A city underground drainage pipe network problem diagnosis system based on big data analysis

Abstract

The application discloses a kind of urban underground drainage pipe network problem diagnosis system based on big data analysis, it is related to intelligent pipe network technical field, comprising the following steps: S1, data acquisition and pretreatment;S2, carry out feature fusion, construct feature vector;S3, construct diagnosis model based on machine learning algorithm to identify whether there is abnormality in drainage pipe network;S4, predict the occurrence trend of potential problem;S5, build optimization model to maintain.The application can not only accurately diagnose common problems such as blockage and leakage in pipe network by deeply fusing multi-source data (such as flow, water level, water quality, geographic information, historical maintenance records, etc.), but also predict the occurrence trend of problems in advance using big data mining and machine learning algorithms.The application greatly improves the operation efficiency and safety of the drainage system, and helps to optimize resource allocation and reduce maintenance costs through intelligent analysis, thereby realizing more refined and dynamic management of drainage pipe network.

Description

Technical Field

[0001] This invention relates to the field of intelligent pipeline technology, specifically to a problem diagnosis system for urban underground drainage pipeline networks based on big data analysis. Background Technology

[0002] With the acceleration of urbanization, the management and maintenance of urban underground drainage networks face enormous challenges. Traditional methods for diagnosing drainage network problems often rely on manual inspections, periodic maintenance, and experience-based judgment, lacking real-time monitoring and intelligent management. This results in many potential problems going undetected, impacting the efficiency and safety of the drainage system. Furthermore, existing technologies have limitations in the accuracy and predictive capabilities of data analysis and anomaly detection, failing to fully utilize big data technologies and advanced analytical algorithms for comprehensive diagnosis.

[0003] Currently, most urban drainage network diagnostic systems rely on single sensor data (such as flow, water level, or water quality monitoring) for problem identification. This approach is susceptible to data noise, localized characteristics, and missing data, and often lacks the fusion of cross-domain information. For example, the correlation between water level changes and water quality deterioration has not been effectively explored, and changes in flow data and the surrounding environment of the pipeline have not been adequately considered. More importantly, existing technologies are often limited to the "post-event" discovery of problems, making it difficult to predict trends and provide proactive early warnings, thus resulting in a large number of potential risks failing to be addressed in advance. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a diagnostic system for urban underground drainage pipe networks based on big data analysis, in order to solve the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a problem diagnosis system for urban underground drainage pipe networks based on big data analysis, comprising the following steps:

[0007] S1. Data acquisition and preprocessing: obtaining sensor data, historical maintenance records, and pipeline geographic information data from different data sources;

[0008] S2. Extract key features from data from different data sources, perform feature fusion, and construct feature vectors;

[0009] S3. Based on feature vectors, construct a diagnostic model to identify whether there are abnormalities in the drainage network using machine learning algorithms;

[0010] S4. After completing the anomaly identification, predict the occurrence trend of potential problems based on historical data and current operating status;

[0011] S5. After predicting the trend of potential problems, build and maintain an optimization model.

[0012] To further optimize this technical solution, in step S1, the collected sensor data includes flow rate, water level, and water quality;

[0013] Preprocessing of sensor data includes missing value imputation, noise removal, outlier detection, and data standardization to obtain standardized sensor data. The standardized formula is as follows:

[0014]

[0015] in,

[0016] The raw sensor data X collected, and These are the mean and standard deviation of the original sensor data X, respectively;

[0017] Standardized sensor data It can eliminate the influence between different dimensions in the raw sensor data, enabling data with different characteristics to be analyzed on the same scale.

[0018] To further optimize this technical solution, in step S2, features that can reflect the operating status of the drainage network are extracted from sensor data, historical maintenance records and pipeline geographic information data obtained from different data sources.

[0019] Time series features, including mean, variance, maximum and minimum values, are extracted from sensor data. Spatial features, including pipeline aging, maintenance frequency and regional distribution, are extracted from historical maintenance records and pipeline geographic information.

[0020] To further optimize this technical solution, feature fusion is performed on the extracted features, integrating features from different data sources into a unified feature vector. ;

[0021] Based on the standardized sensor data obtained in step S1 Historical maintenance records and pipeline geographic information data Then the eigenvector It is constructed by weighted fusion of different features, and the formula is as follows:

[0022]

[0023] in,

[0024] These are weighting coefficients, representing the contribution of sensor data to the feature vector;

[0025] These are weighting coefficients, representing the contribution of historical maintenance records to the feature vector;

[0026] These are weighting coefficients, representing the contribution of pipeline geographic information data to the feature vector;

[0027] Received It is a feature vector that integrates multi-source data and will be used as input for subsequent steps.

