Intelligent bridge support detection and maintenance management system and method

Abstract

The invention discloses an intelligent bridge support detection and maintenance management system and method, and relates to the field of intelligent bridge monitoring and maintenance management, and the system is composed of four main modules: a sensor data acquisition module, a data transmission and storage module, a data processing and analysis module, and a maintenance and management module. Wherein the sensor data acquisition module is provided with a visual sensor, a vibration sensor, a displacement sensor and a crack sensor, the bridge support is monitored in real time through different types of sensors, and dynamic and static health data of the support are acquired. According to the system, various sensors are installed at key parts of the bridge support, real-time monitoring of the health state of the support is achieved, the collected data are analyzed in combination with the artificial intelligence technology to predict the remaining life and the maintenance period of the support, real-time early warning is conducted on the abnormal support, and therefore maintenance management of the bridge support is optimized, and the maintenance efficiency of the bridge support is improved. And long-term safety and stability of the bridge structure are ensured.

Description

Technical Field

[0001] The present application relates to the field of intelligent bridge monitoring and maintenance management, and in particular to a bridge bearing detection and maintenance management system and method based on sensor data collection and artificial intelligence analysis technology. Background Art

[0002] With the rapid development of the transportation industry, bridges, as important transportation infrastructure, carry a large amount of traffic flow and load. With the increase of service life, bridge bearings face the risk of gradual aging and damage, which seriously affects the safety and service life of the bridge structure. In order to ensure the stability and safety of the bridge, timely and effective detection of the bearing status and necessary repairs and maintenance have become important tasks in bridge maintenance and management.

[0003] Traditional bridge bearing inspection methods usually rely on manual inspections, where inspectors detect external damage or deformation of bearings through visual inspections or simple measuring tools. However, this method has many limitations, including: low frequency and poor efficiency of manual inspections, and the possibility of missed inspections or false inspections; manual inspections are difficult to accurately capture dynamic loads and minor damage to bridge bearings; in addition, traditional inspection methods cannot provide scientific assessments and predictions of the remaining life of bearings, resulting in inaccurate maintenance cycles or excessive maintenance.

[0004] In order to solve the above problems, intelligent monitoring technology based on sensors has been introduced in modern bridge engineering. Sensors can collect dynamic and static data of bridge bearings in real time, including vibration data, displacement data, crack data, etc., through which the health status of bearings can be fully understood. However, although sensor technology has been applied to bridge monitoring, traditional sensor data analysis methods are often limited to simple threshold judgments, lack of efficient and accurate algorithm support, and it is difficult to provide a comprehensive assessment and accurate prediction of the health status of bearings.

[0005] In order to further improve the detection accuracy and intelligence level of bridge bearings, the application of artificial intelligence (AI) technology, especially deep learning and machine learning technologies in bridge health monitoring has received widespread attention in recent years. These technologies can extract deep feature information based on the massive data collected by sensors to perform health status assessment, life prediction and maintenance decisions. However, most existing technologies currently rely only on a single data source or simple algorithms, and are unable to conduct comprehensive analysis of multi-source data. In addition, they lack a complete intelligent maintenance management system and cannot achieve intelligent early warning and decision support.

[0006] Therefore, how to combine multiple sensor technologies and advanced artificial intelligence algorithms to provide a comprehensive, accurate, and intelligent bridge bearing detection and maintenance management system has become a technical problem that needs to be solved urgently in the current bridge engineering field. This invention is aimed at this problem and proposes an intelligent bridge bearing detection and maintenance management system and method based on multi-sensor data fusion and artificial intelligence analysis to improve the monitoring accuracy of the bridge, optimize the maintenance cycle, and effectively improve the safety and intelligent management level of the bridge engineering. Summary of the invention

[0007] This application provides an intelligent bridge bearing detection and maintenance management system and method, which aims to comprehensively monitor, evaluate the health and maintain bridge bearings through precise sensor acquisition and artificial intelligence algorithms. The system installs various sensors at key locations of bridge bearings to achieve real-time monitoring of the health status of bearings, and analyzes the collected data in combination with artificial intelligence technology to predict the remaining life and maintenance cycle of bearings, and to provide real-time warnings for abnormal bearings, thereby optimizing the maintenance management of bridge bearings and ensuring the long-term safety and stability of bridge structures.

