage pipe inspection area is determined according to the local natural environmental conditions and the functional type of the drainage area; specifically, when the groundwater level in the sewage pipe inspection area is higher than the buried depth of the sewage pipe, different types of sewage in the sewage pipe inspection area are mixed with groundwater; when there is mountain spring water in the sewage pipe inspection area, different types of sewage in the sewage pipe inspection area are mixed with mountain spring water; when the sewage pipe inspection area is a residential area, domestic sewage is mixed with different types of low-pollution water; when there are catering and commercial businesses in the sewage pipe inspection area at the same time, catering wastewater and commercial wastewater are mixed with different types of low-pollution water.

[0007] In some implementations of the first aspect, when the collected different types of sewage and different types of low-pollutant water are mixed in different proportions, the proportion of low-pollutant water is gradually increased from 0% to 95%, with an increase interval of at least 5%.

[0008] In some implementable methods of the first aspect, before using a three-dimensional fluorescence spectrometer to detect the mixed water sample, the mixed water sample is filtered using a filter membrane with a pore size of 0.22-0.45 μm; when using a three-dimensional fluorescence spectrometer to detect the mixed water sample, a quartz cuvette is used, and the voltage of the three-dimensional fluorescence spectrometer is set to 400 V, the excitation wavelength range is 200-450 nm, the emission wavelength range is 250-550 nm, and the slit width is 2 nm.

[0009] In some implementable embodiments of the first aspect, model training is performed using a water sample type diagnosis training data set to obtain a water sample type diagnosis model, including: Select a machine learning model and use the water sample type diagnosis training dataset to train the model. When the accuracy of the trained model does not meet the requirements, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the requirements, and finally obtain the water sample type diagnosis model.

[0010] In some implementations of the first aspect, the machine learning models that can be selected when using the water sample type diagnosis training data set for model training include support vector machine, convolutional neural network, random forest, k-nearest neighbor algorithm, and gradient boosting machine.

[0011] In some implementable embodiments of the first aspect, model training is performed using a water sample type ratio diagnosis training data set to obtain a water sample type ratio diagnosis model corresponding to a group, including: Select a machine learning model and use the water sample type proportion diagnosis training data set to train the model. When the accuracy of the trained model does not meet the requirements, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the requirements, and finally obtain the water sample type proportion diagnosis model corresponding to the group.

[0012] In some implementations of the first aspect, selectable machine learning models when using a water sample type ratio diagnostic training data set for model training include support vector regression, random forest regression, partial least squares regression, and neural network regression.

[0013] In some implementable methods of the first aspect, the sewage pipe to be diagnosed includes multiple sewage pipe sections to be diagnosed; the fluorescence spectrum image corresponding to the water sample in the inspection well is input into the water sample type diagnosis model to perform water sample type diagnosis, the water samples in the inspection well are divided into groups according to the sewage type and low-pollution water type of the water samples in the inspection well diagnosed, the fluorescence spectrum image corresponding to the water samples in the inspection well is input into the water sample type ratio diagnosis model corresponding to the group to perform water sample type ratio diagnosis, and low-pollution water inflow and infiltration diagnosis is performed on the sewage pipe to be diagnosed according to the sewage type ratio and low-pollution water type ratio of the water samples in the inspection well diagnosed, including: The fluorescence spectral images corresponding to the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed are synchronously input into the water sample type diagnosis model to perform water sample type diagnosis. According to the sewage type and low-pollution water type of the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed, the sewage type and low-pollution water type of the receiving water body of the sewage pipe section to be diagnosed are determined for group division. The fluorescence spectral images corresponding to the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed are input into the water sample type ratio diagnosis model corresponding to the group to perform water sample type ratio diagnosis. According to the sewage type ratio and low-pollution water type ratio of the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed, low-pollution water inflow and infiltration diagnosis is performed on the sewage pipe section to be diagnosed.

