Bridge vibration monitoring and early warning method and system

Abstract

The invention provides a bridge vibration monitoring and early warning method and system, and relates to the technical field of monitoring and early warning, and the method comprises the steps: obtaining to-be-processed vibration data at a plurality of moments of a monitoring period through a sensor combination disposed at a preset position; acquiring processed vibration data according to the to-be-processed vibration data; determining a vibration source, a vibration mode and a vibration strength coefficient according to the processed vibration data; determining a vibration mode abnormal result according to the vibration mode; acquiring environment data at a plurality of moments of the monitoring period through an environment sensor arranged at a preset position; inputting the vibration mode abnormal result, the vibration strength coefficient and the environment data into a trained fault occurrence time prediction model, and determining predicted fault occurrence time; and generating early warning information according to the vibration source and the predicted fault occurrence time. According to the invention, the accuracy of bridge vibration monitoring and early warning can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring and early warning, and particularly to a method and system for bridge vibration monitoring and early warning. Background Art

[0002] In the related art, bridge vibration monitoring and early warning mainly rely on the method of monitoring by drones. During the process of using drones for monitoring, the cycle inspection interval is relatively large and it is difficult for drones to work in strong wind and heavy rain weather. Therefore, using drones for vibration monitoring may not be able to maintain the timeliness of vibration monitoring and the applicability to the monitoring environment.

[0003] The information disclosed in the background art part of this application is only intended to deepen the understanding of the general background art of this application, and should not be regarded as an admission or any form of suggestion that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0004] The present invention provides a method and system for bridge vibration monitoring and early warning, which can solve the technical problems that the related art cannot maintain the timeliness of vibration monitoring and the applicability to the monitoring environment.

[0005] According to a first aspect of the present invention, there is provided a method for bridge vibration monitoring and early warning, including: At multiple moments in a monitoring cycle, obtain to-be-processed vibration data through a sensor combination arranged at a preset position; Obtain processed vibration data according to the to-be-processed vibration data; Determine the vibration source, vibration mode and vibration intensity coefficient according to the processed vibration data; Determine the vibration mode abnormal result according to the vibration mode; At multiple moments in a monitoring cycle, obtain environmental data through environmental sensors arranged at a preset position, wherein the environmental data includes: temperature data, humidity data and wind force data; Input the vibration mode abnormal result, the vibration intensity coefficient and the environmental data into a trained fault occurrence time prediction model to determine the predicted fault occurrence time; Generate a warning message according to the vibration source and the predicted fault occurrence time.

[0006] According to the present invention, determining the vibration source, vibration mode and vibration intensity coefficient according to the processed vibration data includes: Obtain an original vibration signal and a time-frequency spectrogram according to the processed vibration data; Input the original vibration signal and the time-frequency spectrogram into a vibration mode recognition model to determine the vibration mode; Input the original vibration signal and the time-frequency spectrogram into the vibration source identification model to determine the vibration source; Determine the vibration displacement, vibration acceleration, and vibration frequency according to the original vibration signal; Determine the vibration intensity coefficient according to the vibration displacement, the vibration acceleration, and the vibration frequency.

[0007] According to the present invention, determining the vibration intensity coefficient according to the vibration displacement, the vibration acceleration, and the vibration frequency includes: Obtain the measured damping ratio, design damping ratio, and natural frequency of the bridge; Determine the acceleration limit value and displacement limit value; Determine the vibration intensity coefficient according to the measured damping ratio, the design damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value.

[0008] According to the present invention, determining the vibration intensity coefficient according to the measured damping ratio, the design damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value includes: According to the formula

[0009] Determine the vibration intensity coefficient at the i-th moment of the monitoring period at the k-th preset position , where if is a conditional function, , and are preset weights, is the vibration frequency at the i-th moment of the monitoring period at the k-th preset position, is the natural frequency, is the vibration acceleration at the i-th moment of the monitoring period at the k-th preset position, is the acceleration limit value, is the measured damping ratio at the i-th moment of the monitoring period at the k-th preset position, is the design damping ratio, is the vibration displacement at the i-th moment of the monitoring period at the k-th preset position, is the displacement limit value.

[0010] According to the present invention, determining the vibration mode anomaly result according to the vibration mode includes: Obtain the position categories of each preset position; Determine the normal vibration mode according to the position categories; Determine the vibration mode anomaly result according to the vibration mode and the normal vibration mode.

[0011] According to the present invention, the training steps of the fault occurrence time prediction model include: Obtain the historical processed vibration data and historical environmental data at the historical preset positions of other bridges in the historical time period, wherein the historical environmental data includes: historical temperature data, historical wind force data, and historical humidity data; Determine the historical vibration mode and historical vibration intensity coefficient according to the historical processed vibration data; Determine the historical vibration mode anomaly result according to the historical vibration mode; Input the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data into the fault occurrence time prediction model to determine the sample predicted fault occurrence time; Obtain the historical maintenance records of other bridges; Determine the historical nearest maintenance time according to the historical maintenance records; Determine the training loss function of the fault occurrence time prediction model according to the historical nearest maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data; Train the fault occurrence time prediction model according to the training loss function of the fault occurrence time prediction model to obtain the trained fault occurrence time prediction model.

