An intelligent monitoring and analysis method and system for gate opening and closing operating condition data

Abstract

The present invention relates to the technical field of system database management, and specifically to a method and system for intelligent monitoring and analysis of gate opening and closing condition data, comprising obtaining fault records from a historical fault data repository of a specified gate and constructing a fault feature library; obtaining historical monitoring data of a sensor group and performing pattern recognition and statistical analysis to form a fault feature weight table, which is stored in a data cleaning database; utilizing the fault feature weight table, external prediction data, and sensor installation location data, to classify the sensor group through database indexing technology, and storing the classification results in a sensor-level index table. Furthermore, a distributed data intelligent monitoring system is formed by combining N independent data processing units with a database interface protocol. The data processing unit includes a data acquisition layer, a real-time data analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer to ensure efficient data processing and retrieval.

Description

Technical Field

[0001] The present invention relates to the technical field of system database management, and in particular to an intelligent monitoring and analysis method and system for gate opening and closing operating condition data. Background Art

[0002] The core technology of intelligent monitoring and analysis methods for gate opening and closing condition data lies primarily in three aspects: data acquisition, data processing, and intelligent analysis. At the data acquisition level, multiple sensors are typically used in a collaborative manner to comprehensively capture key parameters during gate operation. For example, displacement sensors are used to monitor changes in gate opening, speed sensors are used to capture gate opening and closing speeds, torque sensors are used to measure the output torque of the drive mechanism, vibration sensors are used to detect abnormal vibrations in the gate structure, and temperature sensors are used to monitor temperature rises in key components. These sensors form a distributed or centralized monitoring network that transmits real-time data to a data processing center via wired or wireless communication technologies. At the data processing level, the collected raw data must first be preprocessed, including data cleaning, noise filtering, outlier processing, and data normalization, to ensure data quality and analysis accuracy.

[0003] Based on the gate's structural characteristics and operating mechanism, feature engineering is performed to extract key features that effectively characterize its operating status. These features include statistical characteristics (mean, variance, maximum, minimum, etc.), time-domain characteristics (kurtosis, crest factor, etc.), frequency-domain characteristics (spectral energy, dominant frequency, etc.), and energy characteristics. These features serve as important inputs for intelligent analysis models. At the intelligent analysis level, artificial intelligence algorithms such as machine learning and deep learning are typically used to construct intelligent analysis models for gate operating status. For example, algorithms such as support vector machines (SVM), random forests (RF), and artificial neural networks (ANN) can be used to construct gate status classification models, enabling automatic identification and classification of gate operating states such as normal, abnormal, and faulty. Alternatively, time series analysis models (such as ARIMA and LSTM) or regression models (such as SVR and GBRT) can be used to construct gate performance prediction models to predict key indicators such as the gate's remaining service life and future operating trends, providing a basis for predictive maintenance. Furthermore, hybrid intelligent analysis systems can be constructed by combining expert knowledge with rule-based reasoning to further enhance the accuracy and reliability of analysis.

[0004] To fully understand the true operating conditions of gates during opening and closing, multiple types of sensors are required to collect real-time data on displacement, water level, motor current and voltage, vibration, load stress, and river flow fluctuations. However, in the field of database management technology, data integration is challenging due to the diversity of data formats and communication protocols. Existing systems lack comprehensive mechanisms for database planning and data cleansing, making it difficult to efficiently retrieve and utilize heterogeneous data. Furthermore, current monitoring systems often issue warnings based on simple thresholds, lacking comprehensive analysis and in-depth assessment of multi-source heterogeneous data, making it difficult to promptly identify potential faults and conduct real-time trend forecasting. In practical gate applications, the management and integration of cross-source data remains challenging: in a multi-source heterogeneous data environment, data from different sensors cannot be effectively integrated, resulting in a lack of accurate and real-time analysis of gate safety, and consequently, the inability to promptly identify potential faults and risks, compromising gate safety and decision-making efficiency.

[0005] Therefore, an intelligent monitoring and analysis method and system for gate opening and closing condition data are proposed. Summary of the Invention

[0006] The present invention aims to provide a method and system for intelligent monitoring and analysis of gate opening and closing condition data. The method comprises obtaining fault records from a historical fault data repository for a specified gate and constructing a fault signature library based on these records. The fault signature library is organized using a database structure and supports retrieval and update operations. A fault signature weight table is formed by obtaining historical monitoring data from a sensor group and performing pattern recognition and statistical analysis. The data is stored in a data cleaning database that supports data storage, updating, and retrieval. Using the fault signature weight table, external prediction data, and sensor installation location data, the sensor group is classified using database indexing technology, and the classification results are stored in a sensor-level index table that uses a hierarchical index structure. The method also implements real-time data acquisition and processing through N independent data processing units, combined with a database interface protocol, to form a distributed data intelligent monitoring system. The data processing units include a data acquisition layer, a real-time data analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer to ensure efficient data processing and retrieval.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An intelligent monitoring and analysis method for gate opening and closing condition data, comprising:

[0009] Obtaining fault records from a historical fault data repository of a specified gate and constructing a fault signature database based on the fault records; the fault signature database is organized in a database structure and supports retrieval and update operations;

[0010] Acquiring historical sensor monitoring data collected by the sensor group; performing pattern recognition and statistical analysis on the historical sensor monitoring data to form a fault feature weight table, and storing the historical sensor monitoring data and the fault feature weight table in a data cleaning database, wherein the data cleaning database supports data storage, updating, and retrieval;

[0011] Based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data of the sensor group, the sensor group is classified using a database indexing technology to obtain a key sensor group, a warning sensor group, and a conventional sensor group; and the classification results are written into a sensor-level index table stored in a data cleaning database through a database interface. The sensor-level index table adopts a hierarchical index structure and supports retrieval and update.