[0014] In a second aspect, an embodiment of the present invention provides a sewage pipe low-pollution water inflow and infiltration diagnosis system based on fluorescence spectrum image, the system comprising: A collection module is used to collect different types of sewage and different types of low-pollutant water in the sewage pipe inspection area; The detection module is used to mix different types of sewage and different types of low-pollutant water collected in different proportions according to the mixing conditions of sewage and low-pollutant water in the sewage pipe inspection area, and use a three-dimensional fluorescence spectrometer to detect the mixed water sample to obtain a fluorescence spectrum image corresponding to the mixed water sample; A training module is used to construct a water sample type diagnosis training data set using the fluorescence spectrum image corresponding to the mixed water sample as sample data and the sewage type and low-pollution water type corresponding to the mixed water sample as sample labels, and to perform model training using the water sample type diagnosis training data set to obtain a water sample type diagnosis model; The training module is also used to divide the mixed water samples into groups according to the sewage types and low-polluted water types corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water samples is used as sample data, and the sewage type ratio and low-polluted water type ratio corresponding to the mixed water samples are used as sample labels to construct a water sample type ratio diagnosis training data set, and the water sample type ratio diagnosis training data set is used for model training to obtain a water sample type ratio diagnosis model corresponding to the group; The diagnosis module is used to open the inspection wells one by one and collect water samples in the inspection wells for the sewage pipe to be diagnosed during the peak water consumption on sunny days, and use a three-dimensional fluorescence spectrometer to detect the water samples in the inspection wells to obtain the fluorescence spectrum images corresponding to the water samples in the inspection wells, and input the fluorescence spectrum images corresponding to the water samples in the inspection wells into the water sample type diagnosis model to perform water sample type diagnosis, and divide the water samples in the inspection wells into groups according to the sewage types and low-pollution water types of the diagnosed water samples in the inspection wells, and input the fluorescence spectrum images corresponding to the water samples in the inspection wells into the water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis, and perform low-pollution water inflow and infiltration diagnosis on the sewage pipe to be diagnosed according to the sewage type ratio and low-pollution water type ratio of the diagnosed water samples in the inspection wells.

[0015] Compared with the prior art, the present invention has at least the following technical effects: (1) The present invention uses fluorescence spectroscopy to determine the type of low-polluted water and the inflow-infiltration ratio in the sewage pipe, solving the problem that the current sewage pipe geophysical exploration technology (such as QV, CCTV, etc.) has high diagnostic costs and is difficult to diagnose and trace the source of low-polluted water.

[0016] (2) The present invention preferentially diagnoses the sewage type and low-polluted water type of the main receiving water body in the pipe section through the water sample type diagnosis model, and can select an appropriate water sample type ratio diagnosis model for the subsequent water sample type ratio diagnosis, thereby improving the diagnosis efficiency and accuracy.

[0017] (3) The present invention analyzes the ratio of low-polluted water inflow and infiltration in the sewage pipe section through a machine learning model, thereby overcoming the problem that the fluorescence characteristic peak intensity and the dilution ratio are not linearly related under the conditions of internal filtration effect and coexistence peak interference.

[0018] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present invention and do not constitute a limitation of the present invention. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein: Figure 1 A flow chart of a method for diagnosing low-polluted water inflow and infiltration in a sewage pipe based on fluorescence spectrum images provided by an embodiment of the present invention; Figure 2 ROC diagram of the prediction effect of the water sample type diagnosis model provided by the embodiment of the present invention; Figure 3A diagram showing the prediction effect of the water sample type ratio diagnostic model provided by an embodiment of the present invention; Figure 4 A structural diagram of a sewage pipe low-polluted water inflow and infiltration diagnostic system based on fluorescence spectrum imaging provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In addition, the term "and / or" in the present invention is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects before and after are in an "or" relationship.

[0022] In order to solve the technical problems arising from the background technology, the embodiment of the present invention provides a method and system for diagnosing the inflow and infiltration of low-polluted water in sewage pipes based on fluorescence spectrum images, which can collect different types of sewage and different types of low-polluted water in the sewage pipe inspection area, and according to the mixing of sewage and low-polluted water in the sewage pipe inspection area, the collected different types of sewage and different types of low-polluted water are mixed in different proportions, and the mixed water sample is detected using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum image corresponding to the mixed water sample, and according to the fluorescence spectrum image, sewage type and proportion, low-polluted water type and proportion corresponding to the mixed water sample, a water sample type diagnosis model and a water sample type proportion diagnosis model are trained, and based on the above models, low-polluted water inflow and infiltration diagnosis is performed on the sewage pipe to be diagnosed. In this way, low-polluted water inflow and infiltration diagnosis can be achieved when multiple types of sewage coexist in the sewage pipe, and the diagnostic efficiency and accuracy are effectively improved.

