CN114282398A

CN114282398A - Bridge health monitoring system and method based on big data

Info

Publication number: CN114282398A
Application number: CN202110614994.6A
Authority: CN
Inventors: 张应迁; 郭毅; 付磊; 罗云蓉; 谢云霞
Original assignee: Sichuan University of Science and Engineering
Current assignee: Sichuan University of Science and Engineering
Priority date: 2021-06-02
Filing date: 2021-06-02
Publication date: 2022-04-05

Abstract

The invention belongs to the technical field of deformation for measuring solids, and discloses a bridge health monitoring system and method based on big data, wherein research theories, test data and design data analysis of bridge health monitoring at home and abroad are collected, and hot spots and front edges in the aspects of bridge monitoring at home and abroad are mastered; determining the tension and compression constitutive relation and strength of the bridge concrete; and (3) carrying out a bridge concrete dynamic elastic modulus test: respectively adopting a hyperbolic model and a power function model to obtain characteristic parameters of the concrete constitutive relation; and determining the mass data of the bridge monitoring through big data analysis software WEKA. The invention relates to a theory and a method for bridge structure real-time damage inference and positioning, real-time model correction and safety evaluation based on big data, and a multilevel criterion for structure safety early warning; establishing a standard test model for identifying and evaluating the damage of a typical major engineering structure; and a theoretical and method unified inspection platform is provided for health monitoring and safety early warning of major engineering structures.

Description

Bridge health monitoring system and method based on big data

Technical Field

The invention belongs to the technical field of deformation for measuring solids, and particularly relates to a bridge health monitoring system and method based on big data.

Background

Currently, the current state of the art commonly used in the industry is such that: so far, a bridge health monitoring system is installed on a large bridge with 140 (including a medium bridge and a small bridge 260) seats in China. In the past, due to the limitation of design theory, hidden danger of construction quality and rapid increase of traffic volume brought by economic development of some infrastructures in China, a plurality of bridges are overwhelmed and are in a sub-health state. Especially, large bridges, such as those designed to be "overstrain diseases", will cause huge losses. The bridge structure health monitoring comprises daily usability monitoring, such as early warning bridge sealing when wind speed influences the normal passing of the bridge, and safety monitoring, such as whether a guy cable of a cable bearing bridge is firm enough. The health monitoring system of a large-span bridge is composed of more than 300 sensor measuring points and less than 50, and the cost accounts for 0.25 to 1 percent, even 2 percent of the total manufacturing cost of the bridge. At present, the design concept of the monitoring system focuses on providing support for maintenance management of the bridge, and the design scheme of processing and evaluating test data is less and less accepted by a bridge owner because of the emphasis on monitoring content and technology; the durability of the health monitoring system is more and more emphasized, and the related hardware of the system is required to be replaceable, and the continuity of data acquisition is not influenced during replacement; the health monitoring system after the bridge formation is combined with a monitoring system for controlling the construction process, and the monitoring data content extends to the construction stage; the levels of the structural state evaluation subsystem and the evaluation software are improved, and the intervention of bridge experts in the analysis and evaluation of the bridge structure health monitoring results is emphasized. The health monitoring development of the long-span bridge structure has three trends: deepening and perfecting the traditional monitoring technology and theoretical method; plays an important role in sustainable structural design; the method plays a wider role by means of new information technologies (Internet of things, cloud computing and big data).

About 70 to 80 percent of large-span bridges in China are new bridges, about 30 percent of cable-stayed bridges, about 30 percent of suspension bridges and arch bridges, and the rest are relatively large concrete rigid frame bridges. At present, bridges of health monitoring systems with complete design and complete sensors are applied, 250-300 bridges exist in China, and most bridges are large-span bridges; the small and medium span bridges also have urgent needs for health detection and monitoring. The main factors restricting the application degree are the maturity of the technology and the cost.

According to foreign experience, the actual service life of a bridge with the designed average service life of 75 years is about 40 years on average. For example, 1) United states: 58 thousands of bridges are investigated, 51.4 thousands of bridges are investigated, wherein more than 40% of bridges have damages of different degrees, the structural strength of 9.8 thousands of bridges is reduced, and the bridges are stopped to be used or limited in load and account for 19% of the total number; 2) in Japan: the bridge designed and constructed according to the old standard before 1956 has obviously insufficient bearing capacity. Statistically, such bridges have 5500 seats, wherein the concrete bridge has 4500 seats; 3) germany, a state; approximately half of the bridge superstructures of about 1500 concrete bridges had at least one major damage, with 2/3 having at least one general damage.

On the other hand, the design standard of the bridge in China is lower, the design bearing capacity of the bridge in China is only 68% and 60% of that of Meiying, and the design value of the live load effect bearing standard of the bridge is 40% and 59% lower than that of the Meiying standard. More seriously, because the construction quality is often not guaranteed and the overload phenomenon is generalized, the bridge damage and aging speed in China is very rapid. It is expected that from now on, China will inevitably come to a large-scale bridge aging phenomenon, and most bridges will reach service life in advance if not controlled. According to incomplete statistics, the current 1/3 bridge has various defects, and the bridge is more than 1 ten thousand. Therefore, how to reasonably detect and evaluate the bridges has important significance for maintaining the life and property safety of people, reducing the cost of the whole period and the like.

Bridge health monitoring has been introduced into China for nearly 20 years, and dynamic weighing instruments (dynamically weighed) for testing overload appear more than 10 years ago. However, neither health monitoring techniques nor overload monitoring techniques have gained widespread popularity and use. Currently, monitoring systems are established only in part of large and oversize bridges (about 100 seats), and more than 99.9% of the bridges are in the blank of monitoring. Especially prestressed concrete bridges, which account for the vast majority of bridges, it is almost impossible to see successful monitoring examples. The reason for this is that:

1) high cost, and is only suitable for large and oversize bridges

At present, due to technical and economic reasons, the cost of a single bridge health monitoring system is over millions of yuan, and part of grand bridges reach thousands of yuan or even hundreds of millions (such as hong Kong golden horse bridge). Meanwhile, due to the complexity of the system, the maintenance and operation costs are very high, so that the existing monitoring system cannot be applied to large and medium bridges (accounting for 22% of the total number of bridges), small-span bridges (accounting for 71% of the total number of bridges), and even large-span bridges (accounting for 6.3% of the total number of bridges) in few cases. It is precisely this portion of the bridge that occupies more than 99.8% of the total bridge and occupies almost all of the bridge-crossing accidents.