[0012] Within a specified time period, real-time data collection is performed through N independent data processing units according to the sensor level index table; the data processing units exchange data through a database interface protocol, and together constitute a distributed data intelligent monitoring system;

[0013] The data processing unit includes a data acquisition layer, a data real-time analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

[0014] Furthermore, constructing the fault feature library according to the fault record is specifically as follows:

[0015] The fault record includes a fault type and a fault triggering condition; and the fault feature library is constructed for different fault types and fault triggering conditions.

[0016] Furthermore, the sensor group is used to monitor various working parameters of the designated gate in real time; the sensor group includes a displacement sensor, a water level sensor, a flow sensor, a current sensor, a voltage sensor, a vibration sensor and a load stress sensor.

[0017] Furthermore, the sensor group is specifically:

[0018] The sensor group is composed of a plurality of independent type sensors; wherein, the displacement sensor is composed of a plurality of independent displacement sensors, which are respectively arranged at the first position, the second position and the third position of the gate leaf of the designated gate, and are used to monitor the vertical displacement of the gate leaf of the designated gate, and obtain the first position displacement data, the second position displacement data and the third position displacement data after collection; the water level sensor includes an upstream water level sensor and a downstream water level sensor; wherein, the upstream water level sensor is located in the upstream water area of ​​the designated gate, at a first specified distance from the designated gate, at a second specified depth below the water surface, and is used to monitor the upstream water level of the designated gate; the downstream water level sensor is located in the downstream water area of ​​the designated gate, at a third specified distance from the designated gate, at a fourth specified depth below the water surface, and is used to monitor the downstream water level of the designated gate;

[0019] The flow sensor is arranged in the discharge area of ​​the designated gate, and is used to monitor the flow of the designated gate; the current sensor and the voltage sensor are respectively installed in the motor control circuit of the designated gate hoist, and are used to monitor the current and voltage of the designated gate; the vibration sensor is arranged on the supporting structure of the designated gate, and is used to monitor the vibration of the designated gate;

[0020] The sensor group provides data for the distributed data intelligent monitoring system, and each of the data processing units is independently configured according to the corresponding sensor type.

[0021] Furthermore, classifying the sensor group according to the fault feature library, the fault feature weight table, the external prediction data, and the installation position data of the sensor group includes the following steps:

[0022] Extracting an original feature set based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data; the original feature set includes: sensor working state data, time series features, sensor spatial location information, and real-time working condition data;

[0023] An additional feature set is constructed based on the fault feature weight table, the external prediction data, and the installation location data; the additional feature set includes: fault frequency features, fault mode features, installation location distance features, spatial coordination features, environmental response sensitivity features, and sensor stability features; the environmental response sensitivity features include wind speed response features, wind direction response features, river hydrological response features, and extreme weather response features; the sensor stability features include data fluctuation amplitude features, response time features, and repeatability features;

[0024] The original feature set and the additional feature set are subjected to feature fusion to form a comprehensive feature vector; the comprehensive feature vector is subjected to standardization to obtain a standard comprehensive feature vector;

[0025] Input the standard comprehensive feature vector into the trained XGBoost model and output the classification result;

[0026] The sensor group is divided into the key sensor group, the warning sensor group and the regular sensor group according to the classification result.

[0027] Furthermore, the specified time period specifically refers to the duration of the next opening operation or closing operation of the specified gate.

[0028] Furthermore, the external prediction data includes environmental prediction data and hydrological prediction data within the specified time period in the future.

[0029] Furthermore, the data processing unit is specifically:

[0030] The data acquisition layer acquires real-time monitoring data through corresponding independent type sensors, and the data real-time analysis layer performs real-time analysis on the real-time monitoring data to obtain normal real-time data and abnormal real-time data;

[0031] The data cleaning and quality control layer performs data correction on the abnormal real-time data to obtain corrected real-time data;

[0032] The data storage and retrieval layer uploads the normal real-time data and the corrected real-time data to the database system for storage, and supports data query and analysis operations through a retrieval interface. The database system adopts a multi-dimensional index structure and supports multi-condition retrieval and data processing.

[0033] Furthermore, the data processing unit further includes an early warning module, which is configured to trigger an alarm mechanism and send alarm information to a corresponding coupled data processing unit when performing real-time analysis on the real-time monitoring data and obtaining the abnormal real-time data. The determination of the coupled data processing unit includes:

[0034] The sensor historical monitoring data and the fault records are analyzed to identify all potential coupling relationships among the N independent data processing units; and the coupled data processing unit is determined based on the correlation between the abnormal real-time data and the potential coupling relationships.

[0035] An intelligent monitoring and analysis system for gate opening and closing condition data, comprising:

[0036] A fault data acquisition module is used to obtain fault records from the historical fault data repository of a specified gate and construct a fault feature library based on the fault records; the fault feature library is organized in a database structure and supports retrieval and update operations;

[0037] A sensor data processing unit, configured to obtain historical sensor monitoring data collected by the sensor group;

[0038] a fault pattern recognition and feature extraction module, configured to perform pattern recognition and statistical analysis on the sensor historical monitoring data to form a fault feature weight table, and store the sensor historical monitoring data and the fault feature weight table in a data cleaning database, wherein the data cleaning database supports data storage, updating, and retrieval;

[0039] An external prediction data acquisition module is used to acquire external prediction data;

[0040] a sensor classification module for classifying the sensor group using database indexing technology based on the fault feature weight table, the external prediction data, and the installation location data of the sensor group to obtain a critical sensor group, a warning sensor group, and a conventional sensor group; and writing the classification results into a sensor-level index table stored in a data cleaning database through a database interface. The sensor-level index table adopts a hierarchical index structure and supports retrieval and update;

