A real-time monitoring and early warning system for urban bridges
By deploying a real-time monitoring and early warning system on urban pedestrian overpasses, and using sensors and machine learning algorithms for multi-directional monitoring and data analysis, the problem of insufficient monitoring in the existing technology is solved, and timely safety assessment and hidden danger warning of bridge structures are achieved. , improving the intelligence level of bridge safety management.
Patent Information
- Application Number
- CN202311593980.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-27
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2043-11-27
AI Technical Summary
The monitoring methods of existing urban pedestrian bridges are simple, lacking professional equipment, and not systematic and comprehensive enough to detect bridge safety hazards in a timely manner, resulting in difficulties in formulating accident prevention and control and repair plans.
Design a real-time monitoring and early warning system for urban bridges, including intelligent monitoring module, position determination module, data transmission module, cloud analysis module, early warning module and decision-making support module, using image and sensor equipment, finite element analysis, autonomous crawling robot, 5G network , machine learning algorithms and blockchain technology to realize multi-directional intelligent monitoring and data analysis of bridge structure, timely identify abnormalities and push early warnings.
It has achieved comprehensive and intelligent monitoring of bridge structures, timely discover hidden dangers, actively prevent and control risks, reduce human resources demand, and improve the intelligence and informatization level of bridge safety management.
Smart Images

Figure CN117671921B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of bridge structure health monitoring, in particular to a real-time monitoring and early warning system for urban bridges. Background Art
[0002] At present, with the development of urban transportation, the number of urban pedestrian bridges has increased, the connectivity has improved, the design concept of pedestrian bridges has become more modern, the styles have become richer, and the material technology has also been continuously improved.
[0003] However, many problems have been found. Some old bridges are fatigued and damaged, and some bridge decks are often cracked and collapsed. The settlement of the pier foundation leads to safety hazards. The guardrails have safety design defects. Although there is monitoring of pedestrian bridges, due to the relatively simple monitoring methods, most of them are still in the stage of manual inspection and lack of professional equipment. The monitoring plan is not systematic and comprehensive, and important parts are ignored. The monitoring frequency is not high. In many areas, inspections are conducted only once a year, making it difficult to find hidden dangers. The monitoring personnel are not professional, and there is a lack of scientific basis for the judgment and treatment of damage. The accumulation of monitoring data is insufficient, and many monitoring results are not well recorded and analyzed. There is insufficient linkage between departments, insufficient information sharing, and it is difficult to conduct comprehensive analysis. Insufficient funding affects the breadth and depth of monitoring. There is a lack of follow-up rectification and application of monitoring results, and the rectification of hidden dangers is not timely and in place. Therefore, it has become an urgent problem to discover bridge safety hazards, prevent accidents, conduct scientific assessments of damaged parts, formulate repair plans, and monitor the operation quality of newly built bridges. Summary of the invention
[0004] In view of the above problems existing in the monitoring of the existing pedestrian bridges, the present invention is proposed.
[0005] Therefore, the problem to be solved by the present invention is how to provide a system that can timely monitor and discover bridge safety hazards, prevent accidents, and scientifically evaluate damaged parts and formulate repair plans.
[0006] To solve the above technical problems, the present invention provides the following technical solutions: a real-time monitoring and early warning system for urban bridges, comprising: an intelligent monitoring module, which is used to realize multi-dimensional intelligent monitoring of bridge structure, damage and stress through images and sensor equipment; a position determination module, which is used to use finite element analysis to determine the optimal arrangement of sensors, and design an autonomous crawling robot to realize monitoring of difficult-to-cover areas; a data transmission module, which is used to use a 5G wireless network to realize data transmission between the monitoring equipment and the cloud platform; a cloud analysis module, which is used to use a machine learning algorithm to intelligently process and analyze the monitoring data to realize the assessment of the safety status of the bridge; an early warning module, which is used to analyze and judge the monitoring data and identify the anomalies or risks in the structure; when an abnormal situation is detected, the early warning level is divided; a decision support module, which is used to provide processing suggestions or emergency plans based on the fault reasoning results, and actively push early warning information to relevant personnel through multiple channels; a mobile access module, which is used for relevant personnel to view monitoring and early warning information through APP at any time to facilitate rapid response.