[0012] According to the present invention, determining the training loss function of the fault occurrence time prediction model according to the historical nearest maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data includes: According to the formula

[0013] Determine the training loss function of the fault occurrence time prediction model , where if is a conditional function, is the historical nearest maintenance time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the sample predicted fault occurrence time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the historical vibration intensity coefficient at the t-th moment at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the preset vibration intensity coefficient threshold, is the historical temperature data at the t-th moment at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the first preset temperature threshold, is the second preset temperature threshold, is the historical humidity data at the r-th historical preset position of the e-th other bridge at the j-th historical time period, is the preset humidity threshold, is the historical wind force data at the t-th moment of the r-th historical preset position of the e-th other bridge at the j-th historical time period, is the preset wind force data threshold, is the vibration mode anomaly result at the t-th moment of the r-th historical preset position of the e-th other bridge at the j-th historical time period, , n is the number of historical time periods, j ≤ n, E is the number of other bridges, e ≤ E, S is the number of moments of the historical time period, t ≤ S, R is the number of historical preset positions, r ≤ R, and n, j, E, e, S, t, r, and R are all positive integers.

[0014] According to the present invention, according to the vibration source and the predicted fault occurrence time, an early warning message is generated, including: Determine a preset first time threshold and a preset second time threshold; According to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position, determine the early warning level at each preset position; Generate an early warning message according to the vibration source, the position category of each preset position, and the early warning level at each preset position.

[0015] According to the present invention, according to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position, determining the early warning level at each preset position includes: When the predicted fault occurrence time at the k-th preset position is less than the preset first time threshold, the early warning level at the k-th preset position is a high-level early warning; When the predicted fault occurrence time at the k-th preset position is greater than or equal to the preset first time threshold and less than or equal to the preset second time threshold, the early warning level at the k-th preset position is a medium-level early warning; When the predicted fault occurrence time at the k-th preset position is greater than the preset second time threshold, the early warning level at the k-th preset position is a low-level early warning.

[0016] According to a second aspect of the present invention, there is provided a bridge vibration monitoring and early warning system, including: A vibration data module, configured to obtain vibration data to be processed at multiple moments of a monitoring period through a sensor combination arranged at a preset position; A data processing module for obtaining processed vibration data according to the vibration data to be processed; A vibration mode module for determining the vibration source, vibration mode and vibration intensity coefficient according to the processed vibration data; A mode anomaly module for determining the vibration mode anomaly result according to the vibration mode; An environmental data module for obtaining environmental data through environmental sensors arranged at preset positions at multiple moments in a monitoring period, where the environmental data includes: temperature data, humidity data and wind force data; A time prediction module for inputting the vibration mode anomaly result, the vibration intensity coefficient and the environmental data into a trained fault occurrence time prediction model to determine the predicted fault occurrence time; An early warning information module for generating early warning information according to the vibration source and the predicted fault occurrence time.

[0017] Technical effects: According to the present invention, vibration data can be accurately collected and processed, and based on the processed vibration data, the vibration source, vibration mode and vibration intensity coefficient can be accurately analyzed. Further, according to the environmental data, vibration mode and vibration intensity coefficient, the fault occurrence time at each preset position of the bridge is predicted, and according to the vibration source and the predicted fault occurrence time, early warning information is generated, improving the accuracy of bridge vibration monitoring and early warning. When determining the vibration intensity coefficient, the vibration intensity coefficient can be determined according to the measured damping ratio, design damping ratio, natural frequency, vibration acceleration, vibration displacement, vibration frequency, acceleration limit and displacement limit. During the calculation process, the influence intensity of vibration can be evaluated from three aspects: resonance risk status, sudden instability risk status and displacement status, improving the comprehensiveness and accuracy of the vibration intensity coefficient. When determining the training loss function, the training loss function of the fault occurrence time prediction model can be determined according to the historical recent maintenance time, sample predicted fault occurrence time, historical vibration mode anomaly result, historical vibration intensity coefficient and historical environmental data. The influence of vibration intensity, vibration mode and environmental data on the sample predicted fault occurrence time can be determined, and based on this influence and the relative error of the sample predicted fault occurrence time, the training loss function is set, so that during the training process of the fault occurrence time prediction model, the training loss function is reduced, and the accuracy of the fault occurrence time prediction model is more targeted improved.