[0041] The data acquisition and real-time analysis module is used to perform real-time data acquisition through N independent data processing units according to the sensor-level index table within a specified time period; the data processing units exchange data through a database interface protocol, together forming a distributed data intelligent monitoring system; the data processing units include a data acquisition layer, a real-time data analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. This invention utilizes a sensor array comprised of multiple sensors to achieve comprehensive, real-time monitoring of all gate operating parameters. This multi-dimensional sensor combination, encompassing displacement, water level, flow, current, voltage, vibration, and stress, not only provides accurate operating status data but also effectively reflects the gate's operating conditions and potential faults. This multi-sensor monitoring approach significantly improves the integrity and accuracy of data acquisition, providing reliable data support for subsequent fault analysis, prediction, and maintenance. Furthermore, the combination of a fault signature library, classification algorithms, and data cleansing helps improve the intelligence and processing efficiency of the monitoring system, reducing operating and maintenance costs and ensuring the safe and efficient operation of the gate. This invention innovatively constructs a multi-source, heterogeneous sensing network for the full lifecycle management of gates. It utilizes six sensor types: displacement, water level, flow, current and voltage, vibration, and stress, topologically deployed based on the dynamic characteristics of the gate body. Redundant monitoring nodes are deployed in key stress concentration areas, such as the gate leaf load-bearing beam, support hinge structure, and gate hoist drive shaft. Synchronous data acquisition at 50Hz per second is achieved through a CAN bus and 4G / fiber dual-channel networking. In operation and maintenance practice, this architecture has improved the early identification rate of typical faults such as gate slot jamming, hoist bearing wear, hydraulic rod underpressure, and water stop leakage. Through multi-parameter correlation analysis, it can achieve quantitative diagnosis of internal leakage in the hydraulic system, reducing the number of false alarms by 75% compared with traditional single-point monitoring solutions.

[0044] 2. The present invention achieves comprehensive monitoring of various operating parameters by precisely placing different types of sensors at key locations on designated gates. These sensors, including displacement sensors, water level sensors, flow sensors, current and voltage sensors, and vibration sensors, can respectively collect important data such as gate leaf displacement, upstream and downstream water levels, flow, current, voltage, and vibration in real time, thereby ensuring accurate monitoring of the gate's operating conditions. The distribution and complementary functions of the various sensors make the monitoring data more comprehensive, reflecting all aspects of the gate's operating status and potential faults. By inputting this data into a distributed data intelligent monitoring system, the system can independently process and analyze it based on different sensor types, greatly improving the efficiency and accuracy of data acquisition, enhancing the intelligence level of fault warning and maintenance decision-making, thereby ensuring the safe and reliable operation of the gate and providing a solid data foundation for subsequent intelligent management and optimization. Actual operation data shows that fault prediction accuracy and fault diagnosis response time have also been optimized. The fault prediction accuracy of the experimental group reached 95%, the response time was shortened to 1 minute, and the fault location accuracy was less than ±15cm. Especially under high water level conditions during flood season, the vibration-stress-flow multi-parameter coupling analysis provides data support for formulating dynamic counterweight adjustment plans.

[0045] 3. This invention significantly improves the accuracy and reliability of sensor group classification by extracting and fusing multi-dimensional features from gate opening and closing condition data. By combining a fault feature weight table, external prediction data, and sensor installation location data, this method extracts original features such as sensor operating status, time series characteristics, and spatial location information, as well as additional features such as fault frequency, fault mode, and environmental response. This rich information helps comprehensively reflect the sensor's operating environment and performance. In particular, the introduction of environmental response sensitivity features accurately captures the impact of external environmental changes on sensor performance, enhancing fault prediction and classification capabilities. Through training and classification using an XGBoost model, sensors can be automatically classified into critical, alert, and routine groups, optimizing the intelligence level of the monitoring system. This provides more accurate data support and decision-making basis for equipment maintenance, fault warning, and condition monitoring, thereby improving the safety and stability of the entire monitoring system. Using the XGBoost ensemble learning algorithm (max_depth=6, learning_rate=0.05) to construct the classification model and determine weight assignment based on interpretability analysis, the classification accuracy of key sensors reached 98.7%. Operational maintenance applications have shown that the model can reduce the frequency of routine inspections by 40%, enable preventive maintenance decisions through equipment health scores, and extend the hydraulic system overhaul cycle from 2 years to 3.5 years. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart of an intelligent monitoring and analysis method for gate opening and closing condition data provided by an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of the installation position of the displacement sensor provided in an embodiment of the present invention;

[0048] Figure 3 This is a structural diagram of an intelligent monitoring and analysis system for gate opening and closing operating condition data provided by an embodiment of the present invention.

[0049] In the figure: 1, the door leaf of the designated gate; 2, the first opening sensor; 3, the second opening sensor; 4, the third opening sensor. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 making creative efforts are within the scope of protection of the present invention.

[0051] With the continuous deepening of informatization and intelligent construction of water conservancy projects, the operation and management of gates are gradually moving towards intelligence and refinement. However, the complexity of the gate opening and closing process, the diversity of environmental factors and the standardization of operation requirements make it critical to achieve real-time monitoring of multi-dimensional operating status and accurate fault diagnosis. According to relevant operating procedures, the opening and closing of gates must strictly implement the operation ticket system, and must be operated by certified personnel in accordance with the specifications. 24-hour duty is required during operation. During operation, the gates and opening and closing equipment should be inspected and monitored. If any abnormal situation is found (such as overload, jamming, tilt or abnormal noise), it is necessary to immediately stop the machine for inspection and record it in the work log. The inspection scope covers the gate structure, opening and closing machinery, gate chamber and its monitoring system to ensure that the operating equipment and system are in the best condition. These requirements provide a standardized basis for subsequent system data collection and intelligent analysis.

