A bridge structure damage rapid diagnosis method and system based on modal frequency identification

Abstract

The application discloses a bridge structure damage rapid diagnosis method and system based on modal frequency identification, and particularly relates to the technical field of bridge structure health monitoring and damage identification, and is used for solving the diagnosis ambiguity and misjudgment risk caused by the signal aliasing of environmental fluctuation, material natural aging and damage of the existing method; the method is realized by collecting bridge vibration response signals and extracting at least two different order measured modal frequencies, decoupling and separating the environmental influence component based on the pre-established collaborative evolution mode, then performing long-term trend analysis on the frequency stripped of the environmental influence to separate the structure performance natural time-varying trend component and the potential damage trend component, analyzing the change characteristic difference between different order potential damage trend components, and finally realizing the identification and evaluation of the existence, possible area and development stage of the bridge structure damage.

Description

Technical Field

[0001] This invention relates to the field of bridge structural health monitoring and damage identification technology, and more specifically, to a method and system for rapid diagnosis of bridge structural damage based on modal frequency identification. Background Technology

[0002] In the field of bridge structural health monitoring, damage identification methods based on vibration modal parameters are widely used due to their advantages of being global, non-contact, and implementable online. Among these methods, damage diagnosis using changes in structural modal frequencies is a typical rapid screening technique. The underlying principle is that structural damage leads to a decrease in local stiffness, which in turn causes a drop in the overall modal frequencies of the structure. Therefore, by continuously monitoring and comparing the differences between the bridge's key modal frequencies and its health baseline, it is possible to quickly determine whether there is a significant deterioration in the overall structural stiffness, thus achieving damage early warning. This method is typically used as a preliminary diagnostic tool to trigger more detailed local inspections.

[0003] However, in practical engineering long-term monitoring applications, the observed values ​​of structural modal frequencies are comprehensive signals influenced by multiple factors. Their changes may not only originate from damage, but are also inevitably affected by the natural evolution of structural material properties over time and fluctuations in environmental conditions. During long-term service, the slow time-varying effect of the overall structural stiffness caused by material aging and creep, along with the cyclical and time-varying effects of environmental temperature and humidity changes on structural boundary conditions and material parameters, will generate signal components in the monitoring data that overlap with the frequency change trends caused by slow accumulation of damage. Existing technologies cannot effectively and reliably separate the frequency change components purely contributed by structural damage from this long-term, homogeneous mixed signal trend, leading to inherent ambiguity and the risk of misjudgment in diagnostic conclusions based on changes in a single frequency index. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, the present invention provides a method and system for rapid diagnosis of bridge structural damage based on modal frequency identification to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A rapid bridge structure damage diagnosis method based on modal frequency identification includes the following steps:

[0007] S1. Collect vibration response signals of the bridge under operating environment excitation;

[0008] S2. Extract at least two measured modal frequencies of different orders of the bridge from the vibration response signal;

[0009] S3. Based on the pre-established cooperative evolution mode of measured modal frequencies of each order under the non-damage state, decouple the measured modal frequencies of at least two different orders, separate the environmental influence component, and obtain the structural modal frequencies of each order after removing the environmental influence.

[0010] S4. Perform long-term trend analysis on the structural modal frequencies of each order after removing the environmental influence. Decompose the long-term trend of the structural modal frequencies of each order after removing the environmental influence into the natural time-varying trend component of structural performance and the potential damage trend component, so as to obtain at least two different potential damage trend components corresponding to each order.

[0011] S5. Analyze the differences in variation characteristics between at least two different orders of potential damage trend components;

[0012] S6. Based on the differences in variation characteristics between at least two different orders of potential damage trend components, identify and assess the structural damage of the bridge.

[0013] Furthermore, S1 includes:

[0014] Multiple acceleration sensors are deployed at key locations in the bridge structure to form a sensor network.

[0015] A unified sampling frequency and time reference are set to synchronously trigger each accelerometer in the sensor network, so as to synchronously collect the multi-channel vibration response signal generated by the bridge structure under the excitation of the operating environment.

[0016] During the acquisition process, the amplitude range of the vibration response signal of each channel is monitored and recorded in real time to ensure that the signal is within the effective range of the sensor and the acquisition equipment.

[0017] Furthermore, S2 includes:

[0018] Preprocessing of multi-channel vibration response signals to suppress noise;

[0019] The preprocessed multi-channel vibration response signal is subjected to frequency domain transformation to obtain the spectral characteristics of the multi-channel vibration response signal;

[0020] Based on the spectral characteristics of the multi-channel vibration response signal, multiple significant peaks corresponding to the overall vibration of the bridge structure are identified in the spectrum, and the frequency value corresponding to each significant peak is determined as a measured modal frequency.

