A structural health status early warning method based on bridge associated section monitoring data

Abstract

The present invention discloses a structural health status early warning method based on bridge associated section monitoring data, and relates to the technical field of bridge monitoring. The sensor acquisition data used to monitor the bridge in the target bridge is obtained, and the basic probability assignment of the sensor at each loading point is calculated as a reference response based on the acquired data; a first fusion probability assignment matrix is ​​obtained based on multiple types of reference responses, and the probability function of each loading point is determined; the probability function of each loading point under different monitoring sections is subjected to a second fusion of data to obtain a second fusion probability assignment matrix; the fusion probability function at each loading point is determined based on the second fusion probability assignment matrix, and structural damage diagnosis and safety early warning are performed based on the fusion probability function. By performing fusion early warning calculations on the data of multiple sections and multiple sensors, the associated benchmark of the associated section data is obtained, thereby greatly reducing the false alarm phenomenon of a single section and a single sensor, and effectively improving the accuracy and reliability of bridge early warnings.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge monitoring, and in particular to a structural health status early warning method based on bridge associated cross-section monitoring data. Background Art

[0002] As the throat of traffic, bridges play a vital role in the transportation industry. However, due to unreasonable structural design, construction technology, daily maintenance, and the influence of natural and human factors such as acid rain, earthquakes, and overloading, some bridges have various diseases, which even seriously threaten the daily safe operation of bridges.

[0003] Timely and accurate alarm for bridge safety is the core goal of building a bridge monitoring system. In addition to the conventional graded alarm and high-risk alarm for bridge structure safety hazards, alarm functions are set for extreme environments and overload phenomena, two factors that seriously affect bridge safety.

[0004] The existing alarm method usually uses a single cross-section sensor to set a threshold value for early warning. This method has a single indicator and the structural early warning is too dependent on the accuracy of the threshold setting. Therefore, it is often impossible to detect structural safety hazards in a timely manner, and the accuracy and reliability of bridge early warning are low. Summary of the invention

[0005] The purpose of the present invention is to solve the above-mentioned background technology problems and to propose a structural health status early warning method based on bridge associated section monitoring data.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An embodiment of the present invention provides a structural health status early warning method based on bridge associated section monitoring data, the method comprising:

[0008] Acquire data collected by various types of sensors used for monitoring the bridge in the target bridge, and calculate the basic probability assignment of each sensor at each loading point as a reference response according to the collected data of each sensor; the loading point is the position of a vehicle traveling on the target bridge;

[0009] The benchmark responses of multiple types of sensors are first fused to obtain a first fusion probability assignment matrix of the multi-source sensors; each row of the first fusion probability assignment matrix represents the benchmark responses of the same sensor at different loading points, and each column of the first fusion probability assignment matrix represents the benchmark responses of different types of sensors at the same loading point;

[0010] Determine the probability function of each loading point according to the first fused probability assignment matrix;

[0011] The probability functions of each loading point under different monitoring sections are used as basic probability assignments to perform a second fusion of the data to obtain a second fusion probability assignment matrix of the target bridge; each row of the second fusion probability assignment matrix represents the probability functions at different loading points under the same monitoring section, and each row of the second fusion probability assignment matrix represents the probability functions at the same loading point under different monitoring sections;

[0012] The fusion probability function at each loading point is determined according to the second fusion probability assignment matrix, and structural damage diagnosis and safety warning are performed according to the fusion probability function.

[0013] Optionally, the basic probability assignment of each sensor at each loading point is calculated as a reference response according to the collected data of each sensor, including:

[0014] Calculate the difference between the collected data of each sensor at each loading point and the preset lossless data as the damage value;

[0015] According to the damage value, the basic probability assignment of each loading point is calculated as the benchmark response:

[0016]

[0017] Among them, x j and x k denote the jth and kth loading points respectively, m(x j ) represents x j The basic probability assignment, ΔG(x j ) and ΔG(x k ) represent x j and x k The damage value at , m is the total number of loading points.

[0018] Optionally, basic probability assignments of multiple types of sensors are first fused to obtain a first fusion probability assignment matrix of multi-source sensors, including:

[0019] For each sensor, the reference response of each loading point is divided by the maximum reference response among the loading points to obtain the normalized and dimensionless reference response;

[0020] The normalized and dimensionless reference responses are fused for the first time to obtain the first fusion probability assignment matrix of the multi-source sensors.