[0008] To achieve the above objectives, in the first aspect, the present application discloses an intelligent bridge bearing detection and maintenance management system, which consists of a sensor data acquisition module, a data transmission and storage module, a data processing and analysis module, and a maintenance and management module.

[0009] The sensor data acquisition module includes visual sensors, vibration sensors, displacement sensors and crack sensors, which monitor bridge bearings in real time through different types of sensors and collect dynamic and static health data of the bearings.

[0010] The data obtained by the sensor data acquisition module is sent to the data processing and analysis module via the data transmission and storage module. Specifically, all sensor data are transmitted to the data transmission and storage module via wireless or wired transmission. Each data packet contains the sensor ID, data value and timestamp to ensure the timing and integrity of the data. The data is stored in JSON or CSV format for subsequent processing and analysis.

[0011] The data processing and analysis module obtains data from the data transmission and storage module, processes and analyzes the collected data, and sends the analyzed data to the maintenance and management module. The maintenance and management module displays the support health status and maintenance suggestions based on the data through a graphical user interface.

[0012] Furthermore, the visual sensor: a high-resolution industrial camera is installed on the surface of the support, about 0.5 to 1 meter away from the support surface, and the angle is maintained between 30° and 60°. The sensor is used to capture damage such as cracks, corrosion and wear on the support surface. The image resolution is 0.1mm, and an image is collected every 5 seconds. The image is stored in JPEG or PNG format.

[0013] Furthermore, the vibration sensor: a triaxial accelerometer is used and installed at the bottom of the support or at the connection between the support and the bridge main beam. The vibration sensor has a sampling frequency of 1000Hz, a measurement range of ±2g to ±16g, a resolution of 0.01g, and collects 1000 data points per second to record the acceleration data of the vibration.

[0014] Furthermore, displacement sensor: use laser displacement sensor or optical displacement sensor, installed on the upper surface, lower surface or between the bearing and the bridge deck, to measure the longitudinal, lateral and vertical displacement of the bearing, with a measurement range of 0-100mm, an accuracy of 0.01mm, a sampling frequency of 1Hz, and displacement data collected once per second.

[0015] Furthermore, the crack sensor: an infrared thermal imaging sensor or an ultrasonic sensor is used, which is installed on the surface of the support or inside the support to detect cracks in the support. The measurement accuracy is 0.1mm and the working frequency is to collect an image or data once per second.

[0016] Furthermore, the wireless transmission mode of the data transmission and storage module includes at least one of a cellular network, Wi-Fi or Bluetooth.

[0017] Furthermore, the processing steps of the data processing and analysis module include data preprocessing, feature extraction and analysis model training. It mainly includes the following contents: Data preprocessing: Use digital waveform filters to remove noise from vibration and displacement data. Specifically, vibration data uses high-pass filtering or low-pass filtering to remove low-frequency or high-frequency noise. Image data uses Gaussian blur algorithm to remove environmental noise in the image.

[0018] Standardization processing: Standardize the data collected by all sensors. For example, normalize the amplitude and frequency of vibration data to between 0 and 1, and normalize the displacement data to the range of 0 to 100, to ensure that data from different types of sensors can be processed uniformly.

[0019] Feature extraction: For vibration data, Fourier transform (FFT) is used to extract vibration frequency, amplitude and other features; image data is extracted through convolutional neural network (CNN) to extract crack features; displacement data is extracted through time series analysis to extract the deformation characteristics of the support; crack data is extracted through image processing algorithms to extract crack expansion patterns.

[0020] Health assessment and prediction: Based on AI algorithms (such as LSTM, K-means and SVM), the extracted features are analyzed and evaluated. The LSTM model is used to predict the health status and remaining life of the bearing; the K-means clustering algorithm is used to cluster vibration, displacement, crack and other data to identify abnormal bearings; SVM is used to classify the health status of the bearing and evaluate its damage level (such as healthy, slightly damaged, moderately damaged and severely damaged).

[0021] In the second aspect, the present application discloses an intelligent bridge bearing detection and maintenance management method, which collects, preprocesses, and analyzes data, and combines artificial intelligence algorithms to perform health assessment and prediction, and finally generates maintenance plans and early warning information. The specific steps of the method are as follows: Step 1: Collect the health status data of the bridge bearings in real time by installing visual sensors, vibration sensors, displacement sensors and crack sensors at key locations of the bridge bearings. The sensors collect relevant data regularly and transmit them to the data transmission module for storage via MQTT or HTTP protocol; each data packet contains information such as sensor ID, timestamp, data type and data value, and carries a timestamp to ensure the timing of the data; all data will be stored in a standardized format for subsequent analysis and processing.