2) Complex system and poor reliability

The monitoring system is too complex because the system is sought to be large and complete. A typical bridge monitoring system often includes tens or even hundreds of subsystems, using tens or even hundreds of sensors. While the system cost is increased, the reliability of the system is seriously reduced.

3) Lack of calibration means and poor durability

For various reasons, the sensor has short service life, and the durability of the monitoring system is seriously influenced due to the problems of drift and the like in the use process. In particular, the lack of a calibration means for the monitoring system has made this problem less effective.

Therefore, the bridge health monitoring technology has many advantages, but is difficult to popularize and popularize in practical application. In particular, there is a lack of health monitoring systems for prestressed concrete bridges. The main sources of the bridge monitoring big data comprise three aspects: the health monitoring system generates massive real-time data, design of bridges, construction monitoring and other data, and social data. These data volumes are quite large, and only the data of a large bridge health monitoring system is taken as an example: the data volume generated by the western optical division portal bridge is about 3GB every day, and the data volume of the system is about 1TB every year; the data volume generated by the sutong bridge is about 10GB every day, and the data volume of the system is about 3TB every year.

The data-driven method focuses on monitoring the change rule of the obtained input and output data correlation relationship to identify the mode corresponding to the structural state, and is widely applied to the SHM by means of mature statistical theory. However, due to the limitations of computing power and analysis means, the traditional statistical method can only analyze a small portion of low-dimensional data samples, and cannot efficiently present the analysis result, so that the problem of analyzing massive high-dimensional SHM data is not yet solved.

The big data technology is an emerging technology in recent years, is widely applied to the fields of internet, electronic commerce, medicine and the like, solves the problems of insufficient computing capacity, low efficiency of a data analysis method and the like, and has a wide application prospect in SHM data processing. The method integrates the cognitive ability of people and not good at machines into the analysis process, can improve the efficiency and the accuracy of data analysis, and can visually present high-dimensional data.

The ultimate goal of big data processing and analysis is to make reasonable predictions and judgments about the structure by means of mining the data. The essential function of the novel concept of 'sample-total' applying big data is prediction, and the purpose of predicting the bridge damage is achieved through a data fusion and data mining method. With the arrival of the big data era, the bridge health monitoring technology is also developed rapidly, so that an information data chain interaction platform for analysis, integration, standardization and guidance of national large, medium and small bridge health, disease, prevention and diagnosis and treatment integrated management is constructed, and a reliable, efficient and continuously optimized service health and safety guarantee system is provided.

The bridge health monitoring technology is a new technology developed in recent decades, and mass data and one-sidedness of a single system of the bridge health monitoring technology are not well solved; similarly, big data analysis technology is a brand new technology developed in recent years, and has a wide application prospect in expanding cognitive ability by using big data analysis, and big data analysis technologies such as deep learning, knowledge calculation, visualization and the like are widely applied to different industries and fields. The invention provides a new concept of establishing a bridge health monitoring big data center by using a big data analysis technology, and provides an application prospect and a direction of the big data analysis technology in bridge health monitoring. With the arrival of the big data era, the bridge health monitoring technology is also developed rapidly, so that an information data chain interaction platform for analysis, integration, standardization and guidance of national large, medium and small bridge health, disease, prevention and diagnosis and treatment integrated management is constructed, and a reliable, efficient and continuously optimized service health and safety guarantee system is provided.

Intelligently monitoring the existing bridge structure; the concept of the existing bridge health monitoring and maintenance platform is gradually realized; in the prior art, a bridge technical condition multi-stage degradation model based on inspection and evaluation of big data is provided, and the model mainly stays in an experimental stage; also, big data-based bridge monitoring information clustering analysis is proposed; the prior art provides a bridge monitoring information classification technology research based on big data. Many successful cases for applying big data to bridge monitoring are not available at home and abroad, and research on the aspect is yet to be carried out.

In conclusion, the big data are applied to bridge health monitoring, and a bridge health monitoring big data center is constructed. The center takes bridge health, safety prevention evaluation and diagnosis information as the basis, takes a modern information technology as support, and can provide scientific guidance for health evaluation, diagnosis, prevention and treatment and the like of bridges and groups thereof in time so as to improve the quality and efficiency of comprehensive prevention and treatment. The information data chain interaction platform realizes the analysis, integration, standardization and guidance of the integrated management of the health, disease, prevention and diagnosis and treatment of the bridge.

Two methods, namely a model-based method and a data-driven method, are used for structural state evaluation and damage identification by using SHM data. The model-based structural state assessment has the following disadvantages:

1) different models are required to be established for different structures;

2) the reference state when the structure is built is unknown;

3) most of model correction is a mathematical process, and the physical interpretability of a result is not strong;

4) manual intervention and judgment are required, and the state of the structure is difficult to identify quantitatively.

The model-based method is essentially a process of bridge structure finite element modeling, model correction and system parameter inversion, has high requirements on the precision of a theoretical model and the quality of monitoring data, and has unsatisfactory application effect in actual engineering at present.

In summary, the problems of the prior art are as follows:

1) the cost of the prior art is high, and the method is only suitable for large and super-large bridges.

2) The prior art has complex system and poor reliability.

3) The prior art is lack of a calibration means and poor in durability.

The difficulty and significance for solving the technical problems are as follows:

in the traditional bridge maintenance management, the structure state assessment based on manual detection plays an important role, however, the manual detection workload is large, the subjectivity is strong, and the long-term quantitative tracking of the structure performance is difficult to realize. In recent years, a Structural Health Monitoring (SHM) technique has been widely used for maintenance management of a large-span bridge. The bridge structure health monitoring is characterized in that sensors are installed on the structure to acquire information of bridge site environment and structural response in real time, and real-time and automatic assessment and even safety early warning are carried out on the technical state of the bridge based on the information. At least 240 large-span bridges in China are currently provided with Structural Health Monitoring Systems (SHMS), and after long-term observation, the Monitoring systems accumulate a large amount of data, and effectively interpret the state of the structure and identify possible damage based on the data, which becomes a key problem in the current SHM research.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a bridge health monitoring system and method based on big data.