[0018] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit the present invention. According to the following detailed description of exemplary embodiments with reference to the accompanying drawings, other features and aspects of the present invention will be clearer. Description of the Drawings

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other embodiments can be obtained based on these drawings; Figure 1 Exemplarily shown is a schematic flowchart of a bridge vibration monitoring and early warning method according to an embodiment of the present invention; Figure 2 Exemplarily shown is a schematic diagram for determining the vibration source, vibration mode, and vibration intensity coefficient according to an embodiment of the present invention; Figure 3 Exemplarily shown is a schematic diagram for determining the abnormal result of the vibration mode according to an embodiment of the present invention; Figure 4 Exemplarily shown is a schematic diagram for generating an early warning message according to an embodiment of the present invention; Figure 5 Exemplarily shown is a block diagram of a bridge vibration monitoring and early warning system according to an embodiment of the present invention. Detailed implementation manners

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0021] The following will specifically describe the technical solutions of the present invention with specific embodiments. These several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0022] Figure 1 Exemplarily shown is a schematic flowchart of a bridge vibration monitoring and early warning method according to an embodiment of the present invention, and the method includes: Step S1, at multiple moments in the monitoring period, obtain vibration data to be processed through a sensor combination set at a preset position; Step S2, obtain processed vibration data according to the vibration data to be processed; Step S3, determine the vibration source, vibration mode, and vibration intensity coefficient according to the processed vibration data; Step S4, determine the abnormal result of the vibration mode according to the vibration mode; Step S5, acquiring environmental data at multiple moments in the monitoring period through environmental sensors arranged at preset positions, wherein the environmental data includes: temperature data, humidity data and wind speed data; Step S6, inputting the vibration pattern abnormal result, the vibration intensity coefficient and the environmental data into the trained fault occurrence time prediction model to determine the predicted fault occurrence time; Step S7, generating warning information according to the vibration source and the predicted fault occurrence time.

[0023] According to the bridge vibration monitoring and early warning method of an embodiment of the present invention, vibration data can be accurately collected and processed, and the vibration source, vibration mode and vibration intensity coefficient can be accurately analyzed based on the processed vibration data. Furthermore, according to the environmental data, vibration mode and vibration intensity coefficient, the fault occurrence time at each preset position of the bridge can be predicted, and according to the vibration source and the predicted fault occurrence time, early warning information can be generated, thereby improving the accuracy of bridge vibration monitoring and early warning.

[0024] According to an embodiment of the present invention, in step S1, at multiple moments in a monitoring period, vibration data to be processed is acquired through a sensor combination arranged at a preset position.

[0025] For example, a sensor combination is set at preset positions of the bridge (such as stress concentration areas such as piers, mid-spans, supports and cantilever ends). The sensor combination includes: accelerometers, displacement sensors and vibration sensors. The sensor combination is used to collect vibration data to be processed at each preset position.

[0026] According to an embodiment of the present invention, in step S2, processed vibration data is obtained according to the vibration data to be processed.

[0027] For example, the vibration data to be processed can be subjected to denoising (e.g., using wavelet threshold to separate valid signals), data alignment (e.g., achieving microsecond-level time synchronization based on GPS and PTP protocols), and outlier detection (e.g., using isolation forest algorithm to eliminate abnormal data) to obtain processed vibration data.

[0028] According to an embodiment of the present invention, in step S3, the vibration source, vibration mode and vibration intensity coefficient are determined based on the processed vibration data.

[0029] Figure 2 A schematic diagram exemplarily shows the determination of vibration sources, vibration modes and vibration intensity coefficients according to an embodiment of the present invention.

[0030] According to one embodiment of the present invention, step S3 includes: Step S31: Obtain the original vibration signal and the time-frequency spectrogram according to the processed vibration data; Step S32: Input the original vibration signal and the time-frequency spectrogram into the vibration mode recognition model to determine the vibration mode; Step S33: Input the original vibration signal and the time-frequency spectrogram into the vibration source recognition model to determine the vibration source; Step S34: Determine the vibration displacement, vibration acceleration, and vibration frequency according to the original vibration signal; Step S35: Determine the vibration intensity coefficient according to the vibration displacement, the vibration acceleration, and the vibration frequency.

[0031] For example, obtain the STFT time-frequency spectrogram through a time-frequency analysis tool (such as MATLAB), and obtain the original vibration signal according to the processed vibration data. The original vibration signal is a vibration signal that has been normalized and filtered. The vibration signal that has been normalized and filtered plays an important role in the performance of the machine learning model. The vibration mode recognition model belongs to a type of Transformer model in deep learning models. It captures the global features of the vibration signal through the attention mechanism and can be applied to identify complex vibration modes. The vibration mode recognition model is trained with historical data so that it can identify the vibration modes at each preset position, such as vertical bending vibration and lateral bending vibration. The vibration mode is obtained by using the vibration mode recognition model to identify the original vibration signal and the STFT time-frequency spectrogram. The vibration source recognition model belongs to a type of generative adversarial network model in deep learning models. It can use the generator-discriminator framework to identify unknown vibration sources. The vibration source recognition model is trained with historical data so that it can identify the vibration sources of the vibrations at each preset position, such as vehicles and wind power. According to the original vibration signal, obtain the acceleration signal, velocity signal, and displacement signal. After preprocessing the signals, perform physical quantity conversion to determine the vibration displacement, vibration acceleration, and vibration frequency. According to the vibration displacement, vibration acceleration, and vibration frequency, evaluate the influence of the vibration on the preset position and determine the vibration intensity coefficient. The greater the vibration intensity coefficient, the stronger the influence of the vibration on the preset position.

[0032] According to an embodiment of the present invention, step S35 includes: Step S351: Obtain the measured damping ratio, design damping ratio, and natural frequency of the bridge; Step S352: Determine the acceleration limit value and displacement limit value; Step S353: Determine the vibration intensity coefficient according to the measured damping ratio, the design damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value.