[0022] Step 2: Perform high-pass filtering or low-pass filtering on the vibration data in the health status data to remove interference signals and noise. For image data, use the Gaussian blur algorithm to remove environmental noise in the image.

[0023] The sensor data is standardized, the amplitude and frequency of the vibration data are normalized to between 0 and 1, and the displacement data is normalized to the range of 0 to 100, to ensure that the data of different types of sensors can be processed uniformly.

[0024] Step 3: Use Fourier transform to perform frequency domain analysis on the vibration data, extract characteristics such as vibration frequency and amplitude, and reflect the dynamic response of the support.

[0025] Convolutional neural network (CNN) is used to extract crack features in image data, including the number, location, length and width of cracks.

[0026] Perform time series analysis on the displacement data to extract the deformation characteristics of the bearing, including total displacement, displacement rate and other information, and evaluate whether abnormal deformation of the bearing has occurred.

[0027] Use image processing algorithms (such as edge detection) to extract the crack propagation pattern, and use image deep learning algorithms to analyze the crack propagation trend.

[0028] Step 4: Input the data obtained in step 3 into the LSTM model, and use the K-means clustering algorithm and SVM classification algorithm to analyze the data simultaneously. The LSTM model is a long short-term memory network used to analyze time series data such as vibration, displacement, and cracks, capture long-term dependencies in the data, and predict the health status and remaining life of the bearing. The input of the LSTM model is the historical data of multiple sensors, and the output is the predicted health status value at a certain moment in the future.

[0029] The K-means clustering algorithm is used to perform cluster analysis on vibration, displacement, crack and other data, divide the data into normal and abnormal categories, and identify possible abnormal supports.

[0030] The SVM classification algorithm uses a support vector machine to classify the health status of the bearings and divides the bearing status into "healthy", "slightly damaged", "moderately damaged" and "severely damaged" categories.

[0031] Step 5: Based on the results of step 4, when abnormal data is detected, an early warning is issued via SMS, email or other similar prompting methods to notify relevant personnel to conduct inspection and repair.

[0032] This intelligent bridge bearing detection and maintenance management system realizes all-round monitoring and management of bridge bearings through the cooperation of multiple modules, using multi-sensor technology and artificial intelligence algorithms. Each module plays a vital role in the system, ensuring accurate data collection, effective transmission, precise analysis and intelligent decision-making, thus providing strong protection for the safety and long-term use of bridges.

[0033] The above-listed beneficial effects are not exhaustive of all advantages. Other potential beneficial effects and detailed technical implementations will be further disclosed in the examples or other description parts of this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] After reading the following detailed description in conjunction with the accompanying drawings, you will better understand the various aspects of the present disclosure. The positions, sizes, and ranges of various structures shown in the accompanying drawings sometimes do not represent the actual positions, sizes, and ranges. In the accompanying drawings: Figure 1 It is a system composition block diagram of an embodiment disclosed in this application.

[0035] Figure 2 It is a flowchart of a workflow in an embodiment disclosed in this application. DETAILED DESCRIPTION

[0036] The present disclosure will be described below with reference to the accompanying drawings, which illustrate several embodiments of the present disclosure. However, it should be understood that the present disclosure can be presented in a variety of different ways and is not limited to the embodiments described below; in fact, the embodiments described below are intended to make the disclosure of the present disclosure more complete and fully illustrate the scope of protection of the present disclosure to those skilled in the art. It should also be understood that the embodiments disclosed herein can be combined in various ways to provide more additional embodiments.

[0037] It should be understood that the same reference numerals represent the same elements throughout the drawings. In the drawings, the dimensions of certain features may be distorted for clarity.

[0038] It should be understood that the terms used in the specification are only used to describe specific embodiments and are not intended to limit the present disclosure. All terms (including technical terms and scientific terms) used in the specification have the meanings commonly understood by those skilled in the art unless otherwise defined. For the sake of brevity and / or clarity, the techniques, methods and equipment known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and equipment should be considered as part of the authorization specification.