The invention relates to a bridge health monitoring method based on big data, which comprises the following steps:

determining the tension and compression constitutive relation and strength of the bridge concrete;

and (3) carrying out a bridge concrete dynamic elastic modulus test: respectively adopting a hyperbolic model and a power function model to obtain characteristic parameters of the concrete constitutive relation;

and determining the mass data of the bridge monitoring through big data analysis software WEKA.

Furthermore, the data classification of the bridge health monitoring method based on big data predicts key parameters of bridge health monitoring; factors influencing deflection, such as dynamic elastic modulus of concrete, Poisson ratio, span of a bridge, inertia moment of a section relative to a neutral axis and the like, and the deflection form an arff file identified by WEKA; data are collected according to monitoring results in the early stage, reasonable classifiers are selected for classification in the follow-up stage, and prediction of bridge deflection can be achieved after new parameters are given.

Further, the data clustering of the bridge health monitoring method based on big data sets the key evaluation parameter threshold value of the bridge health monitoring;

clustering the bridge data by combining the data change rate and a clustering method K-Means, dividing the bridge data into a plurality of clusters, and independently obtaining a part of data which can become abnormal; then, the proportion of abnormal data is utilized to judge the accuracy of the data, and finally, the clustering results of a plurality of sensors are combined to carry out overall statistical evaluation on the accuracy of the data;

determining an abnormal threshold value of a data mining model by using deflection data as basic data when a clustering analysis model is trained; setting according to experience when setting an initial threshold; performing cluster analysis by using the threshold, identifying abnormal data in the data group, and comparing the identified abnormal value with the FEA result of finite element analysis; and judging whether the identified abnormal data is real abnormal data or normal data caused by the small setting of the abnormal threshold value in the data mining model is identified as abnormal data according to the result of the data comparison calculation, if the abnormal data is the abnormal data, repeatedly and iteratively adjusting the abnormal threshold value of the data mining model until the abnormal threshold value reaches a critical point, and determining that the abnormal threshold value of the critical point is the abnormal threshold value of the data mining model of the data group.

Further, the big data based bridge health monitoring method for analyzing the data association rule and optimizing the use of the bridge comprises the following steps:

establishing a state evaluation Apriori model of a given bridge, forming a strong association rule among all attributes of bridge data by analyzing monitored historical data, mining a potential association rule among all attributes, and providing more data-supported bases for evaluation of the bridge state;

performing multi-factor analysis on a plurality of attribute data of the bridge, mining the implicit relevance among the attributes, and then obtaining a fitting formula by fitting calculation through a regression model; after the data correlation analysis is completed, selecting characteristic indexes to construct a bridge service performance evaluation model, setting a reasonable value interval of bridge attributes, and performing real-time early warning when abnormal values appear in bridge monitoring data.

Further, a reasonable value interval of the bridge attribute is set, and when abnormal values appear in the bridge monitoring data, real-time early warning is divided into the following 3 parts:

(1) determining the relevance among the bridge attributes, firstly, carrying out primary analysis on data and recognizing the rule of the data; if the data has more dimensions, performing dimension reduction operation on the data by using a principal component analysis method; then analyzing the influence degree among the attributes of the bridge, and deeply mining the invisible correlation by using a data analysis algorithm on the basis of combining the explicit correlation of the professional knowledge of the bridge structure;

(2) establishing a regression model of associated attributes, fitting data by using a regression method, comparing the fitting degrees of various methods, considering overfitting and underfitting of regression when performing regression analysis on the data, improving the fitting degree by adjusting parameters and other methods when performing underfitting, and performing in-depth analysis on the data to find out a method which is suitable for performing regression fitting on the data;

(3) establishing a service performance evaluation model and an early warning model, and evaluating service performance; the deflection of the bridge is expressed as X under a certain temperature and load condition in the first year of use, and the deflection is expressed as Y under the same temperature and load condition in the second year of use; the damage degree of the bridge is defined according to the corresponding relation, and relevant units are reminded to make corresponding bridge inspection and maintenance when the bridge is seriously damaged; meanwhile, the deflection regression model of multi-factor analysis can be combined to verify whether the deflection data accords with the regression model, and if the deflection data has larger deviation, the bridge is prompted to be abnormal.

Furthermore, the establishment of the early warning model needs to consider two factors, namely, the relevance degree of the influencing factors; second, the current damage condition of the bridge structure. The correlation degree of the influence factors is researched to obtain an effective early warning index.

It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:

Another object of the present invention is to provide a big-data based bridge health monitoring system for operating the big-data based bridge health monitoring method, the big-data based bridge health monitoring system comprising:

the overload identification subsystem is used for identifying malignant overload;

and the calibration subsystem is used for periodically calibrating or verifying the sensor or the detection system so as to avoid the drift of the sensor.

The safety evaluation subsystem comprises state safety evaluation and real-time safety evaluation and is used for providing basic data for the periodic detection of the bridge;

the operation management subsystem is used for realizing bridge information management, bridge detection information management, monitoring information comprehensive management and emergency plans;

and the video monitoring module is used for displaying the actual state of the currently monitored bridge and checking the traffic flow and the stress and damage conditions of the key control section of the bridge in real time.

Further, the state security assessment further comprises:

a response allowable value comparison unit for comparing with the response allowable value of the design material;

the bridge rigidity comparison unit is used for comparing the bridge rigidity with the bridge rigidity of the design data;

the monitoring data change unit is used for changing the history of the monitoring data of the bridge;

the bridge comparison unit is similar to a bridge comparison unit and is used for sequencing the health conditions of the bridge according to the response of the bridge.

Further, the big data based bridge health monitoring system further comprises:

the preprocessing unit is used for realizing data acquisition and field preprocessing;

the data transmission unit is used for realizing data transmission of a public wireless network;

the structure safety evaluation unit is used for realizing the structure safety evaluation of dynamic and static response and database technology;

the dynamic load testing unit is used for realizing dynamic response and calibrated dynamic load testing;

the comparison and evaluation unit is used for realizing comparison, search and point acquisition evaluation of bridges with similar types, geology and structural conditions;

the monitoring unit is used for realizing a video monitoring system;

the system construction unit is used for constructing a mobile browsing and management system based on an Internet framework based on a dual-mode system of a BS and a CS;

and the detection unit is used for realizing the periodic detection of the bridge rigidity.