[0033] For example, the measured damping ratios at various positions of the bridge are obtained by the half-power bandwidth method, the designed damping ratios are obtained from the bridge design specifications, and the natural frequencies of the bridge are obtained by the active excitation method (applying a controllable excitation through an external device to stimulate the vibration response of the bridge and extracting the peak values based on the vibration response data); according to industry codes and standards, the acceleration limit is determined to be 0.2g, where g is the acceleration due to gravity, and the displacement limit is 1 / 2000 of the span. For example, for a bridge with a span of 100 m, the displacement limit is 5 cm; based on the measured damping ratio, the designed damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit, and the displacement limit, the vibration intensity coefficient at a preset position is determined.

[0034] According to the first embodiment of the present invention, step S353 includes: determining the vibration intensity coefficient at the k-th preset position at the i-th moment in the monitoring period according to formula (1) , (1) where if is a conditional function, , and are preset weights, is the vibration frequency at the k-th preset position at the i-th moment in the monitoring period, is the natural frequency, is the vibration acceleration at the k-th preset position at the i-th moment in the monitoring period, is the acceleration limit, is the measured damping ratio at the k-th preset position at the i-th moment in the monitoring period, is the designed damping ratio, is the vibration displacement at the k-th preset position at the i-th moment in the monitoring period, is the displacement limit.

[0035] According to an embodiment of the present invention, is the relative difference between the vibration frequency and the natural frequency at the k-th preset position at the i-th moment in the monitoring period. The smaller this ratio is, the closer the vibration frequency is to the natural frequency, is the ratio of the measured damping ratio to the designed damping ratio at the k-th preset position at the i-th moment in the monitoring period. The smaller this ratio is, the relatively lower the measured damping ratio is, is the resonance risk status determined according to the vibration frequency status and the measured damping ratio status. The closer the vibration frequency is to the natural frequency and the relatively lower the measured damping ratio is, the larger this value is, and the higher the risk of resonance is.

[0036] According to an embodiment of the present invention, in formula (1), the value of the conditional function includes the following two cases. When When the condition is met, the vibration acceleration is less than the acceleration limit value, indicating that the acceleration is not exceeded, and the value of the condition function is 0. When the condition is not met, it indicates that the acceleration is exceeded, and the value of the condition function is , is the relative difference between the vibration acceleration and the acceleration limit value at the i-th moment of the monitoring period at the k-th preset position. The larger this ratio, the greater the vibration acceleration. is the sudden instability risk status determined according to the acceleration status and the measured damping ratio status. When the measured damping value is low and the acceleration is exceeded, it may cause local buckling of thin-walled members (such as the web of a steel box girder), and there is not enough damping to suppress the spread of instability, so the sudden instability risk is relatively high.

[0037] According to an embodiment of the present invention, is the ratio of the vibration displacement to the displacement limit value at the i-th moment of the monitoring period at the k-th preset position. The larger this ratio, the greater the vibration displacement, the greater the impact on the structural safety, and the greater the vibration intensity.

[0038] According to an embodiment of the present invention, is to determine the vibration intensity based on three aspects: the resonance risk status, the sudden instability risk status, and the displacement status. Among them, , and can be set to 2, 1.5, and 1 respectively.

[0039] In this way, the vibration intensity coefficient can be determined based on the measured damping ratio, the designed damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value. During the calculation process, the impact intensity of the vibration can be evaluated from three aspects: the resonance risk status, the sudden instability risk status, and the displacement status, improving the comprehensiveness and accuracy of the vibration intensity coefficient.

[0040] According to an embodiment of the present invention, in step S4, according to the vibration mode, the vibration mode abnormal result is determined.

[0041] Figure 3 Exemplarily shows a schematic diagram for determining the vibration mode abnormal result according to an embodiment of the present invention.

[0042] According to an embodiment of the present invention, step S4 includes: Step S41, obtaining the position category of each preset position; Step S42, determining the normal vibration mode according to the position category; Step S43, determining the vibration mode abnormal result according to the vibration mode and the normal vibration mode.

[0043] For example, according to the positions of each preset position in the bridge, determine the position category. For example, if the first preset position is at the main girder of the bridge, the position category of the first preset position is "main structure"; if the second preset position is at the pier of the bridge, the position category of the second preset position is "secondary structure"; if the third preset position is at the bridge deck of the bridge, the position category of the third preset position is "local component". The conventional vibration modes corresponding to the main structure include: low-frequency overall bending vibration, low-frequency torsional vibration, and longitudinal telescopic vibration, etc. The conventional vibration modes corresponding to the secondary structure include medium-frequency cable vibration and medium-frequency pier vibration, etc. The conventional vibration modes corresponding to the local components include high-frequency bridge deck vibration, high-frequency vibration of welds / bolts, etc. If the vibration mode at the preset position does not match its corresponding conventional vibration mode, it indicates that the vibration mode is abnormal, and the vibration mode abnormal result is 1; otherwise, the vibration mode abnormal result is 0.