[0052] The gate's operating modes are divided into gravity diversion (drainage) mode and water lifting diversion (drainage) mode. Its opening and closing operation procedures need to be dynamically adjusted according to the upstream and downstream water level difference, flow demand and unit type. In gravity mode, the gate opening and closing sequence and opening degree need to be reasonably allocated according to the change in water level difference; in water lifting mode, the opening and closing process needs to be completed in stages in combination with the characteristics of the power frequency unit or variable frequency unit and the blade adjustment mode. Each mode has clear regulations on the gate opening degree, time interval and operation sequence. At the same time, the regulations must be strictly followed during operation inspections and shutdown operations to ensure safe and efficient operation. These operating processes provide a data foundation and logical framework for the analysis and optimization of multi-dimensional operating status.

[0053] As the complexity of gate operating environments continues to grow, real-time monitoring and troubleshooting of multi-dimensional operating conditions during the opening and closing process, along with effective organization of massive amounts of data through a robust database structure and data collection mechanism, have become crucial for ensuring system safety and efficient operation. Gate projects are often located in estuaries, facing not only common challenges such as exposure to sunlight and rain, vibration and impact, and siltation, but also the increased risk of loss from river corrosion and flow fluctuations. This can make sensor data susceptible to interference or drift, impacting the accuracy of real-time monitoring.

[0054] See also Figures 1 to 3 The present invention provides an intelligent monitoring and analysis method and system for gate opening and closing condition data, and the technical solution is as follows:

[0055] Example 1

[0056] To improve gate operation efficiency and safety, a pump station management company applied an intelligent monitoring and analysis method for gate opening and closing condition data, including:

[0057] See Figure 1In S10, a fault record is obtained from a historical fault data repository of a specified gate, and a fault feature library is constructed based on the fault record; the fault feature library is organized in a database structure and supports retrieval and update operations;

[0058] Furthermore, constructing the fault feature library based on the fault record is specifically as follows: the fault record includes the fault type and the fault triggering condition; and constructing the fault feature library for different fault types and fault triggering conditions. By constructing a fault feature library, especially combining the fault type with the fault triggering condition, a more accurate fault pattern recognition and prediction capability can be provided for the intelligent monitoring system. The establishment of a fault feature library enables the system to promptly identify potential fault risks during real-time monitoring and respond quickly based on the fault type and triggering conditions. In addition, the use of a fault feature library can also improve the efficiency of data analysis, optimize the classification and processing of sensor data, and thus improve the reliability and accuracy of the overall monitoring system.

[0059] See Figure 1 In S20, historical sensor monitoring data collected by the sensor group is obtained; pattern recognition and statistical analysis are performed on the historical sensor monitoring data to form a fault feature weight table, and the historical sensor monitoring data and the fault feature weight table are stored in a data cleaning database, where the data cleaning database supports data storage, updating, and retrieval;

[0060] Furthermore, the sensor suite is used to monitor various operating parameters of the designated gate in real time; it includes an opening sensor, a water level sensor, a flow sensor, a current sensor, a voltage sensor, a vibration sensor, a load sensor, and a lateral pressure stress sensor. These sensors comprehensively capture the gate's operating status, providing reliable data support for fault warnings, performance evaluation, and maintenance decisions. By integrating the monitoring capabilities of multiple sensor types, the system accurately monitors the gate's dynamic changes, ensuring safe and efficient operation, while also laying the foundation for system optimization and intelligent management.

[0061] Furthermore, the sensor group is specifically:

[0062] In this embodiment, all sensors constitute a sensor group, which is composed of multiple independent types of sensors (such as water level sensors and displacement sensors, etc.); each sensor type may contain multiple independent sensors to monitor different physical quantities or working parameters. Figure 2, the displacement sensor is installed on the gate leaf 1 of the designated gate; the displacement sensor is composed of a plurality of independent displacement sensors, including: a first opening sensor 2, a second opening sensor 3 and a third opening sensor 4; they are respectively arranged at the first position, the second position and the third position of the gate leaf of the designated gate, for monitoring the vertical displacement of the gate leaf of the designated gate, and the first position opening data, the second position opening data and the third position opening data are obtained after collection; the water level sensor includes an upstream water level sensor and a downstream water level sensor; wherein, the upstream water level sensor is located in the upstream water area of ​​the designated gate, at a first specified distance from the designated gate, at a second specified depth below the water surface, for monitoring the upstream water level of the designated gate; the downstream water level sensor is located in the downstream water area of ​​the designated gate, at a third specified distance from the designated gate, at a fourth specified depth below the water surface, for monitoring the downstream water level of the designated gate; the present invention can comprehensively monitor various key working parameters of the gate by accurately configuring multiple independent types of sensor groups. The displacement sensor tracks the vertical displacement of the gate blade in real time, accurately assessing the gate's opening and closing status. The water level sensor monitors upstream and downstream water level changes, effectively reflecting the impact and ensuring the gate's safe operation under varying water levels. The combination of these sensors provides multi-dimensional data support, enhancing real-time monitoring of the gate's operating status while also improving the system's early warning capabilities and fault detection accuracy.

