Bridge structure health degree monitoring and analysis method based on deep learning

Abstract

The invention discloses a deep learning-based bridge structure health degree monitoring and analysis method, which comprises the following steps of: acquiring shape, stress and temperature data by installing an MEMS (Micro Electro Mechanical System) sensor in a bridge, storing the data in real time by using an RFID (Radio Frequency Identification Device) read-write device, performing AES-256 encryption and compression on the data through a multi-section wireless communication module, and storing the data into a database; the method comprises the following steps: firstly transmitting data to a bridge health monitoring center, firstly decompressing the data after the center receives the data, then analyzing the data by using a deep learning model, evaluating the health state of the bridge and setting an early warning mechanism, in addition, the system adopts a cloud native architecture, supports distributed deployment and elastic capacity expansion, and adopts a multi-factor authentication and fine-grained access control strategy, so that the safety of the system is improved. According to the invention, the health condition of the bridge can be effectively monitored, and potential problems can be found in time, so that the safe operation of the bridge is ensured.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a bridge structure health monitoring and analysis method based on deep learning. Background Art

[0002] The existing image vision algorithm detection and engineering monitoring technologies can no longer meet the actual needs of bridge health detection.

[0003] The Chinese patent publication number: CN111428833B discloses the name: A bridge construction monitoring system based on blockchain. The invention aims to solve the problem that when RFID equipment reads tag information, it needs to detect small input signals very sensitively, and the noise generated by the internal circuit of the RFID equipment will cause the instability of the entire system, such as power supply noise, electromagnetic noise and interference between digital and analog. External lightning or static electricity will also bring high-voltage spike noise to the internal circuit of the RFID device to block the reception of normal signals, and even damage the equipment, affecting the normal operation of the entire bridge construction monitoring system. In its core technical solution, the dual-channel amplification and adjustment circuit uses the voltage follower principle to isolate and output the output signal of the collector of the transistor VT1, reduce the interference between digital and analog in the post-stage circuit processing, and use a double T filter to eliminate power supply noise. At the same time, it can effectively eliminate external electromagnetic noise, reduce the impact of spike pulses caused by lightning or static electricity, and ensure the normal reception and processing of signals. The above content uses wireless radio frequency identification reading and writing equipment to store sensor device data in RFID tags as a data recording carrier and cannot solve the market demand for real-time collection of bridge structure health data combined with micro-motor system sensors and multi-band wireless communication modules.

[0004] The present invention proposes a solution to the above shortcomings: when using a wireless radio frequency identification (RFID) reader / writer as a data recording medium, a micro-electromechanical system (MEMS) sensor and a multi-band wireless communication module are combined to collect bridge structure health data, and data compression and encryption technology are used to ensure the security and efficiency of data transmission. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present invention provides a bridge structure health monitoring and analysis method based on deep learning.

[0006] The technical solution of the present invention is implemented as follows: A bridge structure health monitoring and analysis method based on deep learning, comprising:

[0007] S1. Use MEMS sensors in bridges to collect bridge shape, stress, and temperature data, and use RFID reading and writing equipment to store the data collected by the MEMS sensors in real time into RFID tags as a recording medium for bridge health data;

[0008] S2, encrypting the data stored in the RFID tag using the AES-256 encryption algorithm through a multi-segment wireless communication module, and compressing the data before transmitting it to the bridge health monitoring center;

[0009] S3, decompressing the encrypted compressed data transmitted in real time received by the bridge health monitoring center to restore the original data;

[0010] S4. Use deep learning models to analyze the deformation, stress, and temperature data of bridges, assess the health status of bridges, and set up early warning mechanisms;

[0011] S5, using cloud-native architecture for distributed deployment and elastic expansion of bridge health monitoring centers;

[0012] S6. Use multi-factor authentication and fine-grained access control strategies to ensure data security and stability in the bridge health monitoring center.