[0007] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, the specific process of realizing multi-dimensional intelligent monitoring of bridge structure, damage and stress through images and sensor equipment is as follows: electromagnetic wave sensors are set on the bridge deck to monitor the deflection and settlement data of the bridge deck; fiber optic gyroscopes are buried in the main piers to measure the inclination and rotation data of the piers, and the setting rule of the fiber optic gyroscopes is: one is arranged at the top of each main pier; acoustic emission sensors are set on the main beams to detect the stress and fatigue data of the steel structure; and capacitive sensors are distributedly arranged on the cables to detect the size and change data of the cable force.
[0008] As a preferred solution of the real-time monitoring and early warning system for urban bridges of the present invention, the specific operation process of using finite element analysis to determine the optimal arrangement of sensors is as follows: collecting bridge design drawings and extracting geometric structure parameters; establishing a three-dimensional solid model, dividing units and nodes; applying normal loads to the model and solving the displacement field:
[0009] [K]{u}={F}
[0010] Where [K] is the stiffness matrix; {u} is the node displacement vector; {F} is the node force vector; dynamic load is applied to the model and the equation of motion is solved:
[0011] [M]{ü}+[C]{u}+[K]{u}={F(t)}
[0012] Among them, [M] is the mass matrix; {ü} is the node acceleration vector; [C] is the damping matrix; {u} is the node velocity vector; {F(t)} is the time-dependent load vector; the process of designing an autonomous crawling robot to achieve monitoring of difficult-to-cover areas includes analyzing the details of the bridge structure and determining the crawling path; designing the robot joint mechanism to achieve multi-directional crawling; selecting load-bearing materials and designing an automatic obstacle avoidance function.
[0013] As a preferred solution of the real-time monitoring and early warning system for urban bridges of the present invention, the specific process of intelligently processing and analyzing the monitoring data to realize the evaluation of the safety status of the bridge is as follows: the monitoring data of each sensor is collected once every t time, which is recorded as: electromagnetic wave data: E m , Gyroscope data: G y , Acoustic emission data: S e And capacitance data: C a ; Perform denoising and regularization preprocessing on each sensor data to obtain E m ′, G y ′、S e ′ and C a '; Construct an estimation network based on LSTM, the input layer is the data of each sensor, the preprocessed data is input, the middle layer is the LSTM layer, the time series characteristics of the learning data, the hidden layer is the LSTM memory unit, which contains the historical monitoring data of different parts; the output layer is the health status assessment result, expressed as the safety probability P; the training goal is to minimize the loss function, fitting the monitoring data and the actual state; the steps of fitting the monitoring data and the actual state are as follows:
[0014]
[0015] Among them, W is the network parameter weight, b is the network parameter bias, N is the number of training samples, y is the state of the monitoring data predicted by the network, and y′ is the actual state label corresponding to the monitoring data, that is, E m ′, G y ′、S e ′ and C a '; Repeat the training iteration and calculate the gradient of Loss to parameters W, b and Update parameters according to the gradient descent principle:
[0016]
[0017] Among them, ω is the learning rate, which controls the amplitude of parameter update. is the gradient of the bias parameter b to the loss function L; when the Loss function value converges, the model training is completed.
[0018] As a preferred solution of the real-time monitoring and early warning system for urban bridges of the present invention, the health status assessment includes: the output layer calculates the safety probability P based on the input integrated current and historical information, and the specific formula is:
[0019] P=σ(W x *E m ′+W y *G y ′+W z *S e ′+W t *Ca′+b)
[0020] Among them, σ is the activation function, W x From the input layer to the output layer, about E m ′’s weight matrix, W y For the input layer to the output layer about G y ′’s weight matrix, W z For the input layer to the output layer about S e ′’s weight matrix, W t From the input layer to the output layer, about C a ′ is the weight matrix of .
[0021] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, the early warning process of the early warning module is as follows: when 0.95<P≤1, it is displayed as a normal state and monitoring continues; when 0.9<P≤0.95, a level 3 early warning is displayed, and the system issues a prompt "The system detects a slight abnormality, please check the relevant parts, do you need to manually confirm the abnormality?" If the operator selects "Yes", it is upgraded to a level 2 early warning, otherwise it is maintained at level 3; the monitoring frequency is doubled until it returns to a normal state; when 0.85<P≤0.9, a level 2 early warning is displayed, an inspector is sent to inspect, and the use of the bridge is restricted; when P≤0.85, a level 1 early warning is displayed, the use of the bridge is immediately stopped, and an evaluation is conducted; the relevant early warning information is pushed to relevant technical personnel via the APP.