[0063] In this embodiment, when the opening data for the first, second, and third positions are inconsistent, the real-time data analysis layer of the data processing unit first compares the displacement data for each of the three positions one by one, using error detection algorithms (such as standard deviation and deviation analysis) to determine whether the data for each position exceeds a set tolerance range. If the displacement data for a particular position deviates significantly from the other two and exceeds a preset threshold (for example, a deviation greater than 1 mm), the system marks it as abnormal data and triggers the early warning module, immediately sending an alert to the control system or operator. Upon receiving the early warning signal, the on-site operator performs manual measurement and calibration using the opening meter; concurrently, the data cleaning and quality control layer corrects the abnormal data. Correction methods can use interpolation algorithms (such as linear interpolation or polynomial interpolation) to estimate a reasonable value for the abnormal data, or extrapolate displacement values ​​based on similar scenarios in historical data to ensure data consistency and accuracy. For example, if the displacement data for the second position deviates significantly, the system will use the displacement data for the first and third positions for a weighted average, or correct the deviation based on a trend model of historical monitoring data. The corrected data will be saved again in the data storage and retrieval layer, and support subsequent query and analysis through the database interface to ensure the integrity and consistency of the system data and avoid erroneous judgments or control operations due to abnormal data.

[0064] In this embodiment, the water level sensor is located outside the gate vortex area. The distance of the water level sensor is determined according to the size of the gate. The specific distance L of the water level sensor below the water surface is gate height*10.

[0065] The flow sensor is installed in the discharge area of ​​the designated gate to monitor the flow of the designated gate. The current sensor and the voltage sensor are respectively installed in the motor control circuit of the designated gate hoist to monitor the current and voltage of the designated gate. The vibration sensor is arranged on the supporting structure of the designated gate to monitor the vibration of the designated gate. The load sensor is arranged on the gate hoist drum device to monitor the vertical load weight of the designated gate. The lateral pressure sensor is arranged on the side of the gate to monitor the horizontal lateral pressure of the designated gate.

[0066] The sensor group provides data for the distributed data intelligent monitoring system, and each of the data processing units is independently configured according to the corresponding sensor type.

[0067] See Figure 1 In S30, based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data of the sensor group, the sensor group is classified using a database indexing technology to obtain a key sensor group, a warning sensor group, and a regular sensor group; and the classification results are written into a sensor-level index table stored in a data cleaning database through a database interface. The sensor-level index table adopts a hierarchical index structure and supports retrieval and update.

[0068] Furthermore, classifying the sensor group according to the fault feature library, the fault feature weight table, the external prediction data, and the installation position data of the sensor group includes the following steps:

[0069] Extracting an original feature set based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data; the original feature set includes: sensor working state data, time series features, sensor spatial location information, and real-time working condition data;

[0070] An additional feature set is constructed based on the fault feature weight table, the external prediction data, and the installation location data; the additional feature set includes: fault frequency features, fault mode features, installation location distance features, spatial coordination features, environmental response sensitivity features, and sensor stability features; the environmental response sensitivity features include wind speed response features, wind direction response features, hydrological response features, and extreme weather response features; the sensor stability features include data fluctuation amplitude features, response time features, and repeatability features;

[0071] The original feature set and the additional feature set are subjected to feature fusion to form a comprehensive feature vector; the comprehensive feature vector is subjected to standardization to obtain a standard comprehensive feature vector;

[0072] Input the standard comprehensive feature vector into the trained XGBoost model and output the classification result;

[0073] According to the classification result, the sensor group is divided into the key sensor group, the warning sensor group and the regular sensor group. That is, in this embodiment, all sensors in the sensor group are classified into three different categories.

[0074] The present invention intelligently classifies sensor groups by combining a fault feature library, a fault feature weight table, external prediction data, and sensor installation location data. This allows accurate identification of the sensor's operating status and performance, providing an important basis for equipment fault warning and maintenance. By fusing the original feature set with the additional feature set, a comprehensive feature vector is generated, and classification is performed using the XGBoost model. This not only improves classification accuracy, but also enables timely identification of potential failure risks for key sensors. Dividing sensor groups into critical, warning, and conventional sensor groups enables refined monitoring and priority management, improving the system's fault response speed and data processing efficiency, and ensuring the safe and efficient operation of the gate.

[0075] See Figure 1 In S40, within a specified time period, real-time data is collected via N independent data processing units based on the sensor-level index table. The data processing units exchange data via a database interface protocol, collectively forming a distributed intelligent data monitoring system. Furthermore, the specified time period specifically refers to the duration of the next opening or closing operation of the specified gate. During the gate's opening and closing process, particularly at the start and stop moments, the device's operating status may experience significant fluctuations or anomalies. Therefore, limiting data collection to specific time periods is primarily intended to focus on critical moments in the opening and closing process. During these critical time periods, the data collected by the sensors best reflects whether the gate is operating properly, such as changes in displacement, current, and other indicators, allowing for timely detection of potential gate faults or anomalies. By focusing on data from these critical time periods, accurate monitoring and analysis can be performed in the shortest possible time, ensuring that faults are detected and addressed as early as possible while avoiding redundant data collection during non-critical time periods. This improves data processing efficiency and system response speed. This strategy effectively enhances fault detection sensitivity and ensures efficient system operation during critical gate operation moments.

[0076] The data processing unit includes a data acquisition layer, a data real-time analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

[0077] Furthermore, the external prediction data includes environmental prediction data and hydrological prediction data within the specified time period in the future. These prediction data can provide important external environmental background information for the gate opening and closing operations, helping the system to more accurately analyze and predict the gate's operating performance under specific conditions. For example, the impact of changes on the gate is directly related to the gate's opening and closing timing and operating mode. Temperature changes may affect the stability of mechanical components, and environmental prediction data can help foresee possible extreme weather conditions. By combining these external prediction data, the system can identify potential risk factors in advance, thereby optimizing and adjusting the gate before operating it, ensuring that the gate can still operate stably and safely when facing different environmental conditions, reducing the probability of failure, and improving the intelligence level and reliability of the entire system.