[0044] According to an embodiment of the present invention, in step S5, at multiple moments of the monitoring period, environmental data is acquired through environmental sensors arranged at the preset positions, wherein the environmental data includes: temperature data, humidity data, and wind force data.

[0045] For example, temperature data, humidity data, and wind force data at the preset positions are acquired through a temperature sensor, a humidity sensor, and a wind force sensor arranged at the preset positions.

[0046] According to an embodiment of the present invention, in step S6, the vibration mode abnormal result, the vibration intensity coefficient, and the environmental data are input into a trained fault occurrence time prediction model to determine the predicted fault occurrence time.

[0047] For example, the fault occurrence time prediction model is a deep learning neural network model (such as, long short-term memory network), which is a special recurrent neural network. By introducing a gating mechanism and a cell state, it can effectively capture long-term dependencies in time series and can be processed based on the vibration mode abnormal result, the vibration intensity coefficient, and the environmental data to obtain the predicted fault occurrence time at each preset position.

[0048] According to an embodiment of the present invention, the training steps of the fault occurrence time prediction model include: Acquire historical processed vibration data and historical environmental data of historical preset positions of other bridges in a historical time period, wherein the historical environmental data includes: historical temperature data, historical wind force data, and historical humidity data; Determine the historical vibration mode and the historical vibration intensity coefficient according to the historical processed vibration data; Determine the historical vibration mode abnormal result according to the historical vibration mode; Input the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data into the fault occurrence time prediction model to determine the sample predicted fault occurrence time; Obtain the historical maintenance records of other bridges; Determine the historical nearest maintenance time according to the historical maintenance records; Determine the training loss function of the fault occurrence time prediction model according to the historical nearest maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data; Train the fault occurrence time prediction model according to the training loss function of the fault occurrence time prediction model to obtain the trained fault occurrence time prediction model.

[0049] For example, obtain the processed historical vibration data and historical environmental data of other bridges with similar structures and spans to this bridge in the historical time period; determine the historical vibration mode and historical vibration intensity coefficient according to the historical processed vibration data. The acquisition methods of the historical vibration mode and historical vibration intensity coefficient are similar to those of the vibration mode and vibration intensity coefficient, which will not be elaborated here; determine the historical vibration mode anomaly result corresponding to the historical preset position of other bridges according to the historical vibration mode. The determination method of the historical vibration mode anomaly result is similar to that of the vibration mode anomaly result, which will not be elaborated here; process the historical vibration mode anomaly result, historical vibration intensity coefficient, and historical environmental data according to the fault occurrence time prediction model to determine the sample predicted fault occurrence time. The sample predicted fault occurrence time is the duration from the historical time period to the time when each historical preset position needs to be repaired as predicted; obtain the historical maintenance records of other bridges; determine the historical nearest maintenance time according to the historical maintenance records. For example, the location of the first historical preset position of the first other bridge is the pier of the bridge, and the date of the first historical time period is January 1st. Query according to the historical maintenance records, and the nearest maintenance time for repairing the location of this pier to January 1st is January 11th. The duration from January 11th to January 1st is 10 days. Then the historical nearest maintenance time of the first historical preset position of the first other bridge in the first historical time period is 10 days; determine the training loss function of the fault occurrence time prediction model according to the historical nearest maintenance time, sample predicted fault occurrence time, historical vibration mode anomaly result, historical vibration intensity coefficient, and the historical environmental data; train the fault occurrence time prediction model according to the training loss function to obtain the trained fault occurrence time prediction model.

[0050] According to an embodiment of the present invention, determining a training loss function of a fault occurrence time prediction model based on the historical most recent maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data includes: determining the training loss function of the fault occurrence time prediction model according to formula (2). , (2) wherein, if is a conditional function, is the historical most recent maintenance time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the sample predicted fault occurrence time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the historical vibration intensity coefficient at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period, is a preset vibration intensity coefficient threshold, is the historical temperature data at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period, is a first preset temperature threshold, is a second preset temperature threshold, is the historical humidity data at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is a preset humidity threshold, is the historical wind force data at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period, is a preset wind force data threshold, is the vibration mode anomaly result at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period, , n is the number of historical time periods, j ≤ n, E is the number of other bridges, e ≤ E, S is the number of moments of the historical time period, t ≤ S, R is the number of historical preset positions, r ≤ R, and n, j, E, e, S, t, r, and R are all positive integers.

[0051] According to an embodiment of the present invention, is the ratio of the historical vibration intensity coefficient at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period to the preset vibration intensity coefficient threshold. The larger this ratio is, the greater the historical vibration intensity coefficient at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period. The preset vibration intensity coefficient threshold can be set to , indicating that the vibration frequency is quite different from the natural frequency, the acceleration is not exceeded, and the displacement is small, that is, the vibration intensity is in a normal condition, is the vibration mode anomaly result at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period. When there is an anomaly in the vibration mode, the value of is 1. When the vibration mode is normal, the value of is 0, indicating the vibration danger condition at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period. When the vibration mode is abnormal or the vibration intensity coefficient is larger, this value is larger, and the vibration danger condition is more serious, indicating that the vibration danger condition is negatively correlated with the size of the sample predicted fault occurrence time. For example, when the vibration danger condition is more serious, the possibility of a fault occurring is greater, and the sample predicted fault occurrence time is smaller. Therefore, the terms related to the historical vibration intensity coefficient and the vibration mode anomaly result are placed in the denominator position, indicating the larger the value of, the smaller the sample predicted fault occurrence time.