[0022] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, the specific process of the decision support module parsing the early warning information and providing processing suggestions or emergency plans includes: constructing a bridge maintenance knowledge graph to record the entity nodes of the fault type, early warning information and monitoring parameters, and the relationship between the entities; performing natural language processing on the early warning text information to extract keywords and semantic information; searching for keywords in the knowledge graph to find entity nodes that match them; reasoning along the relationship between entities to find the most likely matching fault type; calculating the matching degree between different fault type nodes and early warning keywords, and returning the fault type with the highest matching degree; if the matching result is not unique, comprehensively considering the relevant data of the monitoring parameters and performing summary scoring; outputting the text warning information and the fault type entity with the highest matching degree in the knowledge graph.
[0023] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, the hidden layer forms a curve trend graph with the historical detection values of different parts according to the average data in each monitoring period T. If abnormal data appears in the monitoring value, the monitoring period of the part is marked with a special display, and the abnormal data is used instead of the average data in the monitoring period; the abnormal data is integrated to summarize the regularity of the abnormal data appearing in different time periods; when the number of abnormal values of a certain part of the bridge is ≥ the average value of the abnormal values of all parts of the bridge, it is considered that the part has implicit loss factors, and a warning message is sent to the superior.
[0024] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, the judgment process of the curve trend is as follows: retrieve various data when the bridge is built, and set the normal fluctuation range of data at different parts; if the curve fluctuates within the normal design range and the fluctuation amplitude is less than 20%, it is judged to be normal; if the curve shows a short-term rapid rise or fall of more than 50%, it is judged as a level 1 emergency accident, and a level 1 warning is transmitted to the early warning module to check the data collection process; if the curve has an unreasonable trend of long-term steady rise or fall, it is judged as an abnormal development trend, and the sensor status needs to be confirmed, and a level 3 warning is transmitted to the early warning module; if compared with adjacent nodes in the month-on-month period, there is a fault in a single data, which is inconsistent with the previous and subsequent readings, it is judged that there is a problem in this period, verify and re-collect the data of this part, and transmit a level 2 warning to the early warning module; if the curve data has approached or even exceeded the design maximum threshold limit, such as the stress reaches the yield strength, it is judged as a level 1 emergency accident, and a level 1 warning is transmitted to the early warning module for risk warning; if any judgment condition is met, it is finally judged as abnormal data; if all judgment conditions are not triggered, it is comprehensively judged as normal data.
[0025] As a preferred solution of the real-time monitoring and early warning system for urban bridges described in the present invention, blockchain technology is used to store monitoring and early warning data to ensure the authenticity and traceability of the data.
[0026] The beneficial effects of the present invention are: the intelligent monitoring of the present invention comprehensively covers the key parts of the bridge, realizing active monitoring of the structural safety; ensuring high-speed, low-latency and reliable transmission of monitoring data through the 5G network; realizing more intelligent and efficient monitoring data analysis through machine learning algorithms; providing scientific processing suggestions based on the intelligent decision-making module; improving the overall intelligence, coordination and informatization level of bridge safety management, which is conducive to more timely discovery of hidden dangers, active prevention and control of risks, and at the same time reducing human resource requirements. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:
[0028] Figure 1 This is a system diagram of the urban bridge real-time monitoring and early warning system in Example 1.
[0029] Figure 2 This is a warning process diagram of the urban bridge real-time monitoring and early warning system in Example 1. DETAILED DESCRIPTION
[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.
[0031] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.
[0032] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.
[0033] Example 1
[0034] Reference Figure 1-2, which is the first embodiment of the present invention, provides a real-time monitoring and early warning system for urban bridges, including:
[0035] The intelligent monitoring module is used to realize all-round intelligent monitoring of bridge structure, damage and stress through images and sensor equipment.