[0013] Furthermore, the MEMS sensor parameters in the S1 step are set as follows: accelerometer: sensitivity is ±16g, sampling rate is 10kHz, gyroscope: sensitivity is ±250° / s, sampling rate is 10kHz, temperature sensor: accuracy is ±0.1°C, sampling rate is 1Hz; RFID read / write device parameters: read / write distance: 5cm, read / write speed: 100 times / second, power supply requirement: 5V DC.

[0014] Furthermore, the writing process of the RFID reader / writer is as follows: Encoding: The RFID reader / writer first encodes the received digital data into a format that the RFID tag can understand, including the format specified by the ISO / IEC 15693 or ISO / IEC 14443 standards; Sending: The encoded data is sent out in the form of radio waves through the antenna of the RFID reader / writer; Receiving: The antenna of the RFID tag receives these radio waves and converts them into electrical signals; Writing: The circuit inside the RFID tag converts these electrical signals into data and stores them in the tag's memory.

[0015] Furthermore, in the S2 step, the multi-band wireless communication module is connected to the bridge health monitoring center and adopts LoRa: the communication distance is 10km, the data transmission rate is 10kbps, Wi-Fi: the communication distance is 100m, the data transmission rate is 100Mbps, and Bluetooth: the communication distance is 50m, and the data transmission rate is 1Mbps; data compression and encryption: data compression adopts H.264 video compression algorithm with a compression ratio of 2:1, and data encryption uses AES-256 encryption algorithm with a key length of 256 bits.

[0016] Furthermore, the steps of data encryption and compression are:

[0017] S201, data read from an RFID tag, the data including the written RFID tag;

[0018] S202, split the data into 128-bit (16-byte) blocks and use PKCS#7 padding;

[0019] S203, initialize encryption, select CBC mode, and preset initialization vector IV;

[0020] S204, IV are XORed, and the result after XOR only displays the first 16 bytes;

[0021] S205, storing the encrypted data back to the RFID tag;

[0022] S206, inputting the encrypted data into the H.264 compression algorithm;

[0023] S207, H.264 compression algorithm first compresses the data within the frame, including transformation, quantization and entropy coding; transformation: it is converted into the frequency domain and transformed by a transformation matrix (discrete cosine transform DCT); quantization: the coefficients of the transformed data block in the frequency domain are quantized, and the quantization parameter (QP) determines the quantization step size; entropy coding: the quantized data is further compressed, and entropy coding (Huffman coding or arithmetic coding) is used to reduce redundancy;

[0024] S208, setting the compression ratio to 2:1, and outputting the compressed data after being processed by the H.264 compression algorithm.

[0025] Furthermore, in the data reception and processing in the step S3, the data reception rate is 100 kbps, and the data processing time is 10 ms.

[0026] Furthermore, in the S4 step, a multi-input multi-output (MIMO) deep learning model is adopted, wherein the input layer receives bridge deformation, stress, and temperature data as input, the hidden layer uses convolutional layers, loop layers, and fully connected layers to extract and transform features of the input data, and the output layer outputs the health status assessment results of the bridge, including health scores and fault predictions.

[0027] Furthermore, in the multiple-input multiple-output (MIMO) deep learning model, the input layer input data is: The following formula is preset:

[0028] Deformation data: x = a1*t+a2*t^2;

[0029] Stress data: y = b1*t+b2*t^2;

[0030] Temperature data: z = c1*t+c2*t^2;

[0031] Health status assessment result: h = d1 + d2*x + d3*y + d4*z;

[0032] Among them, a1, a2, b1, b2, c1, c2, d1, d2, d3, d4 are preset coefficients representing the physical characteristics of the bridge;

[0033] Deformation data(x):

[0034] a1: deformation coefficient, which represents the linear response of the bridge to time t in the initial state;

[0035] a2: deformation coefficient, which represents the secondary response of the bridge to time t in the initial state;

[0036] t: time variable, indicating the change of time;

[0037] Stress data(y):

[0038] b1: stress coefficient, representing the linear response of the bridge to time t in the initial state;

[0039] b2: stress coefficient, representing the secondary response of the bridge to time t in the initial state;

[0040] Temperature data(z):

[0041] c1: temperature coefficient, representing the linear response of the bridge to time t in the initial state;

[0042] c2: temperature coefficient, representing the secondary response of the bridge to time t in the initial state;

[0043] Health status assessment results (h):

[0044] d1: health status assessment coefficient, indicating the impact of deformation on health status;

[0045] d2: health status assessment coefficient, indicating the impact of stress on health status;

[0046] d3: health status assessment coefficient, indicating the impact of temperature on health status;

[0047] d4: Health status assessment coefficient, which represents the impact of the interaction between deformation, stress and temperature on the health status.