In summary, the advantages and positive effects of the invention are:

1) uniformly monitoring and managing a plurality of bridges distributed in various places;

2) automatically monitoring environmental conditions and bridge response in real time and visually displaying results;

3) identifying structural damage and carrying out emergency early warning on structural abnormality;

4) identifying overload conditions and linking with systems of related departments;

5) recording the structural state and the long-term change trend, and providing basic data for periodic detection;

an active (self-excited) monitoring system may function and have corresponding features as follows:

1) more sensitive and direct to damage:

2) the test is convenient, the cost is low: the testing distance is variable from several meters to hundreds of meters, no special load is needed, the testing is convenient, the efficiency is high, the testing cost is low, and the traffic can be uninterrupted.

3) High precision and strong objectivity: based on a series of original technologies, the relative error of the test on the elastic modulus can be controlled below 1%, and a solid foundation is provided for evaluating the reliability of the system. Meanwhile, the whole beam can be tested, and the omission problem caused by local detection is effectively prevented;

4) no damage to the bridge: special ballast is not needed, and no damage is caused to the bridge;

5) not only can the bending resistance be tested, but also the shearing resistance can be tested;

6) the long-term detection is convenient: the progress condition of bridge deterioration can be sensitively reflected by a fixed-point detection method;

7) can provide the calibration for the bridge health monitoring system: one of the biggest problems in long-term monitoring systems is the durability of the sensor. During long-term use, the index of the sensor usually changes to some extent. Because the monitoring sensor cannot be simply replaced, the characteristic indexes of the monitoring sensor are often difficult to calibrate, and great influence is brought to the reliability of the monitoring data. By adopting the detection technology, an effective calibration means can be provided for monitoring.

The bridge health monitoring method based on big data provided by the invention realizes data sharing and efficient utilization; the big data is applied to the prediction of the safety and the durability of the concrete structure, so that the accuracy and the speed can be greatly improved; cleaning and correcting the acquired mass data by using a big data mining method, extracting principal components, performing correlation analysis, and predicting the development trend based on a prediction model of machine learning; by applying the big data to the concrete structure health assessment, the functions of structure real-time safety assessment, structure residual service life prediction, structure performance critical point prediction and the like can be realized; therefore, the prediction of the performance critical point of the concrete structure is realized, and scientific suggestions are provided for the maintenance and the repair of the structure.

The invention relates to a bridge health monitoring method based on big data, a theory and a method for deducing and positioning real-time damage of a bridge structure, correcting and evaluating safety of a real-time model based on the big data, and a multilevel criterion for early warning of structural safety; establishing a standard test model for identifying and evaluating the damage of a typical major engineering structure; the method provides theories, methods and a unified inspection platform for health monitoring and safety early warning of major engineering structures.

The bridge health monitoring method based on big data monitors the bridge based on the technology of the Internet of things; carrying out technical evaluation on the bridge based on the bridge bearing capacity of the dynamic elastic modulus; analyzing the bridge data by utilizing big data classification, clustering and association rule data mining technologies; carrying out system evaluation on the safety of the bridge from multiple directions; carrying out system evaluation on the bridge based on the structural damage of the vibration response; and carrying out system evaluation on the bridge based on the state of the monitoring historical data.

Drawings

Fig. 1 is a flowchart of a bridge health monitoring method based on big data according to an embodiment of the present invention.

FIG. 2 is a block diagram of a big data based bridge health monitoring system subsystem according to an embodiment of the present invention.

Fig. 3 is a schematic view of an overload monitoring system (only installed on an approach bridge) provided by an embodiment of the invention.

Fig. 4 is a flowchart of an overload monitoring analysis provided by an embodiment of the invention.

FIG. 5 is a flow chart of a security assessment scheme (as compared to design values) provided by an embodiment of the present invention.

Fig. 6 is a schematic structural diagram of a big data-based bridge health monitoring system according to an embodiment of the present invention.

FIG. 7 is a flow chart of big data analysis of bridge health monitoring according to an embodiment of the present invention.

FIG. 8 is a schematic diagram of dimension reduction of data mining provided by an embodiment of the present invention.

Fig. 9 is a schematic diagram of obtaining an early bridge defect through feature extraction according to an embodiment of the present invention.

Fig. 10 is a conceptual diagram of a system according to an embodiment of the present invention.

FIG. 11 is a closed loop mode of bridge monitoring/detection provided by an embodiment of the present invention.

Fig. 12 is a schematic view of a monitored object according to an embodiment of the present invention.

Fig. 13 is a structural relationship comparison diagram of C50 and C60 concretes commonly used in bridges in 1 year and 10 years according to the embodiment of the invention.

Fig. 14 is a concept of a rapid test (longitudinal test) of a prestressed girder according to an embodiment of the present invention.

Fig. 15 is a concept of a prestressed girder rapid test (transverse, capping beam) provided by an embodiment of the present invention.

Fig. 16 illustrates a concept of rapid bridge pier testing (transmission, top reflection, and side wall reflection) according to an embodiment of the present invention.

FIG. 17 is a schematic diagram of the tangential elastic modulus-x relationship when different grades of concrete are eccentrically tensioned according to an embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The cost for the prior art is high; the system is complex and has poor reliability; lack of calibration means and poor durability. The invention carries out system evaluation on the safety of the bridge from multiple directions; carrying out system evaluation on the bridge based on the structural damage of the vibration response; and carrying out system evaluation on the bridge based on the state of the monitoring historical data.

The following detailed description of the principles of the invention is provided in connection with the accompanying drawings.

As shown in fig. 1, the bridge health monitoring method based on big data provided by the embodiment of the present invention includes the following steps:

s101: the method includes the steps that research theories about bridge health monitoring at home and abroad, relevant test data and design data analysis are collected, and hot spots and frontiers in the aspect of bridge monitoring at home and abroad are mastered;

s102: determining the tension and compression constitutive relation and strength of the bridge concrete;

s103: and (3) carrying out a bridge concrete dynamic elastic modulus test: respectively adopting a hyperbolic model and a power function model to obtain characteristic parameters of the concrete constitutive relation;

s104: and determining the mass data of the bridge monitoring through big data analysis software WEKA.