[0036] Preferably, the specific process includes: setting electromagnetic wave sensors on the bridge deck to monitor the deflection and settlement data of the bridge deck, wherein two electromagnetic wave sensors are arranged for each adjacent span, for example, a bridge length of 100 meters, about 2 spans, requires 4 sensors; burying fiber optic gyroscopes in the main piers to measure the inclination and rotation data of the piers; setting acoustic emission sensors on the main beams to detect the stress and fatigue data of the steel structure, and the placement rule of the acoustic emission sensors is: 2 sensors are arranged at the bottom of the beam in each span, and 1 sensor is arranged at the top of the beam; and distributing capacitance sensors on the cables to detect the size and change data of the cable force, and the placement rule of the capacitance sensors is: 1 sensor is arranged at the middle point of each cable.
[0037] The position determination module is used to determine the optimal arrangement of sensors using finite element analysis and to design an autonomous crawling robot to monitor areas that are difficult to cover.
[0038] Preferably, the specific operation process of using finite element analysis to determine the optimal arrangement of sensors is: collect bridge design drawings and extract geometric structure parameters; establish a three-dimensional solid model and divide units and nodes; apply normal loads, such as deadweight and static force, to the model; solve the displacement field: [K]{u}={F}; where [K] is the stiffness matrix; {u} is the node displacement vector; {F} is the node force vector, and calculate the stress and strain distribution; apply dynamic loads, such as vehicle driving loads, to the model; solve the motion equation:
[0039] [M]{ü}+[C]{u}+[K]{u}={F(t)}
[0040] Where [M] is the mass matrix; {ü} is the node acceleration vector; [C] is the damping matrix; {u} is the node velocity vector; {F(t)} is the time-dependent loading vector; output the dynamic response analysis results; adjust the sensor layout, and repeat the above steps until the optimal configuration is obtained.
[0041] Preferably, the process of designing an autonomous crawling robot to achieve monitoring of difficult-to-cover areas includes analyzing the details of the bridge structure, determining the crawling path, collecting the bridge BIM three-dimensional model, and marking key parts; calculating joint kinematics and determining crawling requirements: using new flexible materials and cable drive mechanisms; achieving 360-degree free attachment and crawling; generating optimized crawling routes to cover all detection points; designing robot joint mechanisms to achieve multi-directional crawling; selecting load-bearing materials to ensure equipment safety; designing automatic obstacle avoidance functions; and integrating cameras and other sensing devices.
[0042] The data transmission module is used to realize data transmission between the monitoring equipment and the cloud platform using the 5G wireless network.
[0043] The cloud analysis module is used to use machine learning algorithms to intelligently process and analyze monitoring data to evaluate the safety status of bridges.
[0044] Furthermore, the specific steps include collecting the monitoring data of each sensor once every t time, which is recorded as: electromagnetic wave data: E m , Gyroscope data: G y , Acoustic emission data: S e And capacitance data: C a ; Perform denoising and regularization preprocessing on each sensor data to obtain E m ′, G y ′、S e ′ and C a ′; Build an estimation network based on LSTM. The input layer is the preprocessed data of each sensor data. The middle layer is the LSTM layer, which learns the time series characteristics of the data. The hidden layer is the LSTM memory unit, which contains the historical monitoring data of different parts. The output layer is the health status assessment result, which is expressed as the safety probability P. The training goal is to minimize the loss function and fit the monitoring data with the actual status.
[0045] Furthermore, the steps of fitting the monitoring data with the actual state are as follows:
[0046]
[0047] Among them, W is the network parameter weight, b is the network parameter bias, N is the number of training samples, y is the state of the monitoring data predicted by the network, and y′ is the actual state label corresponding to the monitoring data, that is, E m ′, G y ′、S e ′ and C a '; Repeat the training iteration and calculate the gradient of Loss to parameters W, b and Update parameters according to the gradient descent principle:
[0048]
[0049] Among them, ω is the learning rate, which controls the amplitude of parameter update. is the gradient of the bias parameter b to the loss function L; when the Loss function value converges, the model training is completed.