[0039] The singular forms "a", "said" and "the" used in the specification include plural forms unless clearly indicated. The terms "include", "comprise" and "contain" used in the specification indicate the presence of the claimed features, but do not exclude the presence of one or more other features. The term "and / or" used in the specification includes any and all combinations of one or more of the relevant listed items. Example

[0040] This system is based on multi-sensor data fusion and artificial intelligence analysis technology. It collects key parameters such as vibration, displacement, cracks, etc. of bridge bearings in real time, and combines LSTM, SVM, K-means and other algorithms to dynamically evaluate the health status of bearings, predict lifespan, and provide maintenance decision support, thereby improving the accuracy and efficiency of bridge safety management.

[0041] The system consists of several core modules, the first of which is the sensor data acquisition module. This module includes visual sensors, vibration sensors, displacement sensors and crack sensors. The visual sensor captures the surface image of the support with high resolution (resolution of 0.1mm, image format of JPEG / PNG), and collects images every 5 seconds to detect surface defects such as cracks, corrosion and wear; the vibration sensor uses a three-axis accelerometer (such as the ADXL series) with a sampling frequency of 1000Hz, an accuracy of 0.01g, and a measurement range of ±2g to ±16g, which is used to monitor the vibration characteristics under dynamic loads; the displacement sensor uses the Keyence LK series laser sensor with an accuracy of 0.01mm and a range of 0-100mm, which can accurately measure the longitudinal, lateral and vertical displacement of the support; the crack sensor uses infrared thermal imaging or ultrasonic sensors with an accuracy of 0.1mm and a sampling frequency of 1Hz, which is used to track the propagation speed and depth of the crack.

[0042] These sensors transmit data to the data transmission and storage module through network protocols such as Wi-Fi, Zigbee or Ethernet. The data is encapsulated in JSON format, and the data packet contains timestamp, sensor ID, data type, value and unit.

[0043] It is important to understand that data packets are the basic unit of sensor data collection, and the system relies on these data packets for data transmission, storage, and processing. The information in the data packet includes timestamp, sensor ID, data type, data value, unit, device status, etc. This data is critical for subsequent processing and health assessment.

[0044] The design of the data packet ensures the effectiveness of data transmission, the consistency of timing, and the accuracy of data processing. The structure of each data packet includes the following key fields: packet_id (data packet ID). Each data packet has a unique identifier to ensure that each data packet can be uniquely identified in the entire system. Through the data packet ID, the system can track the data transmission process, which is convenient for troubleshooting and data management. Example: "packet_id": "12345" sensor_id (sensor ID), the sensor ID is used to uniquely identify the sensor from which the data comes. Each sensor is assigned a unique ID in the system to ensure that the system can distinguish between different types of sensors and their locations. The sensor ID is critical for data analysis, especially in multi-sensor systems, to ensure that the source and origin of the data can be clearly traced. Example: "sensor_id": "Vibration_Sensor_1" timestamp, which is used to record the exact time of data collection. Use the ISO 8601 standard time format to ensure that the time sequence of all data packets is consistent. The timestamp ensures that the system can align the time sequence of data from multiple sensors, so that data from different sensors can be matched and compared according to time. Example: "timestamp": "2025-01-14T12:00:00Z" data_type (data type), the data type field indicates the type of collected data, and can distinguish the data types collected by different sensors. For example, the vibration sensor collects acceleration data, the displacement sensor collects displacement data, and the visual sensor collects image data. The accurate identification of the data type helps the subsequent data processing module to make correct parsing and analysis. Example: "data_type": "acceleration" value, the data value field records the actual measurement data collected by the sensor. Depending on the sensor, the form of the data value is also different: For the vibration sensor, the data value is triaxial acceleration data (such as [0.034, -0.015, 0.002]), in g.

[0045] For a displacement sensor, the data value is the displacement (eg: 5.23mm).

[0046] For image sensors, the data values ​​may be a binary stream of an image or processed image features (such as the boundary location and length of a crack).

[0047] For a crack sensor, the data values ​​might be parameters such as crack location, width, and depth.

[0048] Example (shock sensor data): "value": [0.034, -0.015, 0.002] Example (displacement sensor data): "value": 5.23 Units: The unit field is used to identify the unit of the data value, ensuring that different data can be reasonably compared during analysis. For example, the unit of vibration data is g (gravitational acceleration), and the unit of displacement data is mm (millimeter). This field ensures that the units are not confused during subsequent processing and can correctly reflect the physical meaning of each data.