[0048] Furthermore, in the step S5, the system architecture uses 2 servers, each configured with Intel Xeon E5-2650 v4 CPU, 64 GB RAM, 1 TB SSD, and the database uses MySQL 5.7 with a storage capacity of 10 TB.

[0049] Furthermore, in step S6, the firewall uses Huawei USG6300, which is configured with 4 Gigabit interfaces, the intrusion detection system uses Snort, and the rule update frequency is once a week.

[0050] Beneficial Effects

[0051] In view of the shortcomings of the existing technology that image vision algorithm detection and engineering monitoring can no longer meet the actual needs of bridge health detection, the present invention adopts MEMS sensors to collect bridge shape, stress, and temperature data, and combines RFID reading and writing equipment to store the data in real time in RFID tags as a recording medium for bridge health data. This solves the problem that the existing technology uses wireless radio frequency identification reading and writing equipment to store sensor device data in RFID tags, which as data recording carriers cannot meet the market demand for real-time collection of bridge structure health data by combining micro-electromechanical system sensors and multi-band wireless communication modules. At the same time, the data stored in the RFID tag is encrypted by the AES-256 encryption algorithm through the multi-band wireless communication module, and data compression technology is used to ensure the security and efficiency of data transmission, thereby solving the problem. It solves the problem that RFID devices need to sensitively detect small input signals when reading tag information, and the noise generated by the internal circuit of RFID devices will cause instability of the entire system, such as power supply noise, electromagnetic noise, and interference between digital and analog. In addition, the deformation, stress, and temperature data of the bridge are analyzed by deep learning models, the health status of the bridge is evaluated, and an early warning mechanism is set up, which improves the real-time and accuracy of bridge health detection, effectively prevents the occurrence of emergencies similar to the Guangzhou Lixinsha Bridge, and ensures the safe operation of the bridge. At the same time, a cloud-native architecture is used for distributed deployment and elastic expansion of the bridge health monitoring center, as well as multi-factor authentication and fine-grained access control strategies for data security and stability in the bridge health monitoring center, further improving the reliability and security of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a structural block diagram of a bridge structure health monitoring and analysis method based on deep learning in an embodiment of the present invention;

[0053] Figure 2 It is a flowchart of a method for monitoring and analyzing the health of a bridge structure based on deep learning in an embodiment of the present invention;

[0054] Figure 3 It is a structural block diagram of a deep learning module of a bridge structure health monitoring and analysis method based on deep learning in an embodiment of the present invention;

[0055] Figure 4 This is a structural block diagram of a bridge health monitoring center of a bridge structure health monitoring and analysis method based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] The preferred implementation methods of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these implementation methods are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0058] It should be noted that when an element is considered to be "connected" to another element, it can be directly connected to the other element, or connected to the other element through an intermediate element. In addition, the "connection" in the following embodiments should be understood as "electrical connection", "communication connection", etc. if there is transmission of electrical signals or data between the connected objects.

[0059] When used herein, the singular forms "a", "an", and "said / the" may also include plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "include / comprise" or "have" and the like specify the presence of stated features, wholes, steps, operations, components, parts, or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof. At the same time, the term "and / or" used in this specification includes any and all combinations of the relevant listed items.