[0050] Preferably, the health status assessment includes: the output layer calculates the safety probability P based on the input integrated current and historical information, and the specific formula is:
[0051] P=σ(W x *E m ′+W y *G y ′+W z *S e ′+W t *C a ′+b)
[0052] Among them, σ is the activation function, W x From the input layer to the output layer, about E m ′’s weight matrix, W y For the input layer to the output layer about G y ′’s weight matrix, W z For the input layer to the output layer about S e ′’s weight matrix, W t From the input layer to the output layer, about C a ′ is the weight matrix of .
[0053] Furthermore, the hidden layer forms a curve trend chart with the historical detection values of different parts according to the average data in each monitoring period T. If abnormal data appears in the monitoring value, the monitoring period of the part is marked with a special display, and the abnormal data is used instead of the average data in the monitoring period; the abnormal data is integrated to summarize the regularity of the abnormal data appearing in different time periods; when the number of abnormal values in a certain part of the bridge is ≥ the average value of the abnormal values of all parts of the bridge, it is considered that there are implicit loss factors in this part, and an early warning message is sent to the superior.
[0054] Preferably, the judgment process of the curve trend is as follows: retrieve various data when the bridge is built, and set the normal fluctuation range of data at different parts; if the curve fluctuates within the normal design range and the fluctuation amplitude is less than 20%, it is judged to be normal; if the curve shows a short-term rapid rise or fall of more than 50%, it is judged as a level 1 emergency, and a level 1 warning is transmitted to the early warning module to check the data collection process; if the curve has an unreasonable trend of long-term steady rise or fall, it is judged to be an abnormal development trend, and the sensor status needs to be confirmed, and a level 3 warning is transmitted to the early warning module; if compared with adjacent nodes in the month-on-month period, there is a single data fault, which is inconsistent with the previous and subsequent readings, it is judged that there is a problem in this period, and the data of this part is verified and re-collected, and a level 2 warning is transmitted to the early warning module; if the curve data is close to or even exceeds the design maximum threshold limit, such as the stress reaches the yield strength, it is judged to be a level 1 emergency, and a level 1 warning is transmitted to the early warning module for risk warning; if any judgment condition is met, it is finally judged as abnormal data; if all judgment conditions are not triggered, it is comprehensively judged as normal data.
[0055] The early warning module is used to identify early warnings of abnormal situations and actively push early warning information to relevant personnel through various channels.
[0056] Preferably, the warning process of the early warning module is as follows: when 0.95<P≤1, it is displayed as a normal state and monitoring continues; when 0.9<P≤0.95, a level 3 warning is displayed, and the system issues a prompt "The system detects a slight abnormality, please check the relevant parts, do you need to manually confirm the abnormality?" If the operator selects "Yes", it is upgraded to a level 2 warning, otherwise it is maintained at level 3; the monitoring frequency is doubled until it returns to a normal state; when 0.85<P≤0.9, a level 2 warning is displayed, inspectors are sent to inspect, and the use of the bridge is restricted; when P≤0.85, a level 1 warning is displayed, the use of the bridge is immediately stopped, and an evaluation is conducted; the relevant warning information is pushed to relevant technical personnel via the APP.
[0057] The decision support module is used to analyze the warning information based on the knowledge base and provide processing suggestions or emergency plans.
[0058] The specific process is: construct a bridge maintenance knowledge graph, record entity nodes such as fault type, warning information, monitoring parameters, and the relationship between entities.
[0059] Perform natural language processing on warning text information to extract keywords and semantic information.
[0060] Search for these keywords in the knowledge graph and find the entity nodes that match them.
[0061] Reasoning is carried out along the relationship between entities, such as "warning information-indication-fault type", to find the most likely matching fault type.
[0062] Calculate the matching degree between different fault type nodes and warning keywords, and return the fault type with the highest matching degree.
[0063] If the matching result is not unique, relevant evidence such as monitoring parameters will be considered comprehensively to make a summary score.
[0064] The final output is the text warning information and the fault type entity with the highest matching degree in the knowledge graph.
[0065] The mobile access module allows relevant personnel to view monitoring and early warning information through the APP at any time, facilitating rapid response.
[0066] It should be noted that since the historical data on the storage chain is tamper-proof, the present invention uses blockchain technology to store monitoring and early warning data to ensure the authenticity and traceability of the data.