[0028] To further optimize this technical solution, in step S3, based on Apply machine learning algorithms to identify the presence of anomalies;

[0029] Anomalies include pipe network blockage, abnormal water flow, excessively high or low water levels, and pipe leaks. A diagnostic model is constructed, and algorithms are used to uncover the relationship between features and anomalies, thereby identifying anomalies in real-time data and outputting diagnostic results. This indicates whether an anomaly exists.

[0030] The diagnostic model is further optimized as follows:

[0031]

[0032] in,

[0033] The threshold value represents the boundary between normal and abnormal conditions.

[0034] This is a diagnostic module built using machine learning networks.

[0035] To further optimize this technical solution, the diagnostic module includes a sub-model layer, a fusion layer, and a discrimination layer;

[0036] In the sub-model layer, for each key feature The prediction error is calculated using historical time series data, and a local anomaly score is defined.

[0037]

[0038] in,

[0039] The predicted value at time t;

[0040] This represents the standard deviation of the error in the training set for that dimension.

[0041] The standardized anomaly score for the i-th key feature;

[0042] In the fusion layer, the local anomaly scores of all key features are weighted according to their importance to obtain the total diagnostic score:

[0043]

[0044] in,

[0045] ,satisfy This was obtained through historical diagnostic correlation analysis;

[0046] In the discriminant layer, a threshold is set. The maximum Youden exponent is determined by taking the training data under the ROC curve.

[0047] To further optimize this technical solution, in step S4, known historical data... and diagnostic results The prediction model is shown below, which forecasts the trend of potential problems occurring at time T in the future:

[0048]

[0049] in,

[0050] To predict the trend of potential problems;

[0051] The LSTM algorithm is used to predict the algorithm function;

[0052] The model can predict potential future problems and provide early warning signals for maintenance work.

[0053] To further optimize this technical solution, in step S5, the optimized model is as follows:

[0054]

[0055] in,

[0056] For optimized maintenance strategies;

[0057] This represents the cost of performing the i-th maintenance task under strategy S;

[0058] It is the predicted value for the i-th type of potential problem;

[0059] It is the resource constraint under strategy S;

[0060] It is a regulatory factor used to balance cost and resource constraints;

[0061] This model is designed to minimize maintenance costs and resource waste while maximizing the efficiency of pipeline fault repair.

[0062] To further optimize this technical solution, the system also includes the following functional modules: acquisition and preprocessing module, feature fusion module, anomaly diagnosis module, state prediction module, and strategy optimization module.

[0063] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, they implement the steps of a problem diagnosis system for urban underground drainage pipe network based on big data analysis as described in the first aspect of the present invention.

[0064] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a problem diagnosis system for urban underground drainage pipe networks based on big data analysis as described in the first aspect of the present invention.

[0065] Compared with existing technologies, this invention provides a diagnostic system for urban underground drainage pipe networks based on big data analysis, which has the following beneficial effects:

[0066] This urban underground drainage network problem diagnosis system, based on big data analytics, deeply integrates multi-source data (such as flow rate, water level, water quality, geographic information, and historical maintenance records). Utilizing big data mining and machine learning algorithms, it can not only accurately diagnose common problems in the network, such as blockages and leaks, but also predict the trends of these problems in advance. This transforms urban drainage network management from a traditional "post-event response" approach to a "balanced approach of early warning and prevention," significantly improving the operational efficiency and safety of the drainage system. Simultaneously, intelligent analysis helps optimize resource allocation and reduce maintenance costs, thereby achieving more refined and dynamic drainage network management. Attached Figure Description

[0067] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0068] Figure 1 This is a schematic diagram of a module of an urban underground drainage network problem diagnosis system based on big data analysis proposed in this invention;

[0069] Figure 2 This is a flowchart illustrating a problem diagnosis system for urban underground drainage pipe networks based on big data analysis proposed in this invention. Detailed Implementation

[0070] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0071] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0072] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0073] Example 1:

[0074] Reference Figures 1-2 This is the first embodiment of the present invention, which provides a diagnostic system for urban underground drainage pipe network problems based on big data analysis, including the following steps:

[0075] S1. Data Acquisition and Preprocessing: Sensor data, historical maintenance records, and pipeline geographic information are obtained from various data sources. A key task in the data acquisition process is ensuring the timeliness, accuracy, and completeness of the data to avoid sensor malfunctions or data omissions. Data preprocessing ensures the data is suitable for subsequent analysis and modeling, preventing misdiagnosis due to data quality issues.