[0021] Based on the known modal order of the bridge structure, at least two measured modal frequencies of different orders are selected from the identified measured modal frequencies.

[0022] Furthermore, S3 includes:

[0023] Under the condition of no damage to the bridge, the co-evolution mode between at least two different orders of measured modal frequencies is determined based on historical monitoring data;

[0024] At least two measured modal frequencies of different orders are projected onto the feature space spanned by the cooperative evolution mode to calculate the cooperative change component caused by the environment as the environmental influence component.

[0025] By subtracting the corresponding environmental influence components from at least two different order measured modal frequencies, the structural modal frequencies of each order after removing the environmental influence are obtained.

[0026] Furthermore, under the condition of no damage to the bridge, the co-evolution mode between at least two different orders of measured modal frequencies is determined based on historical monitoring data. This is achieved by performing principal component analysis on the feature vector composed of at least two different orders of measured modal frequencies in the historical monitoring data and extracting the principal component feature directions.

[0027] Furthermore, S4 includes:

[0028] Long-term time series smoothing is performed on the structural modal frequencies of each order after removing environmental influences to highlight long-term variation trends.

[0029] A natural time-varying trend model of structural performance was established based on the long-term performance evolution law of bridge structural materials.

[0030] The long-term variation trend of the modal frequencies of each order of the structure after smoothing was fitted using the natural time-varying trend model of structural performance, and the natural time-varying trend component of structural performance was separated.

[0031] Subtracting the corresponding natural time-varying trend component of structural performance from the long-term variation trend of the structural modal frequencies of each order after smoothing, the remaining part is identified as the potential damage trend component corresponding to each order, thus obtaining at least two potential damage trend components of different orders.

[0032] Furthermore, S5 includes:

[0033] Calculate the rate of change of at least two different orders of potential damage trend components;

[0034] Construct a correlation coefficient matrix based on the rate of change to describe the synchronicity of changes in the trend components of potential damage at different orders;

[0035] Based on the correlation coefficient matrix, the relative rates of change among the potential damage trend components of each order are quantified.

[0036] By combining the synchronicity of changes with the relative rate of change, a holistic characterization of the differences in change characteristics between at least two different orders of potential damage trend components is formed.

[0037] Furthermore, quantifying the relative rate of change among potential damage trend components of each order based on the correlation coefficient matrix is ​​achieved by setting prior weight coefficients for damage sensitivity for each order, calculating the damage sensitivity ranking index of each order in conjunction with the correlation coefficient matrix, and comparing the magnitudes of the damage sensitivity ranking indices.

[0038] Furthermore, S6 includes:

[0039] Compare the variation characteristics between at least two different orders of potential damage trend components with a preset damage characteristic difference threshold;

[0040] When the difference in change characteristics exceeds the preset threshold for difference in damage characteristics, it is determined that the bridge structure is damaged.

[0041] Based on the distribution pattern of the differences in variation characteristics across different orders, and combined with the sensitivity of the modal vibration mode corresponding to the modal frequency of each order to the location of damage, the structural regions where damage may occur are inferred.

[0042] The stage and severity of the injury are graded based on the duration and rate of change of the characteristics of change.

[0043] On the other hand, the present invention provides a rapid bridge structure damage diagnosis system based on modal frequency identification, comprising the following modules:

[0044] The signal acquisition module is used to acquire vibration response signals of the bridge under operating environment excitation;

[0045] The frequency extraction module is used to extract at least two measured modal frequencies of different orders of the bridge from the vibration response signal.

[0046] The frequency stripping module is used to decouple at least two different orders of measured modal frequencies based on the pre-established cooperative evolution mode of measured modal frequencies of each order under a non-damaging state, separate the environmental influence component, and obtain the structural modal frequencies of each order after removing the environmental influence.

[0047] The component generation module is used to perform long-term trend analysis on the structural modal frequencies of each order after removing the environmental influence. It decomposes the long-term change trend of the structural modal frequencies of each order after removing the environmental influence into the natural time-varying trend component of structural performance and the potential damage trend component, thereby obtaining at least two different potential damage trend components corresponding to each order.

[0048] The difference analysis module is used to analyze the differences in the variation characteristics between at least two different orders of potential damage trend components.

[0049] The damage identification module is used to identify and assess structural damage to bridges based on the differences in variation characteristics between at least two different orders of potential damage trend components.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. By establishing a collaborative evolution model of multi-mode frequencies under healthy conditions, efficient decoupling from environmental effects is achieved. This allows the extraction of modal frequencies that primarily reflect the mechanical properties of the structure from measured data. By decomposing long-term trends into two parts: natural material time-varying and potential damage, the signal aliasing generated by two different physical mechanisms—slow aging of structural performance and sudden damage—is effectively separated. This ensures that the potential damage trend component used for damage assessment minimizes interference from environmental fluctuations and material time-varying, providing a purer and more reliable data foundation for damage identification and significantly improving the accuracy of preliminary screening.