In step S104, the big data analysis software WEKA provided by the present invention determines the mass data of bridge monitoring, and includes:

a) data classification, namely predicting key parameters of bridge health monitoring;

explaining by taking deflection as an example, factors influencing the deflection, such as the dynamic elastic modulus of concrete, the Poisson ratio, the span of a bridge, the inertia moment of a section relative to a neutral axis and the like, and the deflection form an arff file identified by WEKA; data are collected according to monitoring results in the early stage, and reasonable classifiers are selected for classification subsequently, so that the bridge deflection can be predicted after new parameters are given.

b) Data clustering, namely setting a key evaluation parameter threshold value for bridge health monitoring;

clustering the bridge data by combining the data change rate and a clustering method K-Means, dividing the bridge data into a plurality of clusters, and independently obtaining a part of data which can become abnormal; and finally, performing overall statistical evaluation on the accuracy of the data by combining clustering results of a plurality of sensors.

And when the clustering analysis model is trained, determining an abnormal threshold value of the data mining model by using the deflection data as basic data. When setting the initial threshold value, the setting is generally performed empirically. Clustering analysis is performed by using the threshold, abnormal data in the data group is identified, and the identified abnormal value is compared with a Finite Element Analysis (FEA) result. And judging whether the identified abnormal data is real abnormal data or normal data caused by the small setting of the abnormal threshold value in the data mining model is identified as abnormal data according to the result of the data comparison calculation, if the abnormal data is the abnormal data, repeatedly and iteratively adjusting the abnormal threshold value of the data mining model until the abnormal threshold value reaches a critical point, and determining that the abnormal threshold value of the critical point is the abnormal threshold value of the data mining model of the data group.

c) And analyzing data association rules to optimize the use of the bridge.

An Apriori model for evaluating the state of a given bridge is established, strong association rules among all attributes of bridge data are formed through analysis of monitored historical data, potential association rules among all attributes are mined, and more data supporting bases are provided for evaluation of the bridge state.

And carrying out multi-factor analysis on a plurality of attribute data of the bridge, mining the implicit relevance among the attributes, and then obtaining a fitting formula by fitting and calculating a regression model. After the data correlation analysis is completed, selecting characteristic indexes to construct a bridge service performance evaluation model, setting a reasonable value interval of bridge attributes, and performing real-time early warning when abnormal values appear in bridge monitoring data. The total number of the components is divided into the following 3 parts:

1) determining associations between bridge attributes

The part firstly carries out preliminary analysis on the data and recognizes the rule of the data. And if the data has more dimensions, performing dimension reduction operation on the data by using a principal component analysis method. And then analyzing the influence degree among the attributes of the bridge, and deeply mining the invisible correlation by using a data analysis algorithm on the basis of combining the explicit correlation of the professional knowledge of the bridge structure.

2) Establishing regression models of associated attributes

The part mainly applies regression method to fit data, compares the fitting degrees of various methods, needs to consider overfitting and underfitting of regression when carrying out regression analysis on the data, needs to improve the fitting degree by methods such as parameter adjustment and the like when carrying out regression underfitting, carries out in-depth analysis on the data, and finds out the method which is suitable for carrying out regression fitting on the data.

3) Establishing a service performance evaluation model and an early warning model

First, the service performance is evaluated. For example, a bridge exhibits X deflection at a certain temperature and load during the first year of use, and Y deflection at the same temperature and load during the second year of use. The damage degree of the bridge is defined according to the corresponding relation, and relevant units are reminded to make corresponding bridge inspection and maintenance when the bridge is seriously damaged. Meanwhile, the deflection regression model of multi-factor analysis can be combined to verify whether the deflection data accords with the regression model, and if the deflection data has larger deviation, the bridge is prompted to be abnormal.

Secondly, two factors need to be considered for establishing the early warning model, namely the relevance degree of the influencing factors; second, the current damage condition of the bridge structure. The correlation degree of the influence factors is researched to obtain an effective early warning index.

As shown in fig. 2, a bridge health monitoring system based on big data includes: an overload identification subsystem, a calibration (bridge rigidity test) subsystem, a safety evaluation subsystem and an operation management subsystem;

(1) overload identification subsystem

The overload identification subsystem (as shown in fig. 3) identifies the malignant overload from the perspective of bridge safety without changing the bridge deck structure and affecting traffic. The overload identification subsystem has the same structure as the bridge health monitoring system. But requires a modest increase in the number of sensors and incorporation of a video monitoring system. The overload monitoring analysis flow is shown in fig. 4.

(2) Calibration (bridge rigidity test) subsystem

One of the most distressing problems in bridge monitoring systems is the drift of the sensors. That is, under the action of various factors, the sensor is aged so that the basic parameters of the sensor are changed. Therefore, in general tests and inspections, it is often required to periodically calibrate or verify the sensor or the inspection system. As a sensor installed on a bridge, it cannot be simply calibrated or verified, so that the reliability of the measured index is often difficult to guarantee, which causes great difficulty in reasonably evaluating the health status of the bridge.

(3) Security assessment subsystem

The safety evaluation system comprises state safety evaluation and real-time safety evaluation, and meanwhile, necessary basic data are provided for the periodic detection of the bridge. The security assessment scheme is shown in fig. 5.

1) And (3) state safety evaluation: according to the tested parameters and the comparison of design data, historical data and similar bridges, comprehensively evaluating the safety state of the bridge;

a) comparing with the response allowable value of the design data (mainly the dynamic deflection, the natural vibration frequency and the like of the beam);

b) comparing with the rigidity of the bridge with the design data (mainly dynamic and static deflection of the beam, etc.);

c) history of changes in the monitored data of the bridge: variation of prominent frequency (particularly pier), average value and maximum value variation of acceleration and dynamic deflection, and the like;

d) comparison of similar bridges: the bridge type of the prestressed concrete girder is relatively simple, and as described above, the prestressed concrete girder may be classified according to its bridge type, single span, linearity, the number of lanes, the type of the girder, the type of pier, whether the pier is crossed, and the like. In the same category, the health conditions of the bridges can be ranked according to the responses of the bridges, and the bridges at the lower level of the order should be paid sufficient attention.

2) And (3) real-time safety evaluation: after an emergent event (such as the passing of a severely overloaded vehicle, collision and earthquake) occurs, rapidly evaluating the health condition and the damage condition of the bridge; the security assessment system supports both online and offline assessment.