[0052] According to an embodiment of the present invention, in formula (2), the conditional function the value of includes the following two cases. When the condition of is satisfied, it indicates that the temperature is within the normal range, and the value of the conditional function is 1. When the condition of is not satisfied, it indicates that the temperature is not within the normal range, and the value of the conditional function is , , can be set to -20 degrees Celsius and can be set to 40 degrees Celsius, is the ratio of the absolute value of the historical temperature data at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period to the average value of the first preset temperature threshold and the second preset temperature threshold. The larger this ratio, the greater the deviation degree of the historical temperature data from the normal temperature range, is the ratio of the historical humidity data at the r-th historical preset position of the e-th other bridge in the j-th historical time period to the preset humidity threshold. The larger this ratio, the larger the historical humidity data. The preset humidity threshold can be set to 80%, is the ratio of the historical wind force data at the t-th moment of the r-th historical preset position of the e-th other bridge in the j-th historical time period to the preset wind force data threshold. The larger this ratio, the larger the historical wind force data. The preset wind force data threshold can be set to 20 m / s, It indicates that the degree of deviation of historical temperature data from the normal temperature range, historical humidity data and historical wind data are negatively correlated with the size of the sample predicted fault occurrence time. For example, when the temperature is higher than 40 degrees Celsius, the elastic modulus of steel decreases under high temperature, resulting in an increase in the vibration amplitude, which increases the possibility of fault occurrence and reduces the sample predicted fault occurrence time. When the temperature is lower than 20 degrees Celsius, the brittleness of steel and concrete increases, and the higher the possibility of brittle fracture caused by vibration, the greater the possibility of fault occurrence and reduces the sample predicted fault occurrence time. When the humidity is higher, the corrosion rate of steel metal increases, which increases the possibility of fault occurrence and reduces the sample predicted fault occurrence time. When the wind is stronger, it may cause aerodynamic instability, which increases the possibility of fault occurrence and reduces the sample predicted fault occurrence time. Therefore, the items related to historical temperature data, historical humidity data and historical wind data are placed in the denominator to indicate , and The larger the value of is, the smaller the sample prediction failure time is.

[0053] According to one embodiment of the present invention, The relative error between the predicted fault occurrence time and the most recent historical maintenance time of the sample at the rth historical preset position of the eth other bridge in the jth historical time period is calculated using and The relative errors of the sample predicted fault occurrence time are weighted averaged to obtain the training loss function. During the training process, the above training loss function is reduced, thereby reducing the error between the sample predicted fault occurrence time and the historical most recent maintenance time, and improving the accuracy of the fault occurrence time prediction model in predicting the fault occurrence time, thereby improving the accuracy of the fault occurrence time prediction model.

[0054] In this way, the training loss function of the fault occurrence time prediction model can be determined according to the historical most recent maintenance time, sample predicted fault occurrence time, historical vibration pattern abnormal results, historical vibration intensity coefficient and historical environmental data. The influence of vibration intensity, vibration pattern and environmental data on the sample predicted fault occurrence time can be determined, and the training loss function can be set based on the influence and the relative error of the sample predicted fault occurrence time. During the training process of the fault occurrence time prediction model, the training loss function can be reduced, thereby improving the accuracy of the fault occurrence time prediction model in a more targeted manner.

[0055] According to an embodiment of the present invention, in step S7, warning information is generated according to the vibration source and the predicted fault occurrence time.

[0056] Figure 4 A schematic diagram of generating early warning information according to an embodiment of the present invention is exemplarily shown.

[0057] According to an embodiment of the present invention, step S7 includes: Step S71, determining a preset first time threshold and a preset second time threshold; Step S72, determining the warning level at each preset position according to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position; Step S73, generating a warning message according to the vibration source, the position categories of each preset position, and the warning level at each preset position.

[0058] For example, set the preset first time threshold to 10 days and the preset second time threshold to 100 days; evaluate the warning level at each preset position according to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position; generate a specific warning message according to the vibration source, the position categories of each preset position, and the warning level at each preset position. For example, there is a high-level warning caused by vehicles at the pier of the bridge.

[0059] According to an embodiment of the present invention, step S72 includes: Step S721, when the predicted fault occurrence time at the kth preset position is less than the preset first time threshold, the warning level at the kth preset position is a high-level warning; Step S721, when the predicted fault occurrence time at the kth preset position is greater than or equal to the preset first time threshold and less than or equal to the preset second time threshold, the warning level at the kth preset position is a medium-level warning; Step S723, when the predicted fault occurrence time at the kth preset position is greater than the preset second time threshold, the warning level at the kth preset position is a low-level warning.