[0067] In summary, the intelligent monitoring of the present invention comprehensively covers the key parts of the bridge, realizing active monitoring of the structural safety; ensuring high-speed, low-latency and reliable transmission of monitoring data through the 5G network; realizing smarter and more efficient monitoring data analysis through machine learning algorithms, and providing scientific processing suggestions based on the intelligent decision-making module; improving the overall intelligence, coordination and informatization level of bridge safety management, which is conducive to more timely discovery of hidden dangers, active prevention and control of risks, and at the same time reducing human resource requirements.
[0068] Example 2
[0069] Referring to Table 1 and Table 2, which are the second embodiment of the present invention, based on the first embodiment, in order to verify its beneficial effects, a verification test of the present invention is provided to verify and illustrate the technical effects adopted in the method.
[0070] First, take a pedestrian bridge in the city as an example, where the bridge is 120 meters long, has 3 spans, 4 main piers, 3 main beams, and 6 cables; a 3D solid model is established in SolidWorks; the model is imported into ANSYS, and the number of unit and node sensors is divided as shown in the following table:
[0071] Table 1
[0072] Location Sensor Type quantity Data Flag Bridge Deck Electromagnetic waves 6 <![CDATA[E m 001-E m 006<!-- 6 --> ]]> main pier Gyroscope 4 <![CDATA[G y 001-G y 004]]> Main beam Acoustic Emission 9 <![CDATA[S e 001-S e 009]]> Cable capacitance 6 <![CDATA[C a 001-C a 006]]>
[0073] Set the steel density, elastic modulus and other properties, add the deadweight load, solve the displacement solution {u} = [3, 5, 8, 10, 6, 4, 7, 9] cm, calculate the stress distribution, and the maximum stress is concentrated at the third sensor; perform dynamic analysis: apply the vehicle load {F(t)}, solve the motion equation, output the time history of the bridge internal force, the stress fluctuation is the largest at the third sensor, adjust the position of the third sensor, and repeat the static and dynamic analysis until the sensors are arranged reasonably.
[0074] After the sensors were arranged, some of the collected data are as follows: m 001: [0.231, 0.221, 0.321, 0.311, 0.248]; E m 002: [0.201, 0.242, 0.223, 0.256, 0.289]; G y 001: [12.5, 11.6, 15.3, 16.1, 14.2]; G y 002: [10.2, 8.5, 9.3, 11.1, 10.5]; S e 001: [532, 512, 521, 517, 525]; S e 002: [642, 685, 661, 610, 628]S e003: [712, 698, 705, 688, 702]; C a 001: [1845, 1931, 2101, 2023, 1942]; C a 002: [1942, 2051, 2102, 2245, 2104]; C a 003: [2122, 2212, 2300, 2315, 2202]; perform denoising and regularization preprocessing on each sensor data, build an estimation network based on LSTM, the input layer is the sensor data, the preprocessed data is input, and the monitoring data is fitted with the actual state. x *E m ′+W y *G y ′+W z *S e ′+W t *Ca′+b) Calculate the safety probability.
[0075] The specific process is: preprocess various types of original data to obtain E m ′, G y ′、S e ′ and C a ′: E m ′=[0.3, 0.25, 0.35, 0.33, 0.28]; G y ′=[0.4, 0.3, 0.5, 0.55, 0.5]; S e ′=[0.6, 0.58, 0.6, 0.58, 0.6]; C a ′=[0.7,0.75,0.8,0.9,0.85];Substitute the preprocessed data into the formula for calculation: P=σ(0.2E m ′+0.1G y ′+0.4S e ′+0.3C a ′+0.1)=σ(0.06+0.045+0.24+0.255+0.1)≈0.91.
[0076] The calculated safety probability P is brought into the evaluation system: 0.85<P≤0.9, where the classification of the early warning module is shown in the following table:
[0077] Table 2
[0078]
[0079] The above table is used to judge that it is a level 2 warning, which will be pushed to the relevant technical personnel through the APP. Inspectors need to be sent to check and the use of the bridge will be restricted.
[0080] In summary, the intelligent monitoring of the present invention comprehensively covers the key parts of the bridge, realizing active monitoring of the structural safety; ensures high-speed, low-latency and reliable transmission of monitoring data through the 5G network; realizes more intelligent and efficient monitoring data analysis through machine learning algorithms; and provides scientific processing suggestions based on the intelligent decision-making module.