[0049] Example (shock sensor units): "units": "g" Example (displacement sensor units): "units": "mm" Status (device status), the device status field records the working status of the sensor (such as normal or faulty). This field is used to indicate whether the sensor is working properly to help the system determine the validity of the data. When the device status is abnormal, the system can take further measures, such as notifying maintenance personnel to check the device or replace the sensor. Example: "status": "normal" Combining the above fields, the following is a complete example of the sensor data packet structure. This example represents the data collected by a vibration sensor. The data packet format is encapsulated in a JSON structure to facilitate transmission and storage in the system: { "packet_id": "12345", "sensor_id": "Vibration_Sensor_1", "timestamp": "2025-01-14T12:00:00Z", "data_type": "acceleration", "value": [0.034, -0.015, 0.002], "units": "g", "status": "normal" } During the transmission process, the data packet is sent from the sensor module to the data transmission and storage module through the network (such as Wi-Fi, Zigbee or Ethernet). The data packet is sent through the communication protocol within the system (such as MQTT or HTTP) and parsed by the receiving end. Each data packet contains the necessary field information to facilitate the system to identify the data source, parse the data content, and integrate with the data from other sensors.

[0050] During data storage, the system stores data packets according to timestamps and aligns data from multiple sensors in chronological order. The storage module uses a time series database to process the data to ensure that the data can be quickly queried and efficiently accessed. The system also determines the validity of the data based on the status of the sensor. For sensor data with a status of "fault", the system will issue an alarm to remind maintenance personnel to check or replace the equipment.

[0051] In this way, the data packet structure ensures the fluidity and consistency of data throughout the system, enabling efficient and safe operation and maintenance management of bridge bearings through continuous data monitoring, real-time analysis and maintenance decision-making.

[0052] The data transmission module ensures the real-time performance of high sampling rate sensors through dynamic load balancing, controls the packet loss rate to <0.1%, and stores data through a time series database, aligns multi-source data by timestamp, and ensures data consistency and integrity.

[0053] The data processing and analysis module is the core part of the system, responsible for preprocessing, feature extraction and analysis of the collected raw data. The preprocessing stage includes noise removal of vibration data, using FIR filter to remove environmental noise, The image data is subjected to Gaussian blur algorithm (standard deviation σ=1.5) to remove noise and improve the accuracy of crack identification. During standardization, the vibration data is normalized to [0-1] and the displacement data is mapped to [0-100] to ensure that the data from different sensors are compared under the same standard. Time alignment uses linear interpolation method to ensure that multi-source data are analyzed under the same time scale.

[0054] In the feature extraction stage, the vibration data is extracted through Fourier transform (FFT) to extract the main frequency, amplitude and frequency band energy distribution, and the dynamic response of the support is analyzed; the image data is processed by a pre-trained convolutional neural network (CNN, such as VGG16 or ResNet) to identify the location, size and other characteristics of the crack; the displacement data is calculated by sliding window statistics to calculate the displacement rate and cumulative deformation amount, and the deformation of the support is analyzed; the crack data is combined with Canny edge detection and Hough transform to predict the propagation direction and depth of the crack.

[0055] In the data analysis phase, the system uses algorithms such as LSTM (Long Short-Term Memory Network), SVM (Support Vector Machine) and K-means to further analyze the extracted features. The role and implementation steps of each algorithm are described below.

[0056] First, LSTM (Long Short-Term Memory Network) is a deep learning algorithm for processing time series data, and is particularly suitable for processing data with temporal dependencies. LSTM can effectively capture long-term dependencies through its unique gating structure (including input gate, forget gate, and output gate), overcoming the gradient vanishing or explosion problems that traditional RNN (recurrent neural network) is prone to encounter in time series modeling. In this system, LSTM is used to analyze the time series of sensor data such as vibration, displacement, and cracks to predict the future health status of the bearing.

[0057] The structural design of the LSTM network consists of three layers of LSTM units, each layer contains 128 LSTM units, and a 0.2 Dropout layer is added after each layer. The function of the Dropout layer is to randomly discard a part of the neurons to reduce the risk of overfitting and improve the generalization ability of the model. The input data of the LSTM includes the health status, vibration data, displacement data, and crack data of the bearing in the past period of time. The output of the LSTM network is the probability prediction value of the health status of the bearing in the next 24 hours. The system determines whether the bearing will be damaged or fail based on these probability prediction values, and takes maintenance measures in advance.