[0060] For example, when the predicted fault occurrence time at the kth preset position is less than 10 days, it indicates that there is a serious safety hazard at this position under vibration, and the warning level is a high-level warning; when the predicted fault occurrence time at the kth preset position is greater than 10 days and less than 100 days, it indicates that there is a certain safety hazard at this position under vibration, and the warning level is a medium-level warning; when the predicted fault occurrence time at the kth preset position is greater than 100 days, it indicates that there may be a safety hazard at this position under vibration, and the warning level is a low-level warning.

[0061] The bridge vibration monitoring and early warning method according to an embodiment of the present invention can accurately collect vibration data, process the vibration data, and accurately analyze the vibration source, vibration mode, and vibration intensity coefficient based on the processed vibration data. Further, according to the environmental data, vibration mode, and vibration intensity coefficient, predict the fault occurrence time at each preset position of the bridge, and generate an early warning message based on the vibration source and the predicted fault occurrence time, improving the accuracy of bridge vibration monitoring and early warning. When determining the vibration intensity coefficient, the vibration intensity coefficient can be determined according to the measured damping ratio, design damping ratio, natural frequency, vibration acceleration, vibration displacement, vibration frequency, acceleration limit, and displacement limit. During the calculation process, the influence intensity of the vibration can be evaluated from three aspects: resonance risk status, sudden instability risk status, and displacement status, improving the comprehensiveness and accuracy of the vibration intensity coefficient. When determining the training loss function, the training loss function of the fault occurrence time prediction model can be determined according to the historical recent maintenance time, sample predicted fault occurrence time, historical vibration mode abnormal result, historical vibration intensity coefficient, and historical environmental data. The influence of vibration intensity, vibration mode, and environmental data on the sample predicted fault occurrence time can be determined, and based on this influence and the relative error of the sample predicted fault occurrence time, the training loss function is set, so that during the training process of the fault occurrence time prediction model, the training loss function is reduced, and the accuracy of the fault occurrence time prediction model is more targeted improved.

[0062] Figure 5 Exemplarily shown is a block diagram of a bridge vibration monitoring and early warning system according to an embodiment of the present invention. The system includes: A vibration data module, configured to obtain vibration data to be processed through a sensor combination arranged at a preset position at multiple moments during a monitoring period; A data processing module, configured to obtain processed vibration data according to the vibration data to be processed; A vibration mode module, configured to determine the vibration source, vibration mode, and vibration intensity coefficient according to the processed vibration data; A mode abnormality module, configured to determine a vibration mode abnormal result according to the vibration mode; An environmental data module, configured to obtain environmental data through environmental sensors arranged at preset positions at multiple moments during a monitoring period, where the environmental data includes: temperature data, humidity data, and wind force data; A time prediction module, configured to input the vibration mode abnormal result, the vibration intensity coefficient, and the environmental data into a trained fault occurrence time prediction model to determine the predicted fault occurrence time; An early warning information module, configured to generate an early warning message according to the vibration source and the predicted fault occurrence time.

[0063] The present invention may be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having thereon computer-readable program instructions for performing various aspects of the present invention.

[0064] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been shown and described in the embodiments, and without departing from the said principles, the embodiments of the present invention may have any variations or modifications.

Claims

1. A method for monitoring and warning bridge vibrations, characterized in that, Including: At multiple moments of a monitoring period, obtain to-be-processed vibration data through a sensor combination set at a preset position; Obtain processed vibration data according to the to-be-processed vibration data; Determine the vibration source, vibration mode, and vibration intensity coefficient according to the processed vibration data; Determine the vibration mode anomaly result according to the vibration mode; At multiple moments of a monitoring period, obtain environmental data through an environmental sensor set at a preset position, where the environmental data includes: temperature data, humidity data, and wind force data; Input the vibration mode anomaly result, the vibration intensity coefficient, and the environmental data into a trained fault occurrence time prediction model to determine the predicted fault occurrence time; Generate a warning message according to the vibration source and the predicted fault occurrence time.

2. The bridge vibration monitoring and early warning method according to claim 1, wherein Determining the vibration source, vibration mode, and vibration intensity coefficient according to the processed vibration data includes: Obtain the original vibration signal and the time-frequency spectrum diagram according to the processed vibration data; Input the original vibration signal and the time-frequency spectrum diagram into a vibration mode recognition model to determine the vibration mode; Input the original vibration signal and the time-frequency spectrum diagram into a vibration source recognition model to determine the vibration source; Determine the vibration displacement, vibration acceleration, and vibration frequency according to the original vibration signal; Determine the vibration intensity coefficient according to the vibration displacement, the vibration acceleration, and the vibration frequency.

3. The bridge vibration monitoring and early warning method according to claim 2, characterized in that, Determining the vibration intensity coefficient according to the vibration displacement, the vibration acceleration, and the vibration frequency includes: Obtain the measured damping ratio, design damping ratio, and natural frequency of the bridge; Determine the acceleration limit value and displacement limit value; Determine the vibration intensity coefficient according to the measured damping ratio, the design damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value.