[0076] Existing technologies for data acquisition and preprocessing largely rely on single-type data sources (e.g., using only flow or water level data) for problem diagnosis, and typically only perform simple data cleaning and preprocessing. Many systems fail to comprehensively consider the fusion of multi-source data, and lack intelligent anomaly detection and supplementation strategies during preprocessing. Therefore, the quality and comprehensiveness of data acquisition are often limited, leading to lower accuracy in diagnostic results. This system, however, combines multi-dimensional data sources (flow, water level, water quality, historical maintenance records, etc.) and performs comprehensive standardization processing, better addressing data quality issues such as missing value imputation and noise removal, thus providing high-quality, reliable data support for subsequent analysis and diagnosis. This comprehensive and intelligent data preprocessing approach significantly improves the accuracy and robustness of subsequent algorithms.

[0077] In this embodiment, the collected sensor data includes flow rate, water level, and water quality;

[0078] Preprocessing of sensor data includes missing value imputation, noise removal, outlier detection, and data standardization to obtain standardized sensor data. The standardized formula is as follows:

[0079]

[0080] in,

[0081] The raw sensor data X collected, and These are the mean and standard deviation of the original sensor data X, respectively;

[0082] Standardized sensor data This process eliminates the influence of different units in the raw sensor data, allowing data with different characteristics to be analyzed on the same scale. Thus, when different types of data (such as flow rate and water quality) are input into subsequent diagnostic models, the influence between different data sources is balanced, avoiding errors caused by different units or data biases. This processing method is particularly crucial in big data analysis and machine learning algorithms, helping to improve the accuracy and stability of the model. The output data serves as input for subsequent steps, laying the foundation for feature engineering, pattern recognition, and anomaly detection, ensuring the accuracy and robustness of subsequent algorithms. Therefore, standardization is not only a core operation of data preprocessing but also key to ensuring the efficient operation of the model.

[0083] S2. Extract key features from data from different data sources, perform feature fusion, and construct feature vectors.

[0084] In existing technologies, feature extraction often focuses on a single data source, such as feature engineering based solely on flow or water level data. This makes traditional methods prone to biasing diagnostic results towards anomalous patterns from a single data source when facing complex, multi-dimensional problems, failing to comprehensively capture the entire system. In contrast, this system combines multiple data sources, including flow, water level, water quality, historical maintenance records, and pipeline geographic information, enabling comprehensive feature extraction from multiple dimensions. It considers not only time-series features but also spatial features, thereby improving the accuracy and depth of diagnosis. By weighted fusion of different features, the system can adaptively adjust based on the importance of different features, achieving a reasonable balance of the influence of multi-source data. This overcomes the limitations of a single data source and enhances the system's ability to detect complex anomaly patterns.

[0085] Features that reflect the operating status of the drainage network are extracted from sensor data, historical maintenance records, and geographic information data of the pipeline network obtained from different data sources.

[0086] Time series features, including mean, variance, maximum and minimum values, are extracted from sensor data. Spatial features, including pipeline aging, maintenance frequency and regional distribution, are extracted from historical maintenance records and pipeline geographic information.

[0087] Feature fusion is performed on the extracted features to integrate features from different data sources into a unified feature vector. ;

[0088] Based on the standardized sensor data obtained in step S1 Historical maintenance records and pipeline geographic information data Then the eigenvector It is constructed by weighted fusion of different features, and the formula is as follows:

[0089]

[0090] in,

[0091] These are weighting coefficients, representing the contribution of sensor data to the feature vector;

[0092] These are weighting coefficients, representing the contribution of historical maintenance records to the feature vector;

[0093] These are weighting coefficients, representing the contribution of pipeline geographic information data to the feature vector;

[0094] The three weighting coefficients can be learned through training using a minimum error optimization algorithm (such as the minimum mean square error method) based on historical diagnostic results to minimize the bias of the comprehensive feature vector towards the known problem classification label. This ensures the final feature vector... This weighted fusion strategy reflects the combined influence of different data sources, preventing any single feature from dominating the diagnostic results. It enables the model to handle complex multi-source data and ensures synergy between data sources, thus providing more comprehensive and accurate pipeline diagnostic information.