(4) Operation management subsystem

a) Bridge information management: bridge position, structure, design data, etc.;

b) bridge detection information management: the method comprises manual inspection and special detection results;

c) comprehensively managing monitoring information;

d) an emergency plan;

(5) and the video monitoring module is used for displaying the actual state of the currently monitored bridge and checking the conditions of traffic flow, stress, damage and the like of the key control section of the bridge in real time.

Further, a bridge health monitoring system based on big data, this system uses vibration monitoring as the core, combines the elastic modulus of periodic test roof beam, board, includes:

1) data acquisition and field pretreatment;

2) data transmission based on a public wireless network;

3) structural security assessment based on dynamic and static responses and database techniques;

4) dynamic load (load carrying load) test based on dynamic response and calibration;

5) comparing, searching and point-picking evaluation of bridges with similar types, geology and structural conditions;

6) a video monitoring system (linked with vibration monitoring and the like);

7) a dual-mode system based on BS and CS, and a mobile browsing and management system based on Internet architecture can be constructed;

8) periodic measurement of bridge rigidity (modulus of elasticity of top plate and web plate of concrete bridge).

As shown in fig. 6, the schematic diagram of the bridge health monitoring system based on big data provided by the present invention;

the bridge health monitoring method comprises the following steps of carrying out health monitoring on a bridge through a plurality of sensors, collecting monitoring data of the bridge through the bridge sensors, transmitting the bridge data collected by the sensors to a bridge monitoring center, transmitting the bridge data to a bridge health monitoring data center through the bridge monitoring center, and summarizing the data of each bridge monitoring center through the bridge health monitoring data center; determining the collected data by utilizing big data analysis software WEKA open source software, and transmitting the data to a bridge health monitoring big data center analysis processing system; data of bridge design, construction, detection, maintenance and the like are input into a bridge health monitoring big data center analysis processing system by using common development software; books, articles and reports of bridge design, construction, accidents, diseases and the like are transmitted to a bridge health monitoring big data center analysis processing system by using open source software such as WEKA and the like through a search engine and a data mining algorithm in the world.

As shown in fig. 7, the big data analysis of the bridge health monitoring includes the following specific processes: data cleaning, data fusion, data dimension reduction, feature extraction, mode identification, prediction and visualization;

first, data cleansing of SHM data

1) Data cleaning: replace, modify, delete incorrect, inaccurate, irrelevant data. Filtering, resampling and other methods are adopted;

2) data normalization: the changing components due to environmental and operational load changes are filtered out from the data of the structural response.

Further, the object is preprocessed: noise, missing values, outliers, drift; the cleaning method comprises the following steps: band-pass filtering, Hi-Huang transform, blind source separation, wavelet transform, moving average; data normalization: the changing components due to environmental and operational load changes are filtered out from the data of the structural response.

Second, data fusion of SHM data

Data fusion (Data fusion) is a process of dominating and integrating multi-source heterogeneous Data (structured Data, semi-structured Data, unstructured Data) by using a logic algorithm, a computer technology and the like so as to obtain a more accurate and comprehensive information result.

The data fusion of the SHM data is mainly to integrate data and information of the bridge maintenance management system and the health monitoring system, which are mainly checked manually, so as to obtain a result with more consistency and reliability. By integrating multi-source heterogeneous data, the multi-dimensional and multi-granularity association relation among data, information and knowledge is established, and multi-level knowledge interaction is realized.

Third, data dimension reduction of SHM data (as shown in FIG. 8)

Dimension reduction (dimensional reduction): the method is a process for reducing the number of random variables under certain limited conditions to obtain a group of irrelevant main variables. The method mainly comprises the following steps:

1) selecting characteristics: redundant and irrelevant features are removed;

2) feature extraction: principal Component Analysis (PCA); singular Value Decomposition (SVD); a Fourier transform (FFT); and (5) performing wavelet transformation.

And the data dimension reduction of the SHM data mainly comprises the following steps:

1) fourier transform;

2) and (5) performing wavelet transformation.

Fourthly, extracting the characteristics of the SHM data

Feature extraction (Feature extraction) for data mining: refers to a process of converting raw data that cannot be recognized by a machine learning algorithm into numerical features that can be recognized by the algorithm. The early bridge defect obtained by feature extraction is shown in fig. 9.

1) Text mining, vocabulary → vectors;

2) picture identification, pixel → matrix;

the feature extraction of the SHM data converts the data into different types of information. Such as modal features extracted by modal analysis, statistical features extracted by statistical analysis, damage indexes of damage identification, and the like. And extracting the index sensitive to the structural damage or the structural state, and determining whether the data analysis is the most important factor.

1) Modal characteristics: frequency, modal strain energy;

2) statistical characteristics: variance, mean, spectral density;

3) frequency domain characteristics: frequency, mode shape, modal curvature;

4) and others: regression residual, wavelet energy, and fitting coefficients;

5) the new technology comprises the following steps: deep learning

The feature selection is mainly to remove redundant, irrelevant features, etc.

As shown in fig. 10, the bridge health monitoring system takes a plurality of prestressed concrete bridges as monitoring objects, and includes two types:

passive type monitoring system: the system is the same as the existing bridge health monitoring system, and the health state of the bridge is inferred by continuously testing various physical quantities generated by traffic and other loads in the service process of the bridge;

active (also called "self-excited") monitoring systems: a novel monitoring system is characterized in that a specific signal is excited and received through a signal excitation device installed on a bridge, and the health state of the bridge is deduced according to the change of parameters of the signal.

The two methods have advantages and disadvantages respectively, and can be used independently or jointly. Meanwhile, the practical value and the cost performance of the system are improved by optimizing the indexes of effectiveness, reliability, economy, maintenance, durability and the like of the system. The closed loop mode of bridge monitoring/detection is shown in fig. 11.

As shown in fig. 12, the main monitoring items of the passive monitoring system are:

1) vibration characteristics of the bridge: analyzing to obtain the natural vibration frequency of the bridge based on the monitored acceleration signal;

2) vibration characteristics of a bridge pier: obtaining the self-vibration frequency of the pier by adopting an independent analysis method on the basis of the monitored acceleration signal;

3) the temperature of the bridge;

4) the digital video is used for monitoring the inclination of the bridge in an auxiliary manner according to the situation;

5) and the elastic modulus of the prestressed beam, the prestressed plate and the prestressed pier is tested periodically.