[0078] Furthermore, the data processing unit is specifically:

[0079] The data acquisition layer acquires real-time monitoring data through corresponding independent type sensors, and the data real-time analysis layer performs real-time analysis on the real-time monitoring data to obtain normal real-time data and abnormal real-time data;

[0080] The data cleaning and quality control layer performs data correction on the abnormal real-time data to obtain corrected real-time data;

[0081] The data storage and retrieval layer uploads the normal real-time data and the corrected real-time data to a database system for storage and supports data query and analysis operations through a retrieval interface. The database system adopts a multidimensional index structure and supports multi-condition retrieval and data processing. The data processing unit effectively improves data quality and reliability by comprehensively collecting, analyzing, and correcting real-time monitoring data. The data acquisition layer ensures accurate real-time data from various independent sensors, while the real-time data analysis layer promptly identifies and processes anomalies in the data, ensuring that the system can distinguish between normal and abnormal data. The correction process for abnormal data further eliminates measurement errors and noise, ensuring data accuracy. The data storage and retrieval layer, through a database system with a multidimensional index structure, achieves efficient storage, rapid retrieval, and flexible query analysis of massive amounts of data. It supports complex multi-condition retrieval operations and provides strong support for subsequent data analysis, prediction, and decision-making. This integrated process improves the intelligence and automation of data processing, not only improving data processing efficiency but also ensuring that the system can respond to various complex working conditions in a timely and accurate manner.

[0082] Furthermore, the data processing unit further includes an early warning module, which is configured to trigger an alarm mechanism and send alarm information to a corresponding coupled data processing unit when performing real-time analysis on the real-time monitoring data and obtaining the abnormal real-time data. The determination of the coupled data processing unit includes:

[0083] The sensor's historical monitoring data and fault records are analyzed to identify all potential coupling relationships among the N independent data processing units; the coupled data processing units are determined based on the correlation between the abnormal real-time data and the potential coupling relationships. By analyzing the monitoring data in real time and promptly identifying abnormal data, the early warning module can quickly trigger an alarm mechanism, promptly notifying operators to take necessary countermeasures and ensure the safe operation of the system. By analyzing the sensor's historical monitoring data and fault records, the system can identify potential coupling relationships between the data processing units. This enables the system to accurately determine the affected coupled data processing units by conducting in-depth analysis of the correlation between the abnormal data and the potential coupling relationships when an anomaly occurs. This not only improves the accuracy of the early warning, but also precisely locates the source of the fault, reduces unnecessary interference and false alarms, and thus more effectively ensures the stability and security of the system. Furthermore, the collaborative operation of the coupled data processing units enables rapid response and handling of anomalies, improving the overall level of intelligent monitoring.

[0084] The present invention constructs a fault feature library from historical fault data and combines it with real-time sensor data for pattern recognition and statistical analysis, which can accurately extract key fault features and abnormal signals, thereby improving the accuracy of fault prediction and risk identification. Secondly, by classifying sensors and assigning different priorities, the system rationally allocates data acquisition frequency, optimizes data processing efficiency, avoids information overload, and ensures priority processing of data from key sensors, thereby improving system response speed and real-time performance. In addition, by adopting database indexing technology and hierarchical indexing structure, the efficiency of data storage and retrieval has been significantly improved, supporting multi-condition retrieval and flexible data analysis, and providing strong support for subsequent decision-making. Ultimately, the intelligent data processing and real-time monitoring capabilities of this method can effectively reduce operating and maintenance costs and improve the safety and reliability of the gate.

[0085] Example 2

[0086] A water conservancy college uses gates as hydraulic engineering equipment to carry out experimental teaching and scientific research projects. By establishing a gate experimental platform, students can not only gain a deep understanding of the design principles and working mechanisms of the gates, but also personally operate and debug related monitoring and control systems, gaining real engineering practical experience. A water conservancy college has applied an intelligent monitoring and analysis system for gate opening and closing operating condition data, such as Figure 3 Shown, including:

[0087] A fault data acquisition module is used to obtain fault records from the historical fault data repository of a specified gate and construct a fault feature library based on the fault records; the fault feature library is organized in a database structure and supports retrieval and update operations;

[0088] A sensor data processing unit, configured to obtain historical sensor monitoring data collected by the sensor group;

[0089] a fault pattern recognition and feature extraction module, configured to perform pattern recognition and statistical analysis on the sensor historical monitoring data to form a fault feature weight table, and store the sensor historical monitoring data and the fault feature weight table in a data cleaning database, wherein the data cleaning database supports data storage, updating, and retrieval;

[0090] An external prediction data acquisition module is used to acquire external prediction data;

[0091] a sensor classification module for classifying the sensor group using database indexing technology based on the fault feature weight table, the external prediction data, and the installation location data of the sensor group to obtain a critical sensor group, a warning sensor group, and a conventional sensor group; and writing the classification results into a sensor-level index table stored in a data cleaning database through a database interface. The sensor-level index table adopts a hierarchical index structure and supports retrieval and update;

[0092] The data acquisition and real-time analysis module is used to perform real-time data acquisition through N independent data processing units according to the sensor-level index table within a specified time period; the data processing units exchange data through a database interface protocol, together forming a distributed data intelligent monitoring system; the data processing units include a data acquisition layer, a real-time data analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

[0093] When the system is running, it can implement all the steps described in the first embodiment.

[0094] The sensor historical monitoring data is collected by a sensor group, which includes: a displacement sensor for monitoring the displacement changes of a specified gate, a water level sensor for monitoring the water level changes, a flow sensor for monitoring the flow, a current sensor and a voltage sensor for monitoring the operating status of the motor, a vibration sensor for monitoring the vibration of the specified gate, and a load stress sensor for monitoring the load stress of the specified gate.