[0052] 2. By analyzing the differences in variation characteristics between at least two different orders of potential damage trend components and inferring the damage location by comprehensively considering the sensitivity of modal modes to the damage location, qualitative identification, preliminary location, and severity assessment of damage are achieved. By utilizing the different sensitivities of different order modal parameters to the damage response, the simple threshold judgment based on a single frequency index is developed into a comprehensive diagnosis based on the differences in multiple frequency change modes. This not only more reliably determines whether damage exists, but also provides directional information on the possible areas of damage and their development stages, thereby enhancing the practicality and reliability of the rapid diagnostic method. Attached Figure Description

[0053] Figure 1 This is a flowchart of a rapid bridge structure damage diagnosis method based on modal frequency identification according to the present invention.

[0054] Figure 2 This is a schematic diagram of the structure of a bridge structure damage rapid diagnosis system based on modal frequency identification according to the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0056] Example 1: Figure 1 This invention presents a rapid bridge structure damage diagnosis method based on modal frequency identification, which includes the following steps:

[0057] S1. Collect vibration response signals of the bridge under operating environment excitation;

[0058] S2. Extract at least two measured modal frequencies of different orders of the bridge from the vibration response signal;

[0059] S3. Based on the pre-established cooperative evolution mode of measured modal frequencies of each order under the non-damage state, decouple the measured modal frequencies of at least two different orders, separate the environmental influence component, and obtain the structural modal frequencies of each order after removing the environmental influence.

[0060] S4. Perform long-term trend analysis on the structural modal frequencies of each order after removing the environmental influence. Decompose the long-term trend of the structural modal frequencies of each order after removing the environmental influence into the natural time-varying trend component of structural performance and the potential damage trend component, so as to obtain at least two different potential damage trend components corresponding to each order.

[0061] S5. Analyze the differences in variation characteristics between at least two different orders of potential damage trend components;

[0062] S6. Based on the differences in variation characteristics between at least two different orders of potential damage trend components, identify and assess the structural damage of the bridge.

[0063] S1. Collect the vibration response signal of the bridge under the excitation of the operating environment. The specific implementation is as follows:

[0064] Multiple accelerometers are deployed at key locations on the bridge structure to form a sensor network. Key locations refer to those where the bridge structure exhibits significant vibration response under operational excitation and reflects the overall dynamic characteristics of the structure, such as the mid-span, quarter-span, top of the piers, and mid-span of the side spans. The sensor network must cover the main vibration modes of the bridge structure. Typically, accelerometers are symmetrically deployed on both the upstream and downstream sides of the main beam. The number of sensors is determined based on the bridge's span and structural form. For example, for a medium-span bridge, at least eight accelerometers should be deployed on the main beam. The selected accelerometers are models suitable for long-term outdoor monitoring. Their measurement range, frequency response range, and sensitivity must match the vibration level of the bridge under normal operational loads. For example, an accelerometer with a range of ±2g and a frequency response range of 0 Hz to 200 Hz should be selected.

[0065] A unified sampling frequency and time reference are set to synchronously trigger all accelerometers in the sensor network to synchronously acquire multi-channel vibration response signals generated by the bridge structure under operational environmental excitation. The unified sampling frequency is set according to the Nyquist sampling theorem and must be more than twice the highest-order modal frequency of the bridge structure to be analyzed. Considering the burden of data storage and transmission, it is usually set to 10 to 20 times the first-order modal frequency of the bridge structure. For example, if the first-order vertical frequency of the bridge is about 2 Hz, the sampling frequency is set to 100 Hz. The time reference is provided by the master clock of the data acquisition equipment. Strict synchronization of all data acquisition channels is achieved through hardware trigger lines or a time synchronization module based on the Global Positioning System to ensure that the time error between the vibration response signals of each channel is less than one sampling interval. The synchronous acquisition process is carried out when the bridge is in normal operating condition, and the acquisition duration must include enough vibration events to obtain stable statistical characteristics. For example, each continuous acquisition time is not less than 30 minutes.