[0095] Received It is a feature vector that integrates multi-source data and will be used as input for subsequent steps.

[0096] S3. Based on feature vectors, the next step is to use big data mining and analysis algorithms to identify potential abnormal patterns in the drainage network, such as blockages and leaks, and to build a diagnostic model based on machine learning algorithms to identify whether there are abnormalities in the drainage network.

[0097] Traditional drainage network anomaly identification mainly relies on manual inspections or static threshold judgments, such as triggering alarms by setting a single variable (e.g., water level exceeding a fixed value). This approach not only has limited ability to identify complex anomalies but also cannot adapt to changes in the dynamic operating environment. In step S3, this system fuses feature vectors... Building upon this foundation, machine learning algorithms (such as random forests, support vector machines, and deep neural networks) are used to learn the complex nonlinear boundaries between "normal" and "abnormal" conditions from historical and real-time data, significantly improving the ability to identify hidden or sudden problems (such as localized blockages or intermittent leaks in pipelines). This method possesses adaptive and self-learning capabilities, dynamically adjusting the identification logic, which is significantly superior to the static and rigid judgment mechanism of traditional threshold methods.

[0098] In this embodiment, based on Apply machine learning algorithms to identify the presence of anomalies;

[0099] Anomalies include pipe network blockage, abnormal water flow, excessively high or low water levels, and pipe leaks. A diagnostic model is constructed, and algorithms are used to uncover the relationship between features and anomalies, thereby identifying anomalies in real-time data and outputting diagnostic results. This indicates whether an anomaly exists.

[0100] The diagnostic model is shown below:

[0101]

[0102] in,

[0103] The threshold value represents the boundary between normal and abnormal conditions.

[0104] This is a diagnostic module built using machine learning networks.

[0105] The diagnostic module includes a sub-model layer, a fusion layer, and a discrimination layer;

[0106] In the sub-model layer, for each key feature Use historical time series data to assess the prediction error (e.g., using ARIMA or Prophet models), and define local anomaly scores:

[0107]

[0108] in,

[0109] The predicted value at time t;

[0110] This represents the standard deviation of the error for that dimension (such as flow rate, water level, water quality, etc.) in the training set.

[0111] The standardized anomaly score for the i-th key feature;

[0112] In the fusion layer, the local anomaly scores of all key features are weighted according to their importance to obtain the total diagnostic score:

[0113]

[0114] in,

[0115] ,satisfy This was obtained through historical diagnostic correlation analysis;

[0116] In this embodiment, if it is found that changes in flow rate and water level are related to the blockage height, then the corresponding... A high value can be set.

[0117] In the discriminant layer, a threshold is set. The maximum Youden exponent is determined by taking the training data under the ROC curve.

[0118] In addition, based on big data analysis methods, the classification of anomaly types (blockage, leakage, water quality anomalies) can be refined. Supervised learning classifiers (such as random forests) can be used to further analyze the anomaly results.

[0119] Compared to building models based solely on single-dimensional data, fusion This significantly improves identification capabilities. For example, in some cases, flow and water level data may appear normal, but by combining historical maintenance frequency and location information of low-lying areas, the system may determine that there is a hidden risk of blockage, thus achieving a higher level of intelligent diagnosis. The diagnostic results output in this step... This will serve as the core input for subsequent trend forecasting.

[0120] S4. After anomaly identification, predict the occurrence trend of potential problems based on historical data and current operating status. By predicting possible future failures, urban drainage network managers can take proactive measures to intervene. The output of this stage will provide a quantitative basis for the formulation of maintenance strategies, ensuring the long-term healthy operation of the drainage system.

[0121] Problem prediction uses big data analytics to combine historical and real-time data to predict potential anomalies or failures in the drainage network at a future point in time or within a specific time period. To achieve accurate trend analysis, a time-series-based prediction model needs to be established, fully considering factors such as the network's long-term operational patterns, seasonal variations, and unforeseen events.