[0060] See also Figure 1-4 As shown, a bridge structure health monitoring and analysis method based on deep learning includes:

[0061] S1. Use MEMS sensors in bridges to collect bridge shape, stress, and temperature data, and use RFID reading and writing equipment to store the data collected by the MEMS sensors in real time into RFID tags as a recording medium for bridge health data;

[0062] S2, encrypting the data stored in the RFID tag using the AES-256 encryption algorithm through a multi-segment wireless communication module, and compressing the data before transmitting it to the bridge health monitoring center;

[0063] S3, decompressing the encrypted compressed data received by the bridge health monitoring system in real time and restoring the original data;

[0064] S4. Use deep learning models to analyze the deformation, stress, and temperature data of bridges, assess the health status of bridges, and set up early warning mechanisms;

[0065] S5, using cloud-native architecture for distributed deployment and elastic expansion of bridge health monitoring centers;

[0066] S6. Use multi-factor authentication and fine-grained access control strategies to ensure data security and stability in the bridge health monitoring center.

[0067] like Figure 2 As shown, in step S1, MEMS sensors are used to collect bridge shape, stress, and temperature data, and RFID reading and writing equipment is used to store the data collected by the MEMS sensors in real time into RFID tags as a recording carrier of bridge health data, wherein the bridge shapes include beam bridges, arch bridges, rigid frame bridges, cable-stayed bridges, suspension bridges, and combined system bridges.

[0068] Specifically, MEMS sensor parameters use:

[0069] Accelerometers: Installed at key locations on the bridge to monitor bridge deformation, collect data in real time at a sampling rate of 10kHz, with a sensitivity of ±16g. They detect changes in acceleration caused by bridge deformation and can provide accurate data to reflect the health status of the bridge;

[0070] Gyroscopes: Installed at key locations on the bridge to monitor its rotation, they also collect data in real time at a sampling rate of 10kHz. With a sensitivity of ±250° / s, they detect changes in angular velocity as the bridge rotates and provide accurate data to reflect the health of the bridge.

[0071] Temperature sensor: installed at key locations on the bridge to monitor the temperature changes of the bridge. It collects data in real time at a sampling rate of 1Hz with an accuracy of ±0.1℃, detects temperature changes on the bridge, and can provide accurate data to reflect the health status of the bridge.

[0072] RFID reading and writing equipment parameters use:

[0073] Reading and writing distance: 5cm, the RFID reading and writing equipment is placed near the bridge to ensure that it can communicate effectively with the RFID tag and store the sensor data into the tag in real time;

[0074] Read and write speed: 100 times / second. The RFID reader / writer processes data at a read and write speed of 100 times / second, ensuring that data can be written and read quickly and efficiently.

[0075] Power requirement: 5V DC, RFID reader / writer requires a stable 5V DC power supply to ensure its normal operation.

[0076] like Figure 2 As shown, in step S2, the data encryption and compression steps are:

[0077] S201, data read from an RFID tag, the data including the written RFID tag;

[0078] S202, split the data into 128-bit (16-byte) blocks and use PKCS#7 padding;

[0079] S203, initialize encryption, select CBC mode, and preset initialization vector IV;

[0080] S204, IV are XORed, and the result after XOR only displays the first 16 bytes;

[0081] S205, storing the encrypted data back to the RFID tag;

[0082] S206, inputting the encrypted data into the H.264 compression algorithm;

[0083] S207, H.264 compression algorithm first compresses the data within the frame, including transformation, quantization and entropy coding; transformation: it is converted into the frequency domain and transformed by a transformation matrix (discrete cosine transform DCT); quantization: the coefficients of the transformed data block in the frequency domain are quantized, and the quantization parameter (QP) determines the quantization step size; entropy coding: the quantized data is further compressed, and entropy coding (Huffman coding or arithmetic coding) is used to reduce redundancy;

[0084] S208, setting the compression ratio to 2:1, and outputting the compressed data after being processed by the H.264 compression algorithm.