[0066] During the acquisition process, the amplitude range of the vibration response signal of each channel is monitored and recorded in real time to ensure that the signal is within the effective range of the sensor and the acquisition device. Real-time monitoring is achieved by setting a real-time oscilloscope or numerical monitoring interface for the vibration response signal of each channel in the data acquisition software. During the acquisition process, the waveform and peak value of the vibration response signal of each channel are continuously observed. Ensuring that the signal is within the effective range of the sensor and the acquisition device is determined by checking whether the instantaneous peak value of the vibration response signal of each channel exceeds 80% of the sensor's calibrated range. This 80% is an empirical proportional threshold set to prevent signal clipping and to reserve a safety margin. If the peak value of the vibration response signal of any channel continuously exceeds 80%, the gain setting or installation position of the sensor for that channel needs to be adjusted before subsequent acquisition. At the same time, the maximum peak value and minimum peak value of the vibration response signal of each channel during the acquisition period are recorded to form a signal amplitude range record. This record is used as a reference for subsequent data quality assessment and preprocessing steps.

[0067] S2. Extract at least two measured modal frequencies of different orders of the bridge from the vibration response signal, specifically as follows:

[0068] The multi-channel vibration response signals acquired in step S1 are preprocessed to suppress noise. The preprocessing operation includes zero-meaning processing of each channel vibration response signal to eliminate DC offset components, and filtering the signals using a digital bandpass filter with an adjustable cutoff frequency. The lower cutoff frequency of the digital bandpass filter is set to 0.5 times the expected lowest modal frequency of the bridge structure, and the upper cutoff frequency is set to 2 times the expected highest modal frequency of the bridge structure. For example, if the expected first few modal frequencies of the bridge are distributed in the range of 1 Hz to 10 Hz, the lower cutoff frequency of the digital bandpass filter can be set to 0.5 Hz, and the upper cutoff frequency can be set to 20 Hz. The digital bandpass filter can be implemented using, for example, an infinite impulse response filter or a finite impulse response filter. The purpose of the preprocessing is to retain the frequency components related to structural vibration and suppress high-frequency electronic noise and low-frequency environmental interference.

[0069] The preprocessed multi-channel vibration response signals are subjected to frequency domain transformation to obtain their spectral characteristics. This transformation is achieved by performing a Fast Fourier Transform (FFT) on each channel's preprocessed vibration response signal. When performing the FFT, appropriate window functions and data lengths must be set, such as using a Hanning window to reduce spectral leakage. The time-domain data length for each analysis is divided into multiple segments containing, for example, 2048 data points. Each segment is transformed, and the spectral results are then averaged to improve the stability of frequency estimation. The power spectral density or amplitude spectrum of each channel's vibration response signal is calculated using the FFT, serving as the spectral characteristics of that channel. The spectral characteristics of the multi-channel vibration response signal refer to the set of spectral characteristics from all channels.

[0070] Based on the spectral characteristics of the multi-channel vibration response signal, multiple significant peaks corresponding to the overall vibration of the bridge structure are identified in the spectrum, and the frequency value corresponding to each significant peak is determined as a measured modal frequency. The process of identifying significant peaks first calculates the average spectrum of the spectral characteristics of all channels to reduce the influence of random noise and highlight the overall vibration response of the structure. Then, on the average spectrum, local extrema points with amplitudes significantly higher than the amplitudes of the frequency points on both sides are found. A local extrema point is determined to be a significant peak if its amplitude exceeds a pre-set background noise amplitude threshold. This background noise amplitude threshold is determined by adding, for example, three times the standard deviation, to the statistical average amplitude of the average spectrum in the frequency range without significant peaks. For each confirmed significant peak, its specific value on the frequency axis, for example in Hertz, is recorded as a measured modal frequency.

[0071] Based on the known modal order sequence of the bridge structure, at least two measured modal frequencies of different orders are selected from the identified multiple measured modal frequencies. The known modal order sequence of the bridge structure is a pre-determined arrangement of modal frequencies from low to high, obtained through comprehensive analysis of bridge design data, finite element model calculations, or multiple frequency identification results under historical health conditions. For example, the first-order vertical bending mode frequency is known to be the lowest, followed by the second-order vertical bending mode frequency, etc. The selection process sorts all the measured modal frequencies identified in the previous step according to their frequency values ​​from low to high, obtaining the measured modal frequencies from the identified modal frequencies. Measure the frequency sequence; perform mode matching on this measured frequency sequence with the known modal order order, and label each frequency value in the measured frequency sequence as the corresponding specific modal order by comparing the relative magnitude and interval of the frequency values. For example, label the lowest frequency value in the sequence as the first-order frequency and the second lowest frequency value as the second-order frequency; select at least two measured modal frequencies belonging to different modal orders from the completed calibration results. For example, select the calibrated first-order measured modal frequency and the third-order measured modal frequency as at least two measured modal frequencies of different orders required for subsequent steps.

[0072] S3. Based on the pre-established cooperative evolution mode of measured modal frequencies of each order under the non-damaging state, decouple the measured modal frequencies of at least two different orders, separate the environmental influence component, and obtain the structural modal frequencies of each order after removing the environmental influence. The specific implementation