[0081] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A real-time monitoring and early warning system for urban bridges, characterized in that: include: Intelligent monitoring module, used to realize multi-dimensional intelligent monitoring of bridge structure, damage and stress through images and sensor equipment; Position determination module, which uses finite element analysis to determine the optimal placement of sensors and designs autonomous crawling robots to monitor hard-to-cover areas; Data transmission module, used to realize data transmission between monitoring equipment and cloud platform using 5G wireless network; The cloud analysis module is used to use machine learning algorithms to intelligently process and analyze monitoring data to evaluate the safety status of bridges; The early warning module is used to analyze and judge the monitoring data and identify the anomalies or risks in the structure; when an abnormal situation is detected, the early warning level is divided; The decision support module is used to provide processing suggestions or emergency plans based on the fault reasoning results, and actively push warning information to relevant personnel through various channels; Mobile access module, for relevant personnel to view monitoring and early warning information through APP at any time; The specific process of intelligently processing and analyzing the monitoring data to evaluate the safety status of the bridge is as follows: The monitoring data of each sensor is collected once every t time, recorded as: Electromagnetic wave data: E m , Gyroscope data: G y , Acoustic emission data: S e And capacitance data: C a ; Each sensor data is preprocessed by denoising and normalization to obtain E m ', G y '、S e ' and C a '; Construct an evaluation network based on LSTM. The input layer is the sensor data, which is the preprocessed data input. The middle layer is the LSTM layer, which learns the time series characteristics of the data. The hidden layer is the LSTM memory unit, which contains the historical monitoring data of different parts. The output layer is the health status assessment result, which is expressed as the safety probability P. The training goal is to minimize the loss function and fit the monitoring data to the actual state; The steps of fitting the monitoring data to the actual state are as follows: Among them, W is the network parameter weight, b is the network parameter bias, N is the number of training samples, y is the state of the monitoring data predicted by the network, and y' is the actual state label corresponding to the monitoring data, that is, E m ', G y '、S e ' and C a '; Repeat the training iterations and calculate the gradient of Loss with respect to the parameters W,b and Update parameters according to the gradient descent principle: Among them, ω is the learning rate, which controls the amplitude of parameter update. is the gradient of the bias parameter b to the loss function L; When the Loss function value converges, the model training is completed; The health status assessment includes: the output layer calculates the safety probability P based on the input integrated current and historical information, and the specific formula is: P=σ(W x *E m '+W y *G y '+W z *S e '+W t *C a '+b) Among them, σ is the activation function, W x From the input layer to the output layer, about E m 'The weight matrix, W y For the input layer to the output layer about G y 'The weight matrix, W z For the input layer to the output layer about S e 'The weight matrix, W t From the input layer to the output layer, about C a 'The weight matrix; The hidden layer forms a curve trend chart with the historical detection values of different parts according to the average data in each monitoring period T. If abnormal data appears in the monitoring value, the monitoring period of the part is marked with a special display, and the abnormal data is used instead of the average data in the monitoring period; the abnormal data is integrated to summarize the regularity of the abnormal data appearing in different time periods; when the number of abnormal values of a certain part of the bridge is ≥ the average value of the abnormal values of all parts of the bridge, it is considered that the part has an implicit loss factor, and a warning message is sent to the superior; The judgment process of the curve trend is as follows: Retrieve various data when the bridge was built, and set the normal fluctuation range of data in different parts; If the curve fluctuates within the normal design range and the fluctuation amplitude is less than 20%, it is judged to be normal; If the curve shows a short-term rapid rise or fall of more than 50%, it is judged as a Level 1 emergency, and a Level 1 warning is transmitted to the warning module to check the data collection process; If the curve shows an unreasonable trend of long-term steady rise or fall, it is judged as an abnormal development trend, and the sensor status needs to be confirmed and a level 3 warning is transmitted to the warning module; If there is a gap in a single data when compared with the adjacent nodes in the month-on-month period, and it is inconsistent with the previous and subsequent readings, it is judged that there is a problem in the period, and the data of the part is verified and recollected, and a level 2 warning is transmitted to the warning module; If the curve data has approached or even exceeded the designed maximum threshold limit, it is judged as an emergency level 1, and a level 1 early warning is transmitted to the early warning module for risk early warning; If any judgment condition is met, it is finally judged as abnormal data; If all judgment conditions are not triggered, it is comprehensively judged as normal data.