[0122] Known historical data and diagnostic results The prediction model is shown below, which forecasts the trend of potential problems occurring at time T in the future:

[0123]

[0124] in,

[0125] To predict the trend of potential problems;

[0126] For prediction algorithms, common algorithms include ARIMA (Autoregressive Integral Moving Average) and LSTM (Long Short-Term Memory). These algorithms can capture trends, cycles, and seasonal variations in time series data, thereby predicting the future state of the system.

[0127] For example, historical data and diagnostic results The input is fed into a trained LSTM model, where the model retains key state evolution information from past inputs through a memory mechanism, thereby predicting what might happen at time T in the future. This prediction result can not only indicate which areas of the pipeline network have potential risks, but also quantitatively represent the strength of the risk trend, which can be used for the formulation of the next stage of maintenance strategy.

[0128] The model can predict potential future problems and provide early warning signals for maintenance work.

[0129] This method not only integrates dynamic evolution of information across multiple time periods and variables, but also establishes temporal dependencies between data through deep learning models (such as LSTM), thereby predicting the location and risk level of potential future faults. This achieves a technological leap from "response-based repair" to "predictive operation and maintenance." This mechanism can provide forward-looking decision-making basis for the maintenance and scheduling of urban drainage systems, improving resource allocation efficiency and risk prevention capabilities.

[0130] S5. After predicting the occurrence trend of potential problems, an optimization model is built for maintenance. Based on the problem prediction results, a specific maintenance plan is generated and optimized to achieve optimal resource allocation and efficient management of the drainage system.

[0131] Most existing drainage system maintenance schemes are based on periodic inspection systems or passive response models, lacking a dynamic strategy generation mechanism that deeply integrates with risk prediction results. Such methods are prone to resource waste (such as over-inspection) or maintenance lag (such as failure to respond promptly to sudden failures).

[0132] The maintenance strategy generated in this invention needs to be based on prediction results, resource availability, and the specific conditions of the drainage network (such as the degree of pipe aging and maintenance history). By combining prediction data, specific maintenance strategies can be formulated for different types of problems (such as blockages and leaks), and the optimal configuration of the strategy can be achieved through optimization. For example, optimization algorithms such as genetic algorithms and simulated annealing can be used to find the optimal maintenance time, maintenance personnel allocation, and equipment scheduling scheme. In addition, the maintenance strategy should also consider risk assessment, that is, quantitatively analyze the risk of potential problems and formulate corresponding response measures according to the risk level.

[0133] In this embodiment, the optimization model is as follows:

[0134]

[0135] in,

[0136] For optimized maintenance strategies;

[0137] This represents the cost of performing the i-th maintenance task under strategy S;

[0138] It is the predicted value for the i-th type of potential problem;

[0139] It is the resource constraint under strategy S;

[0140] It is a regulatory factor used to balance cost and resource constraints;

[0141] This model is designed to minimize maintenance costs and resource waste while maximizing the efficiency of pipeline fault repair.

[0142] The system first retrieves the prediction results from step S4. This includes the probability and distribution trend of each type of problem in different regions. Then, this is compared with the data in the database maintained within the system. and resource information By combining these factors, an objective function is constructed. The optimization engine (such as simulated annealing) seeks the maintenance strategy that minimizes the objective function while satisfying the total resource constraints. This strategy can include specific scheduling decisions such as "when to maintain, which area to maintain, how many resources to use, and what kind of repair solution to adopt".

[0143] This method avoids the inefficiencies of traditional methods that allocate resources evenly by administrative region or based on experience. Instead, it dynamically allocates maintenance resources based on risk. For example, when a significant increase in leakage risk is predicted in a certain area, the model can automatically prioritize the allocation of limited resources to that area while reducing resource investment in low-risk areas, thus achieving a maintenance strategy with optimal global cost and shortest response time.

[0144] The system also includes the following functional modules: acquisition and preprocessing module, feature fusion module, anomaly diagnosis module, state prediction module, and strategy optimization module. These modules are constructed sequentially according to steps S1-5 above.

[0145] Example 2:

[0146] This embodiment also provides a computer device applicable to a problem diagnosis system for urban underground drainage pipe networks based on big data analysis, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the problem diagnosis system for urban underground drainage pipe networks based on big data analysis as proposed in the above embodiment.

[0147] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements a problem diagnosis system for urban underground drainage pipe network based on big data analysis as proposed in the above embodiment.