[0058] SVM (Support Vector Machine) is another machine learning algorithm for classification, which is often used to solve binary or multi-classification problems. In this system, SVM is used to classify the health status of the bearing. Based on the characteristics of data such as vibration, displacement and cracks, the system classifies the health status of the bearing into four categories: "healthy", "slightly damaged", "moderately damaged" and "severely damaged". SVM separates data points of different categories by constructing a hyperplane. In order to improve the classification accuracy, SVM uses the radial basis kernel function (RBF), which can map the original data to a high-dimensional space, making the data linearly separable in the high-dimensional space.

[0059] The training process of SVM uses the cross-validation method to select appropriate hyperparameters. In this system, the hyperparameters of SVM are C=1.0 and γ=0.1, where C controls the width of the classification interval and γ controls the range of the kernel function. Through the selection of these parameters, the SVM model can achieve a classification accuracy of more than 92% and accurately classify the health status of the bearing into different damage levels.

[0060] The K-means clustering algorithm is used for anomaly detection. K-means is an unsupervised learning algorithm that can identify anomalies in the data by dividing the data into K clusters. In this system, K-means clustering divides vibration, displacement, crack and other data into normal and abnormal categories. First, the system selects K=2 (normal and abnormal), and then determines the cluster to which the data point belongs by calculating the Euclidean distance between each data point and the center of its cluster. If the distance between a data point and the center of the cluster exceeds 2 standard deviations (2σ), the data point is judged as abnormal data and the system triggers an early warning. This method helps the system to promptly detect bearings that may have faults and remind maintenance personnel to check.

[0061] Through these three algorithms, the system can comprehensively analyze the health status of the bearing and provide a basis for maintenance decisions. LSTM can provide predictions of the future health status of the bearing, SVM can classify the health status of the bearing, and K-means is used for anomaly detection. These three algorithms work together to ensure that the system can efficiently monitor and evaluate the health of the bearing.

[0062] When implemented, the maintenance management module generates a health report based on the analysis results (the report is automatically generated by the software). The report not only provides the current health status of the bearing, but also predicts the remaining life of the bearing through the LSTM model. The remaining life prediction can help the system identify potential problems in advance and provide a basis for maintenance planning. In the health report, the bearing will be classified into four categories: "healthy", "slightly damaged", "moderately damaged" and "severely damaged", and the system will automatically generate a maintenance plan based on the classification results.

[0063] For example, a bearing in healthy condition would be recommended to undergo routine inspection every 6 months; a bearing with minor damage would be recommended to undergo re-inspection every 3 months and continuously monitor the load; and when a bearing is assessed to be severely damaged, the system would recommend immediate shutdown and maintenance to ensure bridge safety.

[0064] In addition, the system also supports a three-level early warning mechanism. When the data analysis results show that there is an abnormality in the bearing, the system will issue an early warning according to different levels of urgency. The early warning mechanism is divided into three categories: pop-up window, SMS notification and manual intervention, ensuring that maintenance personnel can obtain important information in a timely manner and take necessary maintenance measures. Through these measures, the system can minimize the sudden failure of bridge bearings and ensure the long-term stability of the bridge structure.

[0065] This system has shown significant advantages in practical applications. Compared with traditional manual detection methods, the system has improved detection efficiency by 80% and significantly reduced maintenance costs by 40%. Through real-time monitoring, intelligent analysis and automated decision support, the system can provide more accurate and timely maintenance management. This not only greatly improves the accuracy and efficiency of bridge safety management, but also provides a practical technical solution for intelligent operation and maintenance of infrastructure. Through multi-sensor data fusion and artificial intelligence analysis, the health status of bridge bearings has been more comprehensively and accurately evaluated, promoting the development of bridge health management towards intelligence and automation.

[0066] Although the exemplary embodiments of the present disclosure have been described, it should be understood by those skilled in the art that various changes and modifications can be made to the exemplary embodiments of the present disclosure without departing from the spirit and scope of the present disclosure in essence. Therefore, all changes and modifications are included in the scope of protection of the present disclosure as defined by the claims. The present disclosure is defined by the appended claims, and the equivalents of these claims are also included.