2. The urban bridge real-time monitoring and early warning system according to claim 1, characterized in that: The specific process of realizing the intelligent monitoring of the bridge structure, damage and stress in multiple directions through the image and sensor devices is as follows: An electromagnetic wave sensor is set on the bridge deck to monitor the deflection and settlement data of the bridge deck; Fiber optic gyroscopes are buried in the main bridge piers to measure the inclination and rotation data of the piers. The setting rule of the fiber optic gyroscopes is: 1 is arranged at the top of each main pier; Acoustic emission sensors are set on the main girder to detect the stress and fatigue data of the steel structure; Capacitive sensors are distributed on the suspension cables to detect the cable force magnitude and change data.
3. The urban bridge real-time monitoring and early warning system as claimed in claim 2 is characterized by: The specific operation process of determining the optimal layout of sensors by using finite element analysis is as follows: Collect the bridge design drawings and extract the geometric structure parameters; Establish a three-dimensional solid model, divide elements and nodes; Apply the normal load to the model and solve the displacement field: [K]{u} = {F} where, [K] is the stiffness matrix; {u} is the nodal displacement vector; {F} is the nodal force vector; Apply the dynamic load to the model and solve the motion equation: [M]{ü}+[C]{u}+[K]{u} = {F(t)} where, [M] is the mass matrix; {ü} is the nodal acceleration vector; [C] is the damping matrix; {u} is the nodal velocity vector; {F(t)} is the time-dependent loading vector; The process of designing an autonomous crawling robot to realize the monitoring of difficult-to-cover areas includes analyzing the bridge structure details and determining the crawling path; Design the robot joint mechanism to realize multi-directional crawling; Select the load-bearing material and design the automatic obstacle avoidance function.
4. The urban bridge real-time monitoring and early warning system as claimed in claim 3 is characterized by: The early warning process of the early warning module is as follows: When 0.95 < P ≤ 1, it is displayed as the normal state and continuous monitoring is carried out; When 0.9 < P ≤ 0.95, a level 3 early warning is displayed, and the system issues a prompt "The system detects a mild anomaly. Please check the relevant parts. Do you need to manually confirm the anomaly?" If the operator selects "Yes", it is upgraded to a level 2 early warning, otherwise it remains at level 3; The monitoring frequency is doubled until it returns to the normal state; When 0.85 < P ≤ 0.9, a level 2 early warning is displayed, an inspector is sent to check, and the use of the bridge is restricted; When P ≤ 0.85, a level 1 early warning is displayed, the use of the bridge is immediately stopped, and an assessment is carried out; The relevant early warning information is pushed to the relevant technical personnel through the APP.
5. The urban bridge real-time monitoring and early warning system as claimed in claim 4, characterized in that: The decision support module analyzes the early warning information and gives treatment suggestions or emergency plans. The specific process includes: Construct a bridge maintenance knowledge graph, record the entity nodes of the fault type, early warning information and monitoring parameters, as well as the relationships between the entities; Perform natural language processing on the early warning text information to extract keywords and semantic information; Search for keywords in the knowledge graph to find the matching entity nodes; Expand the reasoning along the relationships between the entities to find the most likely matching fault type; Calculate the matching degree between different fault type nodes and the early warning keywords, and return the fault type with the highest matching degree; If the matching result is not unique, the relevant data of the monitoring parameters are comprehensively considered and a summary score is given; The output text warning information is matched with the fault type entity with the highest matching degree in the knowledge graph.
6. The urban bridge real-time monitoring and early warning system according to claim 5, characterized in that: Use blockchain technology to store monitoring and early warning data to ensure the authenticity and traceability of the data.
Citation Information
Patent Citations
Bridge monitoring decision-making assisting system based on GIS and BIM
CN109686061A
Bridge health state detection and management and maintenance decision-making system and method
CN110633855A
Unit commitment calculation method based on LSTM
CN113420508A
Structural component global strain field inversion reconstruction method, device and equipment and storage medium
CN116822281A
Lithium ion battery fault detection method and system based on knowledge graph
CN117110896A