[0085] like Figure 1-Figure 2 As shown, in step S3, the encrypted compressed data received by the bridge health monitoring center in real time is decompressed to restore the original data;

[0086] Specifically, the process of decompression and recovery of original data is as follows:

[0087] Data reception: The bridge health monitoring center receives the encrypted compressed data transmitted through the wireless communication module;

[0088] Decryption process: Decrypt the encrypted data using the same key and initialization vector (IV) as used for encryption;

[0089] The decryption process is the opposite of the encryption process. First, an XOR operation is performed, and then the XOR result is decrypted using the AES-256 decryption algorithm.

[0090] Data decompression: The decompression process uses the inverse process of the H.264 video compression algorithm to restore the compressed data to its state before encryption;

[0091] The decompression process includes inverse intra-frame compression, inverse inter-frame compression, inverse transformation, inverse quantization and inverse entropy coding;

[0092] Data restoration: After decryption and decompression, the original data is restored, i.e. the health monitoring data of bridge deformation, stress, temperature, etc.

[0093] Data storage or further processing: The restored raw data can be stored in the database of the bridge health monitoring center for subsequent analysis, report generation or triggering of early warning mechanisms.

[0094] like Figure 3 As shown, in step S4, a multi-input multi-output (MIMO) deep learning model is used. The input layer receives the deformation, stress, and temperature data of the bridge as input. The hidden layer uses the convolutional layer, the recurrent layer, and the fully connected layer to extract and transform the input data. The output layer outputs the health status assessment results of the bridge, including the health score and the fault prediction.

[0095] The following formula is preset:

[0096] Deformation data: x = a1*t+a2*t^2;

[0097] Stress data: y = b1*t+b2*t^2;

[0098] Temperature data: z = c1*t+c2*t^2;

[0099] Health status assessment result: h = d1 + d2*x + d3*y + d4*z;

[0100] Among them, a1, a2, b1, b2, c1, c2, d1, d2, d3, d4 are preset coefficients representing the physical characteristics of the bridge;

[0101] Deformation data(x):

[0102] a1: deformation coefficient, which represents the linear response of the bridge to time t in the initial state;

[0103] a2: deformation coefficient, which represents the secondary response of the bridge to time t in the initial state;

[0104] t: time variable, indicating the change of time;

[0105] Stress data(y):

[0106] b1: stress coefficient, representing the linear response of the bridge to time t in the initial state;

[0107] b2: stress coefficient, representing the secondary response of the bridge to time t in the initial state;

[0108] Temperature data(z):

[0109] c1: temperature coefficient, representing the linear response of the bridge to time t in the initial state;

[0110] c2: temperature coefficient, representing the secondary response of the bridge to time t in the initial state;

[0111] Health status assessment results (h):

[0112] d1: health status assessment coefficient, indicating the impact of deformation on health status;

[0113] d2: health status assessment coefficient, indicating the impact of stress on health status;

[0114] d3: health status assessment coefficient, indicating the impact of temperature on health status;

[0115] d4: health status assessment coefficient, which represents the impact of the interaction between deformation, stress and temperature on the health status;

[0116] The specific calculation process is:

[0117] The following preset data is preset:

[0118] Time variable t = 10 (unit: hours)

[0119] Deformation coefficient a1 = 0.01

[0120] Deformation coefficient a2 = -0.005

[0121] Stress coefficient b1 = 0.015

[0122] Stress coefficient b2 = -0.0075

[0123] Temperature coefficient c1=0.01

[0124] Temperature coefficient c2 = -0.005

[0125] Health status assessment coefficient d1 = 0.3

[0126] Health status assessment coefficient d2 = 0.4

[0127] Health status assessment coefficient d3 = 0.2

[0128] Health status assessment coefficient d4 = 0.1

[0129] According to the preset formula, the following values ​​are calculated:

[0130] Deformation data (x): x = a1*t + a2*t^2x = 0.01*10 + (-0.005)*(10^2)x = 0.1-0.005*100x = 0.1-5x = -4.9;

[0131] Stress data (y): y = b1*t + b2*t^2y = 0.015*10 + (-0.0075)*(10^2)y = 0.15-0.0075*100y = 0.15-7.5y = -7.35;

[0132] Temperature data (z): z = c1*t + c2*t^2z = 0.01*10 + (-0.005)*(10^2)z = 0.1-0.005*100z = 0.1-5z = -4.9;