[0148] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0149] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0150] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0151] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0152] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0153] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A problem diagnosis system for urban underground drainage pipe networks based on big data analysis, characterized in that, Includes the following steps: S1. Data acquisition and preprocessing: obtaining sensor data, historical maintenance records, and pipeline geographic information data from different data sources; S2. Extract key features from data from different data sources, perform feature fusion, and construct feature vectors; S3. Based on feature vectors, construct a diagnostic model to identify whether there are abnormalities in the drainage network using machine learning algorithms; S4. After completing the anomaly identification, predict the occurrence trend of potential problems based on historical data and current operating status; S5. After predicting the trend of potential problems, build and maintain an optimization model; In step S3, based on Apply machine learning algorithms to identify the presence of anomalies; Anomalies include pipe network blockage, abnormal water flow, excessively high or low water levels, and pipe leaks. A diagnostic model is constructed, and algorithms are used to uncover the relationship between features and anomalies, thereby identifying anomalies in real-time data and outputting diagnostic results. This indicates whether an anomaly exists; The diagnostic model is shown below: ； in, The threshold value represents the boundary between normal and abnormal conditions. A diagnostic module built using machine learning networks; The diagnostic module includes a sub-model layer, a fusion layer, and a discrimination layer; In the sub-model layer, for each key feature The prediction error is calculated using historical time series data, and a local anomaly score is defined. ； in, The predicted value at time t; The standard deviation of the error of the key features in the training set; The standardized anomaly score for the i-th key feature; In the fusion layer, the local anomaly scores of all key features are weighted according to their importance to obtain the total diagnostic score: ； in, ,satisfy This was obtained through historical diagnostic correlation analysis; In the discriminant layer, a threshold is set. The maximum Youden exponent is determined by taking the training data under the ROC curve; In step S5, the optimized model is as follows: ； in, For optimized maintenance strategies; This represents the cost of performing the i-th maintenance task under strategy S; It is the predicted value for the i-th type of potential problem; It is the resource constraint under strategy S; It is a regulatory factor used to balance cost and resource constraints; This model is designed to minimize maintenance costs and resource waste while maximizing the efficiency of pipeline fault repair.

2. The urban underground drainage network problem diagnosis system based on big data analysis according to claim 1, characterized in that, In step S1, the collected sensor data includes flow rate, water level, and water quality. Preprocessing of sensor data includes missing value imputation, noise removal, outlier detection, and data standardization to obtain standardized sensor data. The standardized formula is as follows: ； in, The raw sensor data X collected, and These are the mean and standard deviation of the original sensor data X, respectively; Standardized sensor data It can eliminate the influence between different dimensions in the raw sensor data, enabling data with different characteristics to be analyzed on the same scale.

3. The urban underground drainage network problem diagnosis system based on big data analysis according to claim 1, characterized in that, In step S2, features that can reflect the operating status of the drainage network are extracted from sensor data, historical maintenance records and pipeline geographic information data obtained from different data sources. Time series features, including mean, variance, maximum and minimum values, are extracted from sensor data. Spatial features, including pipeline aging, maintenance frequency and regional distribution, are extracted from historical maintenance records and pipeline geographic information.

4. The urban underground drainage network problem diagnosis system based on big data analysis according to claim 3, characterized in that, Feature fusion is performed on the extracted features to integrate features from different data sources into a unified feature vector. ; Based on the standardized sensor data obtained in step S1 Historical maintenance records and pipeline geographic information data Then the eigenvector It is constructed by weighted fusion of different features, and the formula is as follows: ； in, These are weighting coefficients, representing the contribution of sensor data to the feature vector; These are weighting coefficients, representing the contribution of historical maintenance records to the feature vector; These are weighting coefficients, representing the contribution of pipeline geographic information data to the feature vector; Received It is a feature vector that integrates multi-source data and will be used as input for subsequent steps.

5. The urban underground drainage network problem diagnosis system based on big data analysis according to claim 1, characterized in that, In step S4, historical data is known. and diagnostic results The prediction model is shown below, which forecasts the trend of potential problems occurring at time T in the future: ； in, To predict the trend of potential problems; The LSTM algorithm is used to predict the algorithm function; The model can predict potential future problems and provide early warning signals for maintenance work.

6. The urban underground drainage network problem diagnosis system based on big data analysis according to claim 1, characterized in that, The system also includes the following functional modules: acquisition and preprocessing module, feature fusion module, anomaly diagnosis module, state prediction module, and strategy optimization module.

Images