Claims

1. An intelligent bridge bearing detection and maintenance management system, characterized in that: It includes sensor data acquisition module, data transmission and storage module, data processing and analysis module and maintenance and management module; The sensor data acquisition module includes a visual sensor, a vibration sensor, a displacement sensor and a crack sensor, which are used to monitor the bridge bearings in real time and collect dynamic and static health data; The data transmission and storage module receives data from the sensor data acquisition module via wireless or wired transmission. Each data packet contains a sensor ID, a data value, and a timestamp. The data storage is in JSON or CSV format. The data processing and analysis module obtains data from the data transmission and storage module, processes and analyzes the data, and sends the analyzed data to the maintenance and management module; The maintenance and management module displays the support health status and maintenance suggestions through a graphical user interface based on the analyzed data; The visual sensor is a high-resolution industrial camera, which is installed on the surface of the support, 0.5m to 1m away from the support surface, with an angle maintained between 30° and 60°, an image resolution of 0.1mm, and an image is collected every 5 seconds, and the image is stored in JPEG or PNG format; The vibration sensor is a three-axis accelerometer installed at the bottom of the support or at the connection between the support and the bridge main beam, with a sampling frequency of 1000Hz, a measurement range of ±2g to ±16g, a resolution of 0.01g, and 1000 data points collected per second; The displacement sensor is a laser displacement sensor or an optical displacement sensor, which is installed on the upper surface, lower surface or between the support and the bridge deck to measure the longitudinal, transverse and vertical displacement of the support. The measurement range is 0-100mm, the accuracy is 0.01mm, and the sampling frequency is 1Hz; The crack sensor is an infrared thermal imaging sensor or an ultrasonic sensor, which is installed on the surface of the support or inside the support, has a measurement accuracy of 0.1 mm, and operates at a frequency of collecting images or data once per second; The wireless transmission mode of the data transmission and storage module includes at least one of a cellular network, Wi-Fi or Bluetooth.

2. An intelligent bridge bearing detection and maintenance management method, characterized in that: The intelligent bridge bearing detection and maintenance management system as claimed in claim 1 comprises the following steps: Step 1: The health status data of the bridge bearings are collected in real time by installing visual sensors, vibration sensors, displacement sensors and crack sensors at key positions of the bridge bearings. The sensors collect data regularly and transmit the data to the data transmission module for storage via the MQTT protocol or HTTP protocol. Each data packet contains the sensor ID, timestamp, data type and data value information, and all data are stored in a standardized format. Step 2: high-pass filtering or low-pass filtering is performed on the vibration data in the health status data to remove interference signals and noise, and a Gaussian blur algorithm is used to remove environmental noise from the image data collected by the visual sensor; all sensor data are standardized; Step 3: Use Fourier transform to perform frequency domain analysis on vibration data to extract vibration frequency and amplitude characteristics; use convolutional neural network to extract crack characteristics in image data, including the number, location, length and width of cracks; perform time series analysis on displacement data to extract total displacement and displacement rate characteristics; use image processing algorithm to extract crack expansion mode, and use image deep learning algorithm to analyze crack expansion trend; Step 4: Input the features extracted in step 3 into the long short-term memory network, use the LSTM model to analyze the time series data of vibration, displacement and crack, and predict the health status and remaining life of the bearing; at the same time, use the K-means clustering algorithm to perform cluster analysis on the vibration, displacement and crack data to divide the normal and abnormal categories, and use the support vector machine to classify the health status of the bearing. The classification results include "healthy", "slightly damaged", "moderately damaged" and "severely damaged"; Step 5: Based on the classification and clustering results of step 4, when abnormal data is detected, a warning message is generated via SMS, email or prompt, and a maintenance plan is generated.

3. The intelligent bridge bearing detection and maintenance management method according to claim 2 is characterized in that: The standardization process in step 2 includes: The amplitude and frequency of the vibration data are normalized respectively so that their value range is limited to between 0 and 1; The displacement data is normalized so that its value range is limited to between 0 and 100.

4. The intelligent bridge bearing detection and maintenance management method according to claim 2 is characterized in that: The image processing algorithm in step 3 includes an edge detection algorithm for extracting the crack propagation pattern.

5. The intelligent bridge bearing detection and maintenance management method according to claim 2 is characterized in that: The input of the LSTM model described in step 4 is the historical data of multiple sensors, and the output is the predicted health status value at a certain moment in the future.

6. The intelligent bridge bearing detection and maintenance management method according to claim 2 is characterized in that: The conditions for generating the warning information described in step 5 are: When the K-means clustering result is an abnormal category, or the SVM classification result is "minor injury", "moderate injury" or "severe injury", an early warning is triggered.

Images