[0133] Health status assessment result (h): h = d1 + d2*x + d3*y + d4*zh = 0.3 + 0.4*(-4.9) + 0.2*(-7.35) + 0.1*(-4.9) h = 0.3-1.96 + (-1.47) -0.49 h = 0.3-1.96-1.47-0.49 h = -3.62;

[0134] Bridge health status settings:

[0135] If h is less than 0, it means that the health of the bridge is very poor;

[0136] If h is greater than 0 and h is less than 7, it means that the bridge is in good health;

[0137] According to the calculation results, the health status assessment result h is -3.62, which indicates that the health status of the bridge is very poor;

[0138] Early warning mechanism setting:

[0139] When the health status assessment result is lower than the threshold of 0, the early warning mechanism is triggered to remind relevant personnel to take measures to maintain the safety of the bridge.

[0140] like Figure 4 As shown, in step S5, a cloud-native architecture is used for distributed deployment and elastic expansion of the bridge health monitoring center;

[0141] Specifically, the system architecture of the bridge health monitoring center includes front-end, back-end, database, and storage components, and ensures the requirements of cloud-native architecture;

[0142] Deploy front-end applications on application servers on cloud platforms, including Amazon EC2, Azure VMs, or Google Compute Engine, to implement the user interface and data interaction of the bridge health monitoring system; use load balancers, including Amazon ELB, Azure Load Balancer, or Google Cloud LoadBalancing, to handle front-end traffic and ensure service continuity and availability;

[0143] Deploy backend services and databases on the computing resources of the cloud platform, including Amazon RDS, Azure SQL Database, or Google Cloud SQL, to process the core business logic and data storage of the bridge health monitoring system; use the storage services of the cloud platform, including Amazon S3, Azure Blue Storage, or Google Cloud Storage, to store bridge health monitoring data to ensure data reliability and security;

[0144] Implement distributed deployment. Utilize the distributed characteristics of the cloud platform to deploy the bridge health monitoring center in multiple geographical locations to achieve local data processing and rapid service response. Set up load balancing and automatic failover to ensure service continuity and availability.

[0145] Achieve elastic expansion, using the automatic expansion function of the cloud platform to automatically increase or decrease resources according to actual needs to cope with traffic fluctuations and business growth, set elastic scaling policies, and automatically adjust resource configuration according to real-time data traffic, such as increasing or decreasing server instances and storage capacity;

[0146] Implement data storage and processing Use the cloud platform's object storage service to store bridge health monitoring data, and use the cloud platform's computing service to process and analyze data, ensuring efficient data processing and analysis;

[0147] Leverage cloud platform security features, including authentication, access control, and data encryption, to ensure data security at the bridge health monitoring center;

[0148] Through the above steps, the bridge health monitoring center can realize the distributed deployment and elastic expansion of the cloud native architecture to improve the scalability, reliability and security of the system, thereby better meeting the needs of bridge health monitoring.

[0149] like Figure 1-Figure 2 As shown, in step S6, multi-factor authentication and fine-grained access control strategies are used to ensure data security and stability in the bridge health monitoring center;

[0150] Specifically, multi-factor authentication uses authentication mechanisms based on multiple factors such as passwords, tokens (RFID, smart cards), and biometrics (fingerprints, facial recognition);

[0151] When users access the Bridge Health Monitoring Center, they need to provide at least two authentication factors, including a password and a smart card;

[0152] Fine-grained access control uses role-based access control (RBAC) to assign access rights based on user roles and responsibilities;

[0153] Define different user roles, including administrator, operator, and auditor;

[0154] Use access control lists (ACLs) to control access to data and resources, ensuring that only authorized users can access sensitive data;

[0155] Implement time-based access control to limit user access to resources during specific time periods;

[0156] Implement auditing and logging functions to record all access attempts and successful operations for easy tracking and auditing;

[0157] Secure transmission and storage: HTTPS protocol is used for data transmission to ensure the security of data during transmission. The database and files stored in the bridge health monitoring center are encrypted to prevent unauthorized access.