[0021] Optionally, determining a probability function of each loading point according to the first fusion probability assignment matrix includes:

[0022] Taking the orthogonal sum of the first fusion probability assignment matrix to obtain the basic probability function of each loading point;

[0023] According to the position coordinates of the sensor and the position coordinates of the loading point, the distance weights corresponding to the sensor and each loading point are calculated;

[0024] The distance weight of each loading point is used to correct the basic probability function of the loading point to obtain the probability function of the loading point.

[0025] Optionally, the first fusion probability assignment matrix is ​​orthogonally summed to obtain the basic probability function of each loading point:

[0026]

[0027] Among them, x j represents the jth loading point, Represents x j The basic probability function, m i (x j ) represents the reference response of the i-th type of sensor at the j-th loading point, n is the number of sensors, and m is the total number of loading points.

[0028] Optionally, calculating the distance weights corresponding to the sensor and each loading point according to the position coordinates of the sensor and the position coordinates of the loading point includes:

[0029] According to the position coordinates of the sensor and the position coordinates of the loading point, the spatial distance between the sensor and the loading point is calculated;

[0030] According to the spatial distance between the sensor and the loading point, the distance weight between the sensor and the loading point is calculated:

[0031]

[0032] Where n is the number of sensors, D ij is the spatial distance between the sensor and the loading point.

[0033] Optionally, the distance weight of each loading point is used to correct the basic probability function of the loading point, and the probability function of the loading point is obtained as follows:

[0034]

[0035] M0(x j ) represents x j The probability function of .

[0036] Optionally, multiple types of sensors include static deflection sensors, acceleration amplitude sensors, and strain sensors.

[0037] Beneficial effects of the present invention:

[0038] The embodiment of the present invention provides a structural health status early warning method based on bridge-associated section monitoring data, which obtains collected data of multiple types of sensors used to monitor the bridge in a target bridge, and calculates the basic probability assignment of the sensor at each loading point as a reference response based on the collected data of each sensor; the loading point is the position of a vehicle traveling on the target bridge; the reference responses of the multiple types of sensors are first fused to obtain a first fusion probability assignment matrix of multi-source sensors; each row of the first fusion probability assignment matrix represents the reference response of the same sensor at different loading points, and each column of the first fusion probability assignment matrix represents the reference response of different types of sensors at the same loading point; the probability function of each loading point is determined according to the first fusion probability assignment matrix; the probability function of each loading point under different monitoring sections is used as the basic probability assignment to perform a second fusion of the data to obtain a second fusion probability assignment matrix of the target bridge; each row of the second fusion probability assignment matrix represents the probability function at different loading points under the same monitoring section, and each row of the second fusion probability assignment matrix represents the probability function at the same loading point under different monitoring sections; the fusion probability function at each loading point is determined according to the second fusion probability assignment matrix, and structural damage diagnosis and safety early warning are performed according to the fusion probability function. By fusing warning calculations on data from multiple sections and multiple sensors, the correlation benchmark of the associated section data is obtained, which greatly reduces the false alarm phenomenon caused by a single section and a single sensor, and effectively improves the accuracy and reliability of bridge warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The present invention will be further described below in conjunction with the accompanying drawings.

[0040] Figure 1 A flow chart of a structural health status early warning method based on bridge associated section monitoring data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Timely and accurate alarm for bridge safety is the core goal of building a bridge monitoring system. In addition to the conventional graded alarm and high-risk alarm for bridge structure safety hazards, alarm functions are set for extreme environments and overload phenomena, two factors that seriously affect bridge safety.

[0043] 1. Bridge monitoring threshold setting

[0044] The sensor alarm thresholds of the bridge safety monitoring system include dynamic thresholds and fixed thresholds.

[0045] The calculation of the fixed threshold is mainly based on the current specification requirements, bridge load test results and model calculation results, and the fixed alarm threshold is determined by mathematical statistics methods.

[0046] The dynamic threshold is calculated in real time using a time-varying fitting method to obtain the threshold setting at the lowest level that characterizes the normal working range of the sensor.

[0047] In order to realize the analysis and alarm of sensor monitoring data, it is necessary to analyze the sensor collected data and derive three levels of thresholds (level one maximum value, level one minimum value, level two maximum value, level two minimum value, level three maximum value, level three minimum value, level one represents the highest level and the most serious situation).