4. The bridge vibration monitoring and early warning method according to claim 3, characterized in that, Determining the vibration intensity coefficient according to the measured damping ratio, the design damping ratio, the natural frequency, the vibration acceleration, the vibration displacement, the vibration frequency, the acceleration limit value, and the displacement limit value includes: According to the formula Determine the vibration intensity coefficient at the k-th preset position at i moments during the monitoring period , where if is a conditional function , and are preset weights is the vibration frequency at the k-th preset position at i moments during the monitoring period is the natural frequency is the vibration acceleration at the k-th preset position at i moments during the monitoring period is the acceleration limit is the measured damping ratio at the k-th preset position at i moments during the monitoring period is the designed damping ratio is the vibration displacement at the k-th preset position at i moments during the monitoring period is the displacement limit 5. The bridge vibration monitoring and early warning method according to claim 1, characterized in that, Determining the vibration mode anomaly result according to the vibration mode includes: Obtain the position categories of each preset position; Determine the conventional vibration mode according to the position categories; Determine the vibration mode anomaly result according to the vibration mode and the conventional vibration mode.

6. The bridge vibration monitoring and early warning method according to claim 1, wherein The training steps of the fault occurrence time prediction model include: Obtain the historical processed vibration data and historical environmental data at the historical preset positions of other bridges during a historical time period, where the historical environmental data includes: historical temperature data, historical wind force data, and historical humidity data; Determine the historical vibration mode and historical vibration intensity coefficient according to the historical processed vibration data; Determine the historical vibration mode anomaly result according to the historical vibration mode; Input the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data into the fault occurrence time prediction model to determine the sample predicted fault occurrence time; Obtain the historical maintenance records of other bridges; Determine the historical most recent maintenance time according to the historical maintenance records; Determine the training loss function of the fault occurrence time prediction model based on the historical most recent maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data; Train the fault occurrence time prediction model according to the training loss function of the fault occurrence time prediction model to obtain the trained fault occurrence time prediction model.

7. The bridge vibration monitoring and early warning method according to claim 6, wherein, Determine the training loss function of the fault occurrence time prediction model based on the historical most recent maintenance time, the sample predicted fault occurrence time, the historical vibration mode anomaly result, the historical vibration intensity coefficient, and the historical environmental data, including: According to the formula Determine the training loss function of the fault occurrence time prediction model , where if is the conditional function, is the historical nearest maintenance time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the sample predicted fault occurrence time at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the historical vibration intensity coefficient at the t-th moment in the j-th historical time period at the r-th historical preset position of the e-th other bridge, is the preset vibration intensity coefficient threshold, is the historical temperature data at the t-th moment in the j-th historical time period at the r-th historical preset position of the e-th other bridge, is the first preset temperature threshold, is the second preset temperature threshold, is the historical humidity data at the r-th historical preset position of the e-th other bridge in the j-th historical time period, is the preset humidity threshold, is the historical wind force data at the t-th moment in the j-th historical time period at the r-th historical preset position of the e-th other bridge, is the preset wind force data threshold, is the vibration mode anomaly result at the t-th moment in the j-th historical time period at the r-th historical preset position of the e-th other bridge, , n is the number of historical time periods, j ≤ n, E is the number of other bridges, e ≤ E, S is the number of moments in the historical time period, t ≤ S, R is the number of historical preset positions, r ≤ R, and n, j, E, e, S, t, r, and R are all positive integers.

8. The bridge vibration monitoring and early warning method according to claim 5, wherein Generate a warning message based on the vibration source and the predicted fault occurrence time, including: Determine a preset first time threshold and a preset second time threshold; Determine the warning level at each preset position according to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position; Generate a warning message based on the vibration source, the position category of each preset position, and the warning level at each preset position.

9. The bridge vibration monitoring and early warning method according to claim 8, wherein Determine the warning level at each preset position according to the predicted fault occurrence time, the preset first time threshold, and the preset second time threshold at each preset position, including: When the predicted fault occurrence time at the k-th preset position is less than the preset first time threshold, the warning level at the k-th preset position is a high-level warning; When the predicted fault occurrence time at the k-th preset position is greater than or equal to the preset first time threshold and less than or equal to the preset second time threshold, the warning level at the k-th preset position is a medium-level warning; When the predicted fault occurrence time at the k-th preset position is greater than the preset second time threshold, the warning level at the k-th preset position is a low-level warning.

10. A bridge vibration monitoring and early warning system, characterized in that, Including: A vibration data module for obtaining vibration data to be processed through a sensor combination set at a preset position at multiple moments in a monitoring period; A data processing module for obtaining processed vibration data according to the vibration data to be processed; A vibration mode module for determining the vibration source, vibration mode, and vibration intensity coefficient according to the processed vibration data; A mode anomaly module for determining the vibration mode anomaly result according to the vibration mode; An environmental data module for obtaining environmental data through environmental sensors set at preset positions at multiple moments in a monitoring period, where the environmental data includes: temperature data, humidity data, and wind force data; A time prediction module for inputting the vibration mode anomaly result, the vibration intensity coefficient, and the environmental data into the trained fault occurrence time prediction model to determine the predicted fault occurrence time; A warning message module for generating a warning message based on the vibration source and the predicted fault occurrence time.

Images