[0158] Through the above technical implementation details, the bridge health monitoring center implements multi-factor authentication and fine-grained access control strategies to ensure data security and stability, thereby protecting the key data and resources of the bridge health monitoring system.

[0159] like Figure 1-Figure 4 As shown, the present invention works in the following manner in the application scenario to solve the shortcomings of the prior art image vision algorithm detection and engineering monitoring that cannot meet the needs of bridge health detection:

[0160] Data collection:

[0161] Install MEMS sensors in the bridge, including accelerometers, gyroscopes and temperature sensors, to collect deformation, stress and temperature data of the bridge in real time;

[0162] Accelerometer parameters: sensitivity ±16g, sampling rate 10kHz;

[0163] Gyroscope parameters: sensitivity ±250° / s, sampling rate 10kHz;

[0164] Temperature sensor parameters: accuracy ±0.1℃, sampling rate 1Hz;

[0165] Data Storage:

[0166] Use RFID reading and writing equipment to store the data collected by the MEMS sensor into the RFID tag in real time. The parameters of the RFID reading and writing equipment are as follows: reading and writing distance 5cm, reading and writing speed 100 times / second, and power supply requirement 5V DC;

[0167] Data transmission:

[0168] The data stored in the RFID tag is encrypted using the AES-256 encryption algorithm through a multi-segment wireless communication module, and the data is compressed and transmitted to the bridge health monitoring center;

[0169] The multi-band wireless communication module adopts LoRa: communication distance is 10km, data transmission rate is 10kbps, Wi-Fi: communication distance is 100m, data transmission rate is 100Mbps, Bluetooth: communication distance is 50m, data transmission rate is 1Mbps;

[0170] Data compression and encryption: Data compression uses the H.264 video compression algorithm with a compression ratio of 2:1, and data encryption uses the AES-256 encryption algorithm with a key length of 256 bits;

[0171] Health status assessment and early warning mechanism:

[0172] Use deep learning models to analyze bridge deformation, stress, and temperature data, assess bridge health status, and set up early warning mechanisms;

[0173] Set warning thresholds based on health status assessment results;

[0174] When the evaluation result exceeds the threshold, the early warning mechanism is triggered to remind relevant personnel to take measures to maintain the safety of the bridge.

[0175] like Figure 1-Figure 4 As shown, the present invention works as follows in the application scenario of solving the disadvantage of the RFID device being unstable when reading tag information:

[0176] Data collection:

[0177] Installing MEMS sensors, including accelerometers, gyroscopes, and temperature sensors, at key locations on the bridge;

[0178] Accelerometer parameters: sensitivity ±16g, sampling rate 10kHz;

[0179] Gyroscope parameters: sensitivity ±250° / s, sampling rate 10kHz;

[0180] Temperature sensor parameters: accuracy ±0.1℃, sampling rate 1Hz;

[0181] Data Storage:

[0182] Use RFID reading and writing equipment to store the data collected by the MEMS sensor into the RFID tag in real time;

[0183] RFID reading and writing equipment parameters: reading and writing distance 5cm, reading and writing speed 100 times / second, power supply requirement 5V DC;

[0184] Data transmission:

[0185] The data stored in the RFID tag is encrypted using the AES-256 encryption algorithm through a multi-segment wireless communication module, and the data is compressed and transmitted to the bridge health monitoring center;

[0186] The multi-band wireless communication module adopts LoRa: communication distance is 10km, data transmission rate is 10kbps, Wi-Fi: communication distance is 100m, data transmission rate is 100Mbps, Bluetooth: communication distance is 50m, data transmission rate is 1Mbps;

[0187] Data compression and encryption: Data compression uses the H.264 video compression algorithm with a compression ratio of 2:1, and data encryption uses the AES-256 encryption algorithm with a key length of 256 bits;

[0188] Health status assessment and early warning mechanism:

[0189] Use deep learning models to analyze the deformation, stress, and temperature data of bridges;

[0190] The input layer receives bridge deformation, stress, and temperature data;

[0191] The hidden layer uses convolutional layers, recurrent layers, and fully connected layers to extract and transform features of the input data;

[0192] The output layer outputs the health status assessment results of the bridge, including health scores and fault predictions;

[0193] Set warning thresholds based on health status assessment results;

[0194] When the evaluation result exceeds the threshold, the early warning mechanism is triggered to remind relevant personnel to take measures to maintain the safety of the bridge.