[0048] 2. Structural safety graded alarm

[0049] The graded alarm for bridge structure safety is the most important function of the alarm system. The health grading standard used in the alarm is to rate the entire bridge and components based on the evaluation and prediction results. The monitoring system has a built-in database of a series of components and units, and grades the structure according to predefined specifications, and also performs graded alarms based on the set grade thresholds. The graded alarm levels are divided into three levels according to the threshold design model:

[0050] (1) Level 3 alarm is an alarm that exceeds the threshold of normal use range, indicating that the structure exceeds the normal use state. The threshold is obtained by mathematical statistics of monitoring data.

[0051] (2) The second level alarm is an alarm that exceeds the normal service limit state, indicating that the structure exceeds the normal limit bearing capacity and is in an unsafe state. The threshold is calculated by the bridge load test or the modified finite element model.

[0052] (3) The first level alarm is an alarm for exceeding the ultimate bearing capacity, indicating that the structure exceeds the ultimate bearing capacity, resulting in an unsafe state. The threshold value is obtained from the specification value or design value.

[0053] (4) When a sudden danger occurs to the bridge structure, the system will issue a high-risk alarm for the bridge structure.

[0054] system

[0055] The three-level threshold settings for the main monitoring indicators are as follows:

[0056] Table 1 Monitoring indicator threshold setting table

[0057]

[0058]

[0059] The embodiment of the present invention provides a structural health status early warning method based on bridge associated section monitoring data. Figure 1 , Figure 1 A flowchart of a structural health status early warning method based on bridge associated section monitoring data provided by an embodiment of the present invention. The method includes:

[0060] S101, acquiring collected data of various types of sensors used for monitoring the bridge in the target bridge, and calculating the basic probability assignment of the sensor at each loading point as a reference response according to the collected data of each sensor.

[0061] S102, performing a first fusion of the reference responses of the various types of sensors to obtain a first fusion probability assignment matrix of the multi-source sensors.

[0062] S103: Determine the probability function of each loading point according to the first fusion probability assignment matrix.

[0063] S104, using the probability function of each loading point under different monitoring sections as the basic probability assignment to perform a second fusion of the data to obtain a second fusion probability assignment matrix for the target bridge.

[0064] S105, determining a fusion probability function at each loading point according to the second fusion probability assignment matrix, and performing structural damage diagnosis and safety warning according to the fusion probability function.

[0065] The loading point is the position of the vehicle traveling on the target bridge; each row of the first fusion probability assignment matrix represents the benchmark response of the same sensor at different loading points, and each column of the first fusion probability assignment matrix represents the benchmark response of different types of sensors at the same loading point; each row of the second fusion probability assignment matrix represents the probability function at different loading points under the same monitoring section, and each row of the second fusion probability assignment matrix represents the probability function at the same loading point under different monitoring sections.

[0066] A structural health status warning method based on bridge associated section monitoring data is provided based on an embodiment of the present invention. By performing fusion warning calculations on data from multiple sections and multiple sensors, an associated benchmark for associated section data is obtained, thereby greatly reducing the false alarm phenomenon caused by a single section and a single sensor, and effectively improving the accuracy and reliability of bridge warnings.

[0067] In one implementation, a plurality of monitoring sections are provided in the target bridge, and a plurality of types of sensors are deployed in each monitoring section.

[0068] In one implementation, a vehicle traveling on a target bridge can obtain the real-time position of the vehicle as a loading point through a GPS-RTK module.

[0069] In one implementation, the data fusion of multiple sensors of the associated cross section is not limited to two-level data fusion, and this method can be used for combined fusion when there are enough sensors.

[0070] In one embodiment, the multiple types of sensors include static deflection sensors, acceleration amplitude sensors, and strain sensors.

[0071] In one embodiment, the basic probability assignment of each sensor at each loading point is calculated as a reference response based on the collected data of each sensor, including:

[0072] Step 1: Calculate the difference between the collected data of each sensor at each loading point and the preset lossless data as the damage value;

[0073] Step 1: Calculate the basic probability assignment of each loading point as the benchmark response based on the damage value:

[0074]

[0075] In formula (1), x j and x k denote the jth and kth loading points respectively, m(x j ) represents x j The basic probability assignment, ΔG(x j ) and ΔG(x k ) represent x j and x k The damage value at , m is the total number of loading points.