[0095] In addition, environmental prediction data such as wind speed, wind direction, and temperature prediction data are used to assist the system in analyzing and predicting the working status of the gate.

[0096] Environmental forecast data includes wind speed forecast data, wind direction forecast data, river hydrological forecast data, and extreme weather forecast data; hydrological forecast data includes temperature change forecast, daytime and nighttime temperature difference forecast, extreme temperature forecast, and forecast of the relationship between air temperature and water temperature;

[0097] Among them, the real-time operating condition data collected by the sensor group in real time includes real-time displacement data, real-time water level data, real-time flow data, real-time current data, real-time voltage data and real-time load stress data; the time series characteristics are statistical information of the real-time operating condition data, including mean, standard deviation, maximum value, minimum value, root mean square, skewness, kurtosis and periodicity; the sensor spatial position information refers to the installation position of the sensor, including the specific position of the corresponding sensor on the gate; the sensor's working status data includes start-stop status and working mode.

[0098] The XGBoost model is a three-category model trained using a historical fault feature weight table, external prediction data, and installation location data. The historical data is labeled by experts. The expert labeling process first collects historical monitoring data from all sensors, including fault records, sensor operating status, external environmental data, and sensor installation location information. Experts analyze multiple dimensions, including the sensor's impact on gate opening and closing conditions, fault history, and external environmental factors. Combining actual operational requirements and experience, they assign a category label (critical sensor group, warning sensor group, and conventional sensor group) to each sensor. The critical sensor group is the high-risk group, the warning sensor group is the medium-risk group, and the conventional sensor group is the low-risk group. The high-risk group is crucial to the operation and safety of the gate and requires priority handling in the event of a fault. The medium-risk group has a certain impact on the gate but does not immediately cause a fault. Failures in the low-risk group have minimal impact on the system and generally have little impact on normal operation.

[0099] The specific steps of performing real-time data collection through N independent data processing units according to the sensor level index table are as follows:

[0100] Based on the classification results of the critical sensor group, warning sensor group, and regular sensor group in the sensor level index table, different sampling priorities are assigned to each sensor group. The critical sensor group has the highest priority and its sampling frequency is set to the first sampling frequency; the warning sensor group has the second highest priority and its sampling frequency is set to the second sampling frequency, which is lower than the first sampling frequency; the regular sensor group has the lowest priority and its sampling frequency is set to the third sampling frequency, which is lower than the second sampling frequency. Based on the sampling priorities, N independent data processing units collect real-time monitoring data from the corresponding sensor groups according to the assigned sampling frequencies.

[0101] In this embodiment, a set of comparative experiments was also conducted. The experimental group used the intelligent monitoring and analysis system for gate opening and closing operating condition data proposed by the present invention. The differences between the control group and the experimental group are as follows:

[0102] Referring to Table 1, by comparing the experimental data, the experimental group (the intelligent monitoring and analysis system of the present invention) outperformed the control group (traditional sensor arrangement) in multiple key performance indicators. First, the experimental group showed accurate detection performance in vertical displacement monitoring, while the traditional method did not cover this dimension, and the accuracy of water level and flow monitoring was also significantly improved, thereby providing more accurate data support for the experimental group. Secondly, the data acquisition frequency and system response time of the experimental group were greatly improved, and it was able to monitor changes in gate working conditions in real time and respond quickly. The fault prediction accuracy and fault diagnosis response time have also been optimized. The fault prediction accuracy of the experimental group is as high as 95%, and the response time is shortened to 1 minute, which significantly improves the system's early warning capability and operation and maintenance efficiency. In summary, the intelligent monitoring and analysis system of the experimental group shows significant advantages in accuracy, real-time performance, fault prediction and diagnosis, and effectively improves the safety and operation efficiency of the gate.

[0103] Table 1. Comparative experimental results data table

[0104]

[0105]

[0106] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring and analysis method for gate opening and closing condition data, characterized in that: include: Obtaining fault records from a historical fault data repository of a designated gate, and building a fault signature library based on the fault records; Obtain historical sensor monitoring data collected by the sensor group; Performing pattern recognition and statistical analysis on the historical monitoring data of the sensor to form a fault feature weight table, and storing the historical monitoring data of the sensor and the fault feature weight table in a data cleaning database; Classifying the sensor group by database indexing technology according to the fault feature library, the fault feature weight table, the external prediction data and the installation location data of the sensor group to obtain a key sensor group, a warning sensor group and a conventional sensor group; and write the classification results into the sensor-level index table stored in the data cleaning database through the database interface; Extracting an original feature set based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data; the original feature set includes: sensor working state data, time series features, sensor spatial location information, and real-time working condition data; An additional feature set is constructed based on the fault feature weight table, the external prediction data, and the installation location data; the additional feature set includes: fault frequency features, fault mode features, installation location distance features, spatial coordination features, environmental response sensitivity features, and sensor stability features; the environmental response sensitivity features include wind speed response features, wind direction response features, river hydrological response features, and extreme weather response features; the sensor stability features include data fluctuation amplitude features, response time features, and repeatability features; The original feature set and the additional feature set are subjected to feature fusion to form a comprehensive feature vector; the comprehensive feature vector is subjected to standardization to obtain a standard comprehensive feature vector; Input the standard comprehensive feature vector into the trained XGBoost model and output the classification result; dividing the sensor group into the key sensor group, the warning sensor group and the regular sensor group according to the classification result; Within a specified time period, real-time data collection is performed through N independent data processing units according to the sensor level index table; the data processing units exchange data through a database interface protocol, and together constitute a distributed data intelligent monitoring system; The data processing unit includes a data acquisition layer, a data real-time analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

2. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 1 is characterized in that: Constructing the fault feature library according to the fault records is specifically as follows: The fault record includes a fault type and a fault triggering condition; and the fault feature library is constructed for different fault types and fault triggering conditions.

3. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 1 is characterized in that: The sensor group is used to monitor the various working parameters of the designated gate in real time; the sensor group includes a displacement sensor, a water level sensor, a flow sensor, a current sensor, a voltage sensor, a vibration sensor and a load stress sensor.

4. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 3 is characterized in that: The sensor group specifically includes: The sensor group is composed of a plurality of independent type sensors; wherein, the displacement sensor is composed of a plurality of independent displacement sensors, which are respectively arranged at the first position, the second position and the third position of the gate leaf of the designated gate, and are used to monitor the vertical displacement of the gate leaf of the designated gate, and obtain the first position displacement data, the second position displacement data and the third position displacement data after collection; the water level sensor includes an upstream water level sensor and a downstream water level sensor; wherein, the upstream water level sensor is located in the upstream water area of ​​the designated gate, at a first specified distance from the designated gate, at a second specified depth below the water surface, and is used to monitor the upstream water level of the designated gate; the downstream water level sensor is located in the downstream water area of ​​the designated gate, at a third specified distance from the designated gate, at a fourth specified depth below the water surface, and is used to monitor the downstream water level of the designated gate; The flow sensor is arranged in the discharge area of ​​the designated gate, and is used to monitor the flow of the designated gate; the current sensor and the voltage sensor are respectively installed in the motor control circuit of the designated gate hoist, and are used to monitor the current and voltage of the designated gate; the vibration sensor is arranged on the supporting structure of the designated gate, and is used to monitor the vibration of the designated gate; The sensor group provides data for the distributed data intelligent monitoring system, and each of the data processing units is independently configured according to the corresponding sensor type.

5. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 1 is characterized in that: The specified time period specifically refers to the duration of the next opening or closing operation of the specified gate.

6. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 1 is characterized in that: The external prediction data includes environmental prediction data and hydrological prediction data within the specified time period in the future.

7. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 1 is characterized in that: The data processing unit is specifically: The data acquisition layer acquires real-time monitoring data through corresponding independent type sensors, and the data real-time analysis layer performs real-time analysis on the real-time monitoring data to obtain normal real-time data and abnormal real-time data; The data cleaning and quality control layer performs data correction on the abnormal real-time data to obtain corrected real-time data; The data storage and retrieval layer uploads the normal real-time data and the corrected real-time data to the database system for storage, and supports data query and analysis operations through a retrieval interface. The database system adopts a multi-dimensional index structure and supports multi-condition retrieval and data processing.

8. The intelligent monitoring and analysis method for gate opening and closing condition data according to claim 7 is characterized in that: The data processing unit further includes an early warning module, which is configured to trigger an alarm mechanism and send alarm information to a corresponding coupled data processing unit when abnormal real-time data is obtained through real-time analysis of the real-time monitoring data. The determination of the coupled data processing unit includes: The sensor historical monitoring data and the fault records are analyzed to identify all potential coupling relationships among the N independent data processing units; and the coupled data processing unit is determined based on the correlation between the abnormal real-time data and the potential coupling relationships.

9. An intelligent monitoring and analysis system for gate opening and closing condition data, characterized in that: include: A fault data acquisition module is used to obtain fault records from a historical fault data repository of a specified gate and to construct a fault feature library based on the fault records; The fault signature database is organized in a database structure and supports retrieval and update operations; A sensor data processing unit, configured to obtain historical sensor monitoring data collected by the sensor group; a fault pattern recognition and feature extraction module, configured to perform pattern recognition and statistical analysis on the sensor historical monitoring data to form a fault feature weight table, and store the sensor historical monitoring data and the fault feature weight table in a data cleaning database, wherein the data cleaning database supports data storage, updating, and retrieval; An external prediction data acquisition module is used to acquire external prediction data; a sensor classification module, configured to classify the sensor group according to the fault feature weight table, the external prediction data, and the installation location data of the sensor group by using a database indexing technology to obtain a key sensor group, a warning sensor group, and a conventional sensor group; The classification results are written into a sensor-level index table stored in a data cleaning database through a database interface. The sensor-level index table adopts a hierarchical index structure and supports retrieval and update. Extracting an original feature set based on the fault feature library, the fault feature weight table, the external prediction data, and the installation location data; the original feature set includes: sensor working state data, time series features, sensor spatial location information, and real-time working condition data; An additional feature set is constructed based on the fault feature weight table, the external prediction data, and the installation location data; the additional feature set includes: fault frequency features, fault mode features, installation location distance features, spatial coordination features, environmental response sensitivity features, and sensor stability features; the environmental response sensitivity features include wind speed response features, wind direction response features, river hydrological response features, and extreme weather response features; the sensor stability features include data fluctuation amplitude features, response time features, and repeatability features; The original feature set and the additional feature set are subjected to feature fusion to form a comprehensive feature vector; the comprehensive feature vector is subjected to standardization to obtain a standard comprehensive feature vector; Input the standard comprehensive feature vector into the trained XGBoost model and output the classification result; dividing the sensor group into the key sensor group, the warning sensor group and the regular sensor group according to the classification result; The data acquisition and real-time analysis module is used to perform real-time data acquisition through N independent data processing units according to the sensor-level index table within a specified time period; the data processing units exchange data through a database interface protocol, together forming a distributed data intelligent monitoring system; the data processing units include a data acquisition layer, a real-time data analysis layer, a data cleaning and quality control layer, and a data storage and retrieval layer.

Images