TABLE 1 Effect of the sensors

The technical effects of the present invention will be further described in detail with reference to specific tests below:

the size of the model is length multiplied by width multiplied by thickness (185 multiplied by 20 multiplied by 10cm), the load direction is parallel to the pouring surface, namely the structural size in the load direction is 20 cm; reinforcing the model (steel plate reinforcement and carbon fiber Cloth (CFRP) reinforcement);

and (3) load test: and obtaining dynamic elastic modulus of the bridge concrete C50 and C60 by applying 30KN static load for 5 days, 30KN static load for 15 days, no static load for 15 days and 30KN static load for 5 days.

The invention relates to a theory and a method for bridge structure real-time damage inference and positioning, real-time model correction and safety evaluation based on big data, and a multilevel criterion for structure safety early warning; establishing a standard test model for identifying and evaluating the damage of a typical major engineering structure; the method provides theories, methods and a unified inspection platform for health monitoring and safety early warning of major engineering structures.

1. Theory and method for structure real-time damage inference, positioning and model correction

(1) Environmental time-varying action model:

(a) determining the amplitude of day-night temperature difference change and seasonal temperature difference change and the number of times of circulating action of the amplitude, and determining the influence of temperature on the static and dynamic characteristics of the continuous rigid frame bridge;

(b) influence of the prestress on the dynamic properties of the statically indeterminate structure.

(2) The method for identifying the sub-structure and the decentralization of the damage of the complex structure comprises the following steps:

the method is used for determining the inference of local damage, substructure damage and scattered damage of a structure aiming at large-scale guyed bridge structures and other major engineering structures with obvious substructure characteristics (such as relatively independent and organically-linked substructure systems such as guys, bridge decks, bridge towers and the like).

(3) The structural damage identification method based on the non-physical model comprises the following steps:

a structural damage identification method based on a non-physical model is determined by adopting a modern signal processing technology and an artificial intelligence method, and mainly comprises a wavelet packet transformation analysis method, a Hilbert-Huang transformation analysis method, a neural network method and the like.

(4) Theory and method of structural model correction:

and determining an optimized objective function and constraint conditions corrected by the structural model on the basis of the structure damage inference and positioning.

2. Theory and method for assessing structural health state

(1) Load standard for structural safety assessment:

bridge evaluation is obviously different from bridge design, and design load standard is comprehensively obtained based on statistical analysis. Researching the structure extreme value environment effect based on the environment condition monitoring; determining the probability exceeding criterion of random environmental loads of the structural design service life and the subsequent service life and the structural safety evaluation load standard based on the criterion; and researching specific evaluation load standards of earthquake, strong wind, sea wave and the like.

(2) Relationship between accumulated damage and resistance attenuation of typical major engineering structures:

the method comprises the steps of determining a rule of accumulated damage of a key structural member, a degradation rule of the structural member and the overall performance and a resistance attenuation model aiming at a large-scale stay cable bridge structure and a fixed steel jacket ocean platform structure.

(3) And (3) carrying out real-time safety evaluation on the structure:

the relationship between a vulnerable component and an important component and a structural failure mode and the corresponding limit bearing capacity is determined by combining a typical major engineering structure; and determining a structural safety evaluation method based on the current structural damage condition and the evaluation load standard.

(4) Predicting the residual service life of the bridge:

(a) establishing a bridge system resistance model;

(b) a calculation method and an identification technology for determining the occurrence probability of the structure-dominated failure mode.

3. Multilevel criterion of structural safety pre-warning:

(1) structure early warning level decision: determining a failure mechanism, a failure mode and the minimum safety margin of the structure by combining a typical major engineering structure; according to the functions of different states of the structure, determining the establishment criterion and standard of the multilevel safety early warning level of the structure and a threshold value adjusting method based on damage process control;

(2) the structure safety early warning method based on the prior knowledge comprises the following steps: determining a theory and a method for rapidly predicting the catastrophe response of the structure according to the structural vulnerability analysis, the structural failure path and critical state and the structural safety early warning method of the damage fingerprint;

(3) and establishing a long and large bridge safety early warning system.

The technical scheme of the invention is further explained by combining the following embodiments:

the damage of the prestressed concrete bridge has many aspects, in terms of the macroscopic view of the beam body, the damage is reflected in that the downwarping quantity is increased and the bearing capacity (strength) is reduced, and the two aspects supplement each other. From the microscopic level, the method is represented by:

(1) loss of prestress: due to the corrosion, the relaxation and the like of the steel strand, the effective prestress is reduced;

(1) deterioration of concrete material: under the action of fatigue load and external environment, the rigidity and strength of the concrete are reduced;

(3) corrosion of reinforcing steel bars

The prestress loss is an important characteristic of a prestressed concrete bridge different from a reinforced concrete bridge.

Since the bridge is subjected to a large cyclic load, the concrete material is subject to a deterioration phenomenon, so-called "fatigue", under the repeated load and the environmental load.

The concrete material is a heterogeneous multi-phase medium, and a large number of micro cracks, micro gaps and the like are generated inside the concrete material due to hydration heat, uneven quality and the like in the construction period. The defects are initial damage of concrete, and the number and the positions of the initial cracks are related to the type of cement, the mineral property and the geometric characteristic of aggregate, the water cement ratio, the curing condition and the like. After the concrete structure with the microcracks is subjected to cyclic load, the microcracks are changed, expanded and connected and are stabilized on a certain cracking level or finally destroyed according to the condition of the load. When the cyclic load is small, the micro-cracks only have small deformation and expansion, and the concrete can be considered to be in an elastic working stage; when the cyclic load exceeds a certain level, the stress gradient around some micro cracks in the structure is increased, so that the instability is expanded, and the crack can be connected and communicated with an adjacent crack to form a crack with a larger scale.

Meanwhile, the improvement of the internal stress level of the structure can also cause the bonding micro-damage of some bonding weak areas and generate new micro-cracks. When the microcracks interconnect and penetrate to some extent, macrocracks are formed. At this time, if the cyclic load continues to act, new microcracks continue to appear and develop, new macrocracks are formed, and meanwhile, the formed macrocracks continue to expand. When the cycle number is increased to a certain value, the macrocrack propagation rate is accelerated, and the unstable stage is entered, so that the concrete is rapidly damaged. Thus, the entire fatigue process is a process in which damage gradually accumulates.

Bridge deterioration (reduced load bearing capacity) is manifested in a number of ways, such as increased downwarp, increased vibration, cracking, etc. The phenomena are finally concluded, and the reduction of the rigidity of the bridge is reflected.