[0076] In one implementation, a data fusion algorithm for multiple sensors of the associated cross section is established. Assuming that L(x) is the structural reference generalized influence line response GQIL (Generalized Quasi-tatic Influence Line), where "generalized" refers to the reference deflection influence line, reference rotation influence line, and reference strain influence line, the first-order difference of the reference influence line can be expressed as formula (2):

[0077] G(x)=L(x)-L(x-Δx) (2)

[0078] It is now determined that G(x) obtained by lossless baseline processing of the initial state of the structure is G(x) BEL , after the structure is damaged, it is G(x) DAMAGE , the GQIL before and after the damage is processed using formula (2) to obtain formula (3).

[0079] ΔG(x)=G(x) BEL -G(x) DAMAGE (3)

[0080] Define the identification frame Θ as each measurement point x j The benchmark response information of (j=1,···,n) is:

[0081] Θ={x1,x2,x3,···,x m-1 ,x m} (4)

[0082] According to the response measurement point x j The amplitude of ΔG(x) at determines the basic probability assignment vector in the identification framework as:

[0083] [m(x1),m(x2),···,m(x j ),···,m(x m )] (5)

[0084] Then in x j The basic probability assignment vector at is defined as the above formula (1). In formula (1), ΔG(x j ) means that the larger the L(x j )The greater the difference, the j The basic probability assignment m(x j ) is larger, and the probability of damage occurring here is also greater.

[0085] Now construct the basic probability assignment matrix under the reference influence line response of multiple sensors as shown in formula (6):

[0086]

[0087] In formula (6), the first row represents the reference response of the first sensor under the 1-mth loading step, and the first column represents the response of the first sensor under the 1-nth loading step. That is, the subscript m represents the number of the sensor, and the subscript x represents the number of the reference loading step.

[0088] In one embodiment, basic probability assignments of multiple types of sensors are first fused to obtain a first fusion probability assignment matrix of multi-source sensors, including:

[0089] For each sensor, the reference response of each loading point is divided by the maximum reference response among the loading points to obtain the normalized and dimensionless reference response;

[0090] The normalized and dimensionless reference responses are fused for the first time to obtain the first fusion probability assignment matrix of the multi-source sensors.

[0091] In one embodiment, determining the probability function of each loading point according to the first fused probability assignment matrix includes:

[0092] The first fusion probability assignment matrix is ​​orthogonally summed to obtain the basic probability function of each loading point;

[0093] According to the position coordinates of the sensor and the position coordinates of the loading point, the distance weights corresponding to the sensor and each loading point are calculated;

[0094] The distance weight of each loading point is used to correct the basic probability function of the loading point to obtain the probability function of the loading point.

[0095] In one implementation, the frequency of each sensor is consistent with the positioning frequency of the vehicle's GPS-RTK module, which can ensure that the collected data is aligned with the loading point in time and space.

[0096] In one implementation, according to the above formulas (1)-(6), the confidence function can be expressed as:

[0097] BEL=[M(x1),M(x2),···,M(x j ),···,M(x m )] (7)

[0098] Among them, based on the orthogonal sum rule, x j The basic probability function at is defined as:

[0099]

[0100] In one embodiment, the first fusion probability assignment matrix is ​​orthogonally summed to obtain the basic probability function of each loading point:

[0101]

[0102] Among them, x j represents the jth loading point, M0(x j ) represents x j The basic probability function, m i (x j ) represents the reference response of the i-th type of sensor at the j-th loading point, n is the number of sensors, and m is the total number of loading points.

[0103] In one embodiment, according to the position coordinates of the sensor and the position coordinates of the loading point, calculating the distance weights corresponding to the sensor and each loading point includes:

[0104] According to the position coordinates of the sensor and the position coordinates of the loading point, the spatial distance between the sensor and the loading point is calculated;

[0105] According to the spatial distance between the sensor and the loading point, the distance weight between the sensor and the loading point is calculated:

[0106]

[0107] Where n is the number of sensors, D ij is the spatial distance between the sensor and the loading point.

[0108] In one implementation, since the precise confidence measurement information of each measuring point is limited, the damage identification index based on the reference influence line is easier to identify local damage close to the measuring point. Therefore, when fusing information from multiple measuring points, it is necessary to consider the difference in sensitivity of the spatial distance between the measuring point and the loading point to local damage, and then introduce a weight coefficient to improve the confidence level of evidence closer to the measuring point.

[0109] In formula (10), the inverse square of the spatial distance is used as the distance weight to adjust the distance D between different measuring points and loading points of the basic probability function. ij The impact on the strength of evidence.

[0110] In one implementation, D is calculated according to formula (11): ij :

[0111]

[0112] The position coordinates of the sensor are (x i ,y i ,z i ), the position coordinates of the loading point are (x j ,y j ,z j ) There are also rules for distance weights as follows:

[0113]

[0114] It is not difficult to know from the above formula that the distance weight of each measuring point is affected differently by the position of the measuring point and the loading point, but the distance weight of the determined measuring point is also a deterministic parameter, which not only weakens the sensitivity of the indicator to the influence of the measuring point distance, but also enhances the error resistance of the measuring point response information.

[0115] In one embodiment, the distance weight of each loading point is used to correct the basic probability function of the loading point, and the probability function of the loading point is obtained as follows:

[0116]

[0117] M0(x j ) represents x j The probability function of .

[0118] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A structural health status early warning method based on bridge associated section monitoring data, characterized in that: The method comprises: Acquire the collected data of various types of sensors used to monitor the bridge in the target bridge, and calculate the basic probability assignment of the sensor at each loading point as the benchmark response based on the collected data of each sensor, including: Calculate the difference between the collected data of each sensor at each loading point and the preset lossless data as the damage value; According to the damage value, the basic probability assignment of each loading point is calculated as the benchmark response: ， Among them, x j and x k denote the jth and kth loading points respectively, m(x j ) represents x j The basic probability assignment, ΔG(x j ) and ΔG(x k ) represent x j and x k The damage value at , m is the total number of loading points, and the loading point is the position of the vehicle traveling on the target bridge; The basic probability assignment matrix under the baseline influence line response of multiple sensors is constructed as follows: ， Where, the first row represents the reference response under the 1st to mth loading steps collected by the 1st sensor, and the first column represents the response collected by the 1st to nth sensors when the 1st reference loading step is applied, that is, the subscript m represents the number of the sensor, and the subscript x represents the number of the reference loading step; The reference responses of various types of sensors are first fused to obtain the first fusion probability assignment matrix of multi-source sensors, including: For each sensor, the reference response of each loading point is divided by the maximum reference response among the loading points to obtain the normalized and dimensionless reference response; The normalized and dimensionless reference responses are first fused to obtain the first fusion probability assignment matrix of the multi-source sensor; Each row of the first fusion probability assignment matrix represents a reference response of the same sensor at different loading points, and each column of the first fusion probability assignment matrix represents a reference response of different types of sensors at the same loading point; Determining the probability function of each loading point according to the first fusion probability assignment matrix includes: The orthogonal sum of the first fusion probability assignment matrix is ​​calculated to obtain the basic probability function of each loading point, which is specifically: ， Among them, x j represents the jth loading point, M0(x j ) represents x j The basic probability function, m i (x j ) represents the reference response of the i-th type of sensor at the j-th loading point, n is the number of sensors, and m is the total number of loading points; According to the position coordinates of the sensor and the position coordinates of the loading point, the distance weights corresponding to the sensor and each loading point are calculated, including: According to the position coordinates of the sensor and the position coordinates of the loading point, the spatial distance between the sensor and the loading point is calculated; According to the spatial distance between the sensor and the loading point, the distance weight between the sensor and the loading point is calculated: ， Where n is the number of sensors, D ij is the spatial distance between the sensor and the loading point; The distance weight of each loading point is used to correct the basic probability function of the loading point to obtain the probability function of the loading point, which is specifically: ， M0(x j ) represents x j The probability function of The probability functions of each loading point under different monitoring sections are used as basic probability assignments to perform a second fusion of the data to obtain a second fusion probability assignment matrix of the target bridge; each row of the second fusion probability assignment matrix represents the probability functions at different loading points under the same monitoring section, and each row of the second fusion probability assignment matrix represents the probability functions at the same loading point under different monitoring sections; The fusion probability function at each loading point is determined according to the second fusion probability assignment matrix, and structural damage diagnosis and safety warning are performed according to the fusion probability function.

2. The structural health status early warning method based on bridge associated section monitoring data according to claim 1 is characterized in that: The various types of sensors include static deflection sensors, acceleration amplitude sensors, and strain sensors.

Images