[0195] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0196] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and rules of the present invention shall be included in the protection scope of the present invention.

Claims

1. A bridge structure health monitoring and analysis method based on deep learning, characterized in that: The following steps are involved: S1. Use MEMS sensors in bridges to collect bridge shape, stress, and temperature data, and use RFID reading and writing equipment to store the data collected by the MEMS sensors in real time into RFID tags as a recording medium for bridge health data; S2, encrypting the data stored in the RFID tag using the AES-256 encryption algorithm through a multi-segment wireless communication module, and compressing the data before transmitting it to the bridge health monitoring center; S3, decompressing the encrypted compressed data transmitted in real time received by the bridge health monitoring center to restore the original data; S4. Use deep learning models to analyze the deformation, stress, and temperature data of bridges, assess the health status of bridges, and set up early warning mechanisms; S5, using cloud-native architecture for distributed deployment and elastic expansion of bridge health monitoring centers; S6. Adopt multi-factor authentication and fine-grained access control strategies for bridge health monitoring centers.

2. A bridge structure health monitoring and analysis method based on deep learning according to claim 1, characterized in that: The writing process of the RFID reader / writer is as follows: Encoding: The RFID reader / writer first encodes the received digital data into an RFID tag format, including the formats specified by ISO / IEC 15693 and ISO / IEC 14443 standards; Sending: The encoded data is sent out in the form of radio waves through the antenna of the RFID reader / writer; Receiving: The RFID tag’s antenna receives these radio waves and converts them into electrical signals; Writing: Circuitry inside the RFID tag converts these electrical signals into data and stores it in the tag's memory.

3. The bridge structure health monitoring and analysis method based on deep learning according to claim 1 is characterized by: The data encryption and compression steps in step S2 are: S201, data read from an RFID tag, the data including the written RFID tag; S202, split the data into 128-bit blocks and use PKCS#7 padding; S203, initialize encryption, select CBC mode, and preset initialization vector IV; S204, IV are XORed, and the result after XOR only displays the first 16 bytes; S205, storing the encrypted data back to the RFID tag; S206, inputting the encrypted data into the H.264 compression algorithm; S207, H.264 compression algorithm first compresses the data within the frame, including transforming, quantizing and entropy coding; Transformation: It is converted into the frequency domain and transformed by a transformation matrix; Quantization: The coefficients of the transformed data block in the frequency domain are quantized, and the quantization parameter determines the quantization step size; Entropy coding: quantized data is further compressed; S208, setting the compression ratio to 2:1, and outputting the compressed data after being processed by the H.264 compression algorithm.

4. The bridge structure health monitoring and analysis method based on deep learning according to claim 1 is characterized by: The S4 step adopts a multiple-input multiple-output (MIMO) deep learning model. The input layer receives bridge deformation, stress, and temperature data as input. The hidden layer uses convolutional layers, loop layers, and fully connected layers to extract and transform input data. The output layer outputs the health status assessment results of the bridge, including health scores and fault predictions.

5. The bridge structure health monitoring and analysis method based on deep learning according to claim 1 is characterized by: In the S5 step, the system architecture uses two servers, each configured with Intel Xeon E5-2650v4 CPU, 64GB RAM, 1TB SSD, and the database uses MySQL 5.7 with a storage capacity of 10TB.

6. The bridge structure health monitoring and analysis method based on deep learning according to claim 1 is characterized by: In step S6, the firewall uses Huawei USG6300, which is configured with 4 Gigabit interfaces, and the intrusion detection system uses Snort, and the rule update frequency is once a week.

Images