[0023] In conjunction with the accompanying drawings, a method and system for diagnosing inflow and infiltration of low-polluted water in sewage pipes based on fluorescence spectrum images provided by an embodiment of the present invention will be described in detail through specific embodiments.

[0024] Figure 1 A flowchart of a method for diagnosing low-polluted water inflow and infiltration in a sewage pipe based on fluorescence spectral images is provided in an embodiment of the present invention, such as Figure 1 As shown, the sewer pipe low-polluted water inflow infiltration diagnosis method 100 may include: S110, collecting different types of sewage and different types of low-pollutant water in the sewage pipe inspection area.

[0025] Among them, different types of sewage are determined according to the types of water users in the sewage pipe inspection area, including but not limited to domestic sewage, catering wastewater, commercial wastewater and / or papermaking wastewater; different types of low-pollution water are determined according to the geographical characteristics of the sewage pipe inspection area, including but not limited to groundwater, mountain spring water, sea water and / or river water.

[0026] S120, according to the mixing conditions of sewage and low-pollutant water in the sewage pipe inspection area, the collected different types of sewage and different types of low-pollutant water are mixed in different proportions, and the mixed water samples are tested using a three-dimensional fluorescence spectrometer to obtain a fluorescence spectrum image corresponding to the mixed water samples.

[0027] Among them, the mixing of sewage and low-pollution water in the sewage pipe inspection area is determined according to the local natural environmental conditions and the functional type of the drainage area. Specifically, when the groundwater level in the sewage pipe inspection area is higher than the buried depth of the sewage pipe, different types of sewage in the sewage pipe inspection area are mixed with groundwater; when there is mountain spring water in the sewage pipe inspection area, different types of sewage in the sewage pipe inspection area are mixed with mountain spring water; when the sewage pipe inspection area is a residential area, domestic sewage is mixed with different types of low-pollution water; when there are both catering and commercial businesses in the sewage pipe inspection area, catering wastewater and commercial wastewater are mixed with different types of low-pollution water.

[0028] As an example, when the collected different types of sewage and different types of low-polluted water are mixed in different proportions, the proportion of low-polluted water needs to be gradually increased from 0% to 95%, and the increase interval should be at least 5%. Furthermore, in order to avoid errors caused by the instability of the light source of the three-dimensional fluorescence spectrometer and filtration pretreatment, the mixing ratio of low-polluted water shall not be less than 5% of the total volume. Before using the three-dimensional fluorescence spectrometer to detect the mixed water sample, it is necessary to use a filter membrane with a pore size of 0.22-0.45μm to filter the mixed water sample. When using a three-dimensional fluorescence spectrometer to detect the mixed water sample, a quartz cuvette is required, and the voltage of the three-dimensional fluorescence spectrometer is set to 400V, the excitation wavelength range is 200-450nm, the emission wavelength range is 250-550nm, and the slit width is 2nm.

[0029] S130, using the fluorescence spectrum image corresponding to the mixed water sample as sample data, and the sewage type and low-pollution water type corresponding to the mixed water sample as sample labels, constructing a water sample type diagnosis training data set, using the water sample type diagnosis training data set for model training, and obtaining a water sample type diagnosis model.

[0030] Specifically, the fluorescence spectrum image corresponding to the mixed water sample can be used as sample data, and the sewage type and low-pollution water type corresponding to the mixed water sample can be used as sample labels to construct a water sample type diagnosis training data set, select a machine learning model, and use the water sample type diagnosis training data set for model training based on this. When the accuracy of the trained model does not meet the standard, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the standard, and finally obtain a water sample type diagnosis model. Optionally, the machine learning models that can be selected during training mainly include support vector machines, convolutional neural networks, random forests, k-nearest neighbor algorithms, and gradient boosting machines.

[0031] S140, divide the mixed water samples into groups according to the sewage types and low-pollution water types corresponding to the mixed water samples. For any group, use the fluorescence spectrum image corresponding to the mixed water samples as sample data, and use the sewage type ratio and low-pollution water type ratio corresponding to the mixed water samples as sample labels to construct a water sample type ratio diagnosis training data set. Use the water sample type ratio diagnosis training data set to perform model training to obtain a water sample type ratio diagnosis model corresponding to the group.

[0032] Specifically, the mixed water samples can be divided into groups according to the sewage type and low-polluted water type corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water samples is used as the sample data, and the sewage type ratio and low-polluted water type ratio corresponding to the mixed water samples are used as sample labels to construct a water sample type ratio diagnosis training data set, select a machine learning model, and use the water sample type ratio diagnosis training data set for model training based on this. When the accuracy of the trained model does not meet the standard, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the standard, and finally obtain the water sample type ratio diagnosis model corresponding to the group. Optionally, the machine learning models that can be selected during training mainly include support vector regression, random forest regression, partial least squares regression, and neural network regression.

[0033] S150, for the sewage pipe to be diagnosed, open the inspection wells one by one and collect water samples in the inspection wells during the peak water consumption period on sunny days, and use a three-dimensional fluorescence spectrometer to detect the water samples in the inspection wells to obtain fluorescence spectrum images corresponding to the water samples in the inspection wells, input the fluorescence spectrum images corresponding to the water samples in the inspection wells into the water sample type diagnosis model to perform water sample type diagnosis, divide the water samples in the inspection wells into groups according to the diagnosed sewage types and low-pollution water types of the water samples in the inspection wells, input the fluorescence spectrum images corresponding to the water samples in the inspection wells into the water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis, and perform low-pollution water inflow and infiltration diagnosis on the sewage pipe to be diagnosed according to the sewage type ratio and low-pollution water type ratio of the water samples in the inspection wells.

[0034] Among them, the sewage pipe to be diagnosed includes multiple sewage pipe sections to be diagnosed. In view of this, the fluorescence spectrum images corresponding to the water samples in the adjacent inspection wells of the same sewage pipe section to be diagnosed can be synchronously input into the water sample type diagnosis model to perform water sample type diagnosis, and the sewage type and low-polluted water type of the main receiving water body of the sewage pipe section to be diagnosed are determined according to the sewage type and low-polluted water type of the water samples in the adjacent inspection wells of the same sewage pipe section to be diagnosed, so as to perform group division, and the fluorescence spectrum images corresponding to the water samples in the adjacent inspection wells of the same sewage pipe section to be diagnosed are input into the water sample type ratio diagnosis model corresponding to the group to perform water sample type ratio diagnosis, and the low-polluted water inflow and infiltration diagnosis of the sewage pipe section to be diagnosed is performed according to the sewage type ratio and low-polluted water type ratio of the water samples in the adjacent inspection wells of the same sewage pipe section to be diagnosed, and the diagnosis result is output.

[0035] For further understanding, the above-mentioned sewage pipe low-polluted water inflow and infiltration diagnosis method 100 is described in detail below in conjunction with a specific embodiment: (1) Collect different types of sewage and different types of low-pollutant water in the sewage pipe inspection area.

[0036] It is assumed here that the sewage pipe inspection area is mainly residential areas and catering and commercial areas, and the average elevation of the sewage pipe is lower than the elevation of groundwater and river water along the line.

[0037] Based on the above situation, domestic sewage and catering wastewater in the sewage pipe inspection area are collected as the main sewage types, and groundwater and river water in the sewage pipe inspection area are collected as the main low-pollution water types.

[0038] (2) According to the mixing conditions of sewage and low-pollutant water in the sewage pipe inspection area, the different types of sewage and low-pollutant water collected are mixed in different proportions, and the mixed water samples are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum image corresponding to the mixed water samples.

[0039] Refer to Table 1, set different groups, and gradually increase the proportion of low-polluted water from 0% to 95%, with an interval of 5%. If there are more than two types of sewage in the mixed water sample, different sewage proportions should be set according to the actual situation. In this example, catering wastewater accounts for up to 30% of the mixed water sample. In this embodiment, the proportions of different types of sewage and low-polluted water (i.e., volume proportions) are shown in Table 1: Table 1

[0040] (3) Using the fluorescence spectrum image corresponding to the mixed water sample as sample data, and the sewage type and low-pollution water type corresponding to the mixed water sample as sample labels, a water sample type diagnosis training dataset is constructed. The water sample type diagnosis training dataset is used for model training to obtain a water sample type diagnosis model.

[0041] Among them, the ROC curve of the water sample type diagnosis model is as follows: Figure 2 As shown in the figure, its AUC is 0.95, which shows that its water sample type diagnosis effect is good and the model has high reliability.

[0042] (4) The mixed water samples are divided into groups according to the sewage type and low-pollution water type corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water samples is used as the sample data, and the sewage type ratio and low-pollution water type ratio corresponding to the mixed water samples are used as the sample labels to construct a water sample type ratio diagnosis training data set. The water sample type ratio diagnosis training data set is used for model training to obtain a water sample type ratio diagnosis model corresponding to the group.

[0043] Among them, the prediction effect of the water sample type ratio diagnosis model is as follows: Figure 3 As shown, its R 2 It is 0.91, indicating that the diagnostic result of the water sample type ratio is good and the model has high reliability.

[0044] (5) For the sewage pipes to be diagnosed, the inspection wells are opened one by one during the peak water consumption period on sunny days and water samples are collected from the inspection wells. The water samples in the inspection wells are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum images corresponding to the water samples in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type diagnosis model to perform water sample type diagnosis. The water samples in the inspection wells are divided into groups according to the sewage types and low-pollution water types diagnosed in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis. According to the sewage type ratio and low-pollution water type ratio of the water samples in the inspection wells, low-pollution water inflow and infiltration diagnosis is performed on the sewage pipes to be diagnosed.

[0045] Furthermore, the fluorescence spectrum images corresponding to the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed are taken as input in pairs, and the output results are shown in Table 2: Table 2

[0046] In Table 2, the upstream and downstream of each pipe section are the diagnostic results of the upstream and downstream inspection wells respectively.

[0047] Refer to Table 2. For the sewage pipes to be diagnosed, there is no obvious inflow and infiltration of low-polluted water in sections one to three; there is groundwater infiltration in section four, and the infiltration ratio is 20%, indicating that section four is damaged and needs structural repair; there is river water inflow in section five, and the inflow ratio is 10%, and the interception points along the coast need to be checked.

[0048] In summary, the present invention achieves at least the following technical effects: (1) The present invention uses fluorescence spectroscopy to determine the type of low-polluted water and the inflow-infiltration ratio in the sewage pipe, solving the problem that the current sewage pipe geophysical exploration technology (such as QV, CCTV, etc.) has high diagnostic costs and is difficult to diagnose and trace the source of low-polluted water.

[0049] (2) The present invention preferentially diagnoses the sewage type and low-polluted water type of the main receiving water body in the pipe section through the water sample type diagnosis model, and can select an appropriate water sample type ratio diagnosis model for the subsequent water sample type ratio diagnosis, thereby improving the diagnosis efficiency and accuracy.

[0050] (3) The present invention analyzes the ratio of low-polluted water inflow and infiltration in the sewage pipe section through a machine learning model, thereby overcoming the problem that the fluorescence characteristic peak intensity and the dilution ratio are not linearly related under the conditions of internal filtration effect and coexistence peak interference.

[0051] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0052] The above is an introduction to the method embodiment. The following is a further explanation of the solution of the present invention through a system embodiment.

[0053] Figure 4 A structural diagram of a sewage pipe low-pollution water inflow and infiltration diagnosis system based on fluorescence spectrum imaging provided by an embodiment of the present invention, such as Figure 4 As shown, the sewer pipe low-pollution water inflow infiltration diagnostic system 400 may include: The collection module 410 is used to collect different types of sewage and different types of low-polluted water in the sewage pipe inspection area.

[0054] The detection module 420 is used to mix the collected different types of sewage and different types of low-pollutant water in different proportions according to the mixing conditions of sewage and low-pollutant water in the sewage pipe inspection area, and use a three-dimensional fluorescence spectrometer to detect the mixed water sample to obtain a fluorescence spectrum image corresponding to the mixed water sample.

[0055] The training module 430 is used to construct a water sample type diagnosis training data set using the fluorescence spectrum image corresponding to the mixed water sample as sample data and the sewage type and low-pollution water type corresponding to the mixed water sample as sample labels, and use the water sample type diagnosis training data set to perform model training to obtain a water sample type diagnosis model.

[0056] The training module 430 is also used to divide the mixed water samples into groups according to the sewage type and low-pollution water type corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water sample is used as sample data, and the sewage type ratio and low-pollution water type ratio corresponding to the mixed water sample are used as sample labels to construct a water sample type ratio diagnosis training data set, and the water sample type ratio diagnosis training data set is used for model training to obtain a water sample type ratio diagnosis model corresponding to the group.

[0057] Diagnostic module 440 is used to open the inspection wells one by one and collect water samples in the inspection wells during the peak water consumption period on sunny days for the sewage pipe to be diagnosed, and use a three-dimensional fluorescence spectrometer to detect the water samples in the inspection wells to obtain fluorescence spectrum images corresponding to the water samples in the inspection wells, input the fluorescence spectrum images corresponding to the water samples in the inspection wells into a water sample type diagnosis model to perform water sample type diagnosis, divide the water samples in the inspection wells into groups according to the diagnosed sewage types and low-pollution water types of the water samples in the inspection wells, input the fluorescence spectrum images corresponding to the water samples in the inspection wells into a water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis, and perform low-pollution water inflow and infiltration diagnosis on the sewage pipe to be diagnosed according to the sewage type ratio and low-pollution water type ratio of the diagnosed water samples in the inspection wells.

[0058] Understandably, Figure 4 The modules / units in the sewage pipe low-polluted water inflow infiltration diagnostic system 400 shown in the figure have the functions of realizing Figure 1 The functions of the various steps in the sewage pipe low-polluted water inflow and infiltration diagnostic method 100 shown and their corresponding technical effects are not repeated here for the sake of brevity.

[0059] It should be noted that the above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for diagnosing low-pollution water inflow and infiltration in sewage pipes based on fluorescence spectral images, characterized in that: The method comprises: Collect different types of sewage and different types of low-pollutant water in the sewage pipe inspection area; According to the mixing conditions of sewage and low-pollutant water in the sewage pipeline inspection area, the collected different types of sewage and different types of low-pollutant water are mixed in different proportions, and the mixed water samples are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum image corresponding to the mixed water samples; The fluorescence spectrum image corresponding to the mixed water sample is used as sample data, and the sewage type and low-pollution water type corresponding to the mixed water sample are used as sample labels to construct a water sample type diagnosis training data set. The water sample type diagnosis training data set is used for model training to obtain a water sample type diagnosis model. The mixed water samples are divided into groups according to the sewage type and low-pollution water type corresponding to the mixed water samples. For any group, the fluorescence spectrum image corresponding to the mixed water samples is used as sample data, and the sewage type ratio and low-pollution water type ratio corresponding to the mixed water samples are used as sample labels to construct a water sample type ratio diagnosis training data set. The water sample type ratio diagnosis training data set is used for model training to obtain a water sample type ratio diagnosis model corresponding to the group. For the sewage pipe to be diagnosed, the inspection wells are opened one by one during the peak water consumption on sunny days and water samples in the inspection wells are collected. The water samples in the inspection wells are tested using a three-dimensional fluorescence spectrometer to obtain the fluorescence spectrum images corresponding to the water samples in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type diagnosis model to perform water sample type diagnosis. The water samples in the inspection wells are divided into groups according to the diagnosed sewage types and low-pollution water types of the water samples in the inspection wells. The fluorescence spectrum images corresponding to the water samples in the inspection wells are input into the water sample type ratio diagnosis model corresponding to the groups to perform water sample type ratio diagnosis. According to the sewage type ratio and low-pollution water type ratio of the diagnosed water samples in the inspection wells, low-pollution water inflow and infiltration diagnosis is performed on the sewage pipe to be diagnosed.

2. The method according to claim 1, characterized in that The different types of sewage are determined according to the types of water users in the sewage pipe inspection area, including domestic sewage, catering wastewater, commercial wastewater and / or papermaking wastewater; the different types of low-pollution water are determined according to the geographical characteristics of the sewage pipe inspection area, including groundwater, mountain spring water, sea water and / or river water.

3. The method according to claim 1, characterized in that: The mixing of sewage and low-pollution water in the sewage pipe inspection area is determined according to the local natural environmental conditions and the functional type of the drainage area; specifically, when the groundwater level in the sewage pipe inspection area is higher than the buried depth of the sewage pipe, different types of sewage in the sewage pipe inspection area are mixed with groundwater; when there is mountain spring water in the sewage pipe inspection area, different types of sewage in the sewage pipe inspection area are mixed with mountain spring water; when the sewage pipe inspection area is a residential area, domestic sewage is mixed with different types of low-pollution water; when there are catering and commercial businesses in the sewage pipe inspection area at the same time, catering wastewater and commercial wastewater are mixed with different types of low-pollution water.

4. The method according to claim 1, characterized in that When the collected different types of sewage and different types of low-pollutant water are mixed in different proportions, the proportion of low-pollutant water shall be gradually increased from 0% to 95%, with an increase interval of at least 5%.

5. The method according to claim 1, characterized in that Before using the three-dimensional fluorescence spectrometer to detect the mixed water sample, the mixed water sample is filtered with a filter membrane with a pore size of 0.22-0.45 μm; when using the three-dimensional fluorescence spectrometer to detect the mixed water sample, a quartz cuvette is used, and the voltage of the three-dimensional fluorescence spectrometer is set to 400 V, the excitation wavelength range is 200-450 nm, the emission wavelength range is 250-550 nm, and the slit width is 2 nm.

6. The method according to claim 1, characterized in that The method of using the water sample type diagnosis training data set to perform model training to obtain a water sample type diagnosis model includes: Select a machine learning model and use the water sample type diagnosis training dataset to train the model. When the accuracy of the trained model does not meet the requirements, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the requirements, and finally obtain the water sample type diagnosis model.

7. The method according to claim 6, characterized in that The machine learning models that can be selected when using the water sample type diagnosis training dataset for model training include support vector machine, convolutional neural network, random forest, k-nearest neighbor algorithm, and gradient boosting machine.

8. The method according to claim 1, characterized in that The water sample type ratio diagnosis training data set is used to perform model training to obtain a water sample type ratio diagnosis model corresponding to the group, including: Select a machine learning model and use the water sample type proportion diagnosis training data set to train the model. When the accuracy of the trained model does not meet the requirements, replace the machine learning model or adjust the model hyperparameters until the accuracy meets the requirements, and finally obtain the water sample type proportion diagnosis model corresponding to the group.

9. The method according to claim 8, characterized in that The machine learning models that can be selected when using the water sample type ratio diagnosis training dataset for model training include support vector regression, random forest regression, partial least squares regression, and neural network regression.

10. The method according to claim 1, characterized in that The sewage pipeline to be diagnosed includes a plurality of sewage pipe sections to be diagnosed; the fluorescence spectrum image corresponding to the water sample in the inspection well is input into the water sample type diagnosis model to perform water sample type diagnosis, the water samples in the inspection well are divided into groups according to the sewage type and low-pollution water type of the water samples in the inspection well diagnosed, the fluorescence spectrum image corresponding to the water samples in the inspection well is input into the water sample type ratio diagnosis model corresponding to the group to perform water sample type ratio diagnosis, and low-pollution water inflow and infiltration diagnosis is performed on the sewage pipeline to be diagnosed according to the sewage type ratio and low-pollution water type ratio of the water samples in the inspection well diagnosed, including: The fluorescence spectral images corresponding to the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed are synchronously input into the water sample type diagnosis model to perform water sample type diagnosis. According to the sewage type and low-pollution water type of the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed, the sewage type and low-pollution water type of the receiving water body of the sewage pipe section to be diagnosed are determined for group division. The fluorescence spectral images corresponding to the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed are input into the water sample type ratio diagnosis model corresponding to the group to perform water sample type ratio diagnosis. According to the sewage type ratio and low-pollution water type ratio of the water samples in adjacent inspection wells of the same sewage pipe section to be diagnosed, low-pollution water inflow and infiltration diagnosis is performed on the sewage pipe section to be diagnosed.

Images