Therefore, if the elastic modulus of the concrete material and the change condition of the crack surface can be tested, the rigidity of the bridge and the damage condition can be estimated and evaluated. Fig. 13 shows a structural relationship comparison graph of C50 and C60 concretes commonly used in bridges in 1 year and 10 years.

The development of the method for testing the shock elastic waves provides a theoretical and technical basis for the test. According to the theory of elastic waves, there is a clear relationship between the wave velocity and the dynamic tangent elastic modulus of the material, and at the same time, there is a close correlation between the elastic wave velocity, attenuation and cracks due to the interruption of the crack surface. Therefore, by testing the wave velocity and attenuation of the shock elastic wave, the elastic modulus and the structural crack state of the concrete can be theoretically calculated. The method specifically comprises the following steps:

(1) testing of the elastic modulus of concrete based on the velocity of the impact elastic wave P wave, as shown in fig. 14;

(2) a test for the presence or absence of a concrete crack based on the change in the wave velocity of the P-wave of the ballistic elastic wave, as shown in fig. 15;

(3) test of concrete crack depth based on shock elastic wave, as shown in fig. 16.

TABLE 2 typical parameters of concrete strength (strength unit: MPa)

Meaning and method for testing parameters

(1) Testing of concrete modulus of elasticity

The elastic modulus of the concrete can reflect the quality and damage condition of the material and the stress level of the concrete. The testing method is based on the P wave velocity of the impact elastic wave and carries out corresponding steel bar and shape correction.

(2) Test for Presence of concrete crack

The elastic wave is shielded when meeting the fracture surface in the propagation process. Therefore, the propagation wave velocity of the elastic wave is reduced when the elastic wave passes through the fracture region. Since the compressive modulus of the fracture faces is generally much greater than the tensile modulus thereof, the fracture faces less resistance to the compression elastic wave than to the tension elastic wave, so that the wave velocity of the tension elastic wave is generally lower than that of the compression elastic wave. On the other hand, in the absence of cracks, the wave velocities of the two elastic waves are the same. Accordingly, the presence or absence and the degree of the crack surface can be estimated by analyzing the change in the wave velocity of the tensile and compressive elastic waves.

(3) Testing of concrete crack depth

The reasonable grasp of the crack depth has very important significance in the aspects of judging the damage degree of the structure, calculating the section inertia moment, bearing capacity and the like. The tangential elastic modulus-x relationship when the concrete of different grades is eccentrically pulled is shown in FIG. 17.

The rapid evaluation method of the prestressed concrete bridge has the characteristics that:

the test is convenient, the cost is low: the testing distance is variable from several meters to hundreds of meters, no special load is needed, the testing is convenient, the efficiency is high, the testing cost is low, and the traffic can be uninterrupted.

High precision and strong objectivity: based on a series of original technologies, the relative error of the test on the elastic modulus can be controlled below 1%, and a solid foundation is provided for evaluating the reliability of the system. Meanwhile, the whole beam can be tested, and the omission problem caused by local detection is effectively prevented;

no damage to the bridge: special ballast is not needed, and no damage is caused to the bridge;

the long-term detection is convenient: the progress condition of bridge deterioration can be sensitively reflected by a fixed-point detection method;

can provide the calibration for the bridge health monitoring system: one of the biggest problems in long-term monitoring systems is the durability of the sensor. During long-term use, the index of the sensor usually changes to some extent. Because the monitoring sensor cannot be simply replaced, the characteristic indexes of the monitoring sensor are often difficult to calibrate, and great influence is brought to the reliability of the monitoring data. By adopting the detection technology, an effective calibration means can be provided for monitoring.

The bridge health monitoring main content based on big data comprises the following steps:

clustering threshold parameters of key parameters (deflection, frequency and the like) of bridge deformation development by utilizing WEKA from a large amount of historical data and social data, and providing related parameters for bridge health monitoring in the later period;

obtaining performance parameters according to a mechanical experiment of concrete;

on the basis of historical bridge monitoring parameters, forecasting of bridge monitoring parameters is realized by utilizing a WEKA big data time sequence function, and once the monitoring parameters reach a critical threshold value, an alarm is given;

and (3) integrating bridge monitoring key parameters such as deflection, natural vibration frequency, concrete dynamic elastic modulus and the like, constructing an arff file supported by WEKA, and realizing prediction of the residual service life of the bridge.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims

1. A bridge health monitoring method based on big data is characterized by comprising the following steps:

2. The big-data-based bridge health monitoring method according to claim 1, wherein the big-data-based bridge health monitoring method is characterized by data classification and prediction of key bridge health monitoring parameters; factors influencing deflection, such as dynamic elastic modulus of concrete, Poisson ratio, span of a bridge, inertia moment of a section relative to a neutral axis and the like, and the deflection form an arff file identified by WEKA; data are collected according to monitoring results in the early stage, reasonable classifiers are selected for classification in the follow-up stage, and prediction of bridge deflection can be achieved after new parameters are given.

3. The big data based bridge health monitoring method according to claim 1, wherein the data clustering of the big data based bridge health monitoring method is used for setting a key evaluation parameter threshold value for bridge health monitoring;

4. The big-data-based bridge health monitoring method of claim 1, wherein the big-data-based bridge health monitoring method analyzes data association rules and optimizes bridge usage comprises:

5. The bridge health monitoring method based on big data as claimed in claim 4, wherein a reasonable value interval of the bridge attribute is set, and the real-time early warning is divided into the following 3 parts when the bridge monitoring data has an abnormal value:

6. The bridge health monitoring method based on big data as claimed in claim 4, wherein the early warning model is established by considering two factors, namely correlation degree of the influence factors; secondly, the current damage condition of the bridge structure is researched, and the correlation degree of the influence factors is researched to obtain effective early warning indexes.

7. A computer device, characterized in that the computer device comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of:

8. A big-data-based bridge health monitoring system for operating the big-data-based bridge health monitoring method according to any one of claims 1 to 6, wherein the big-data-based bridge health monitoring system comprises:

9. The big-data based bridge health monitoring system of claim 8, wherein the state safety assessment further comprises:

10. The big-data based bridge health monitoring system of claim 8, further comprising:

the monitoring unit is used for realizing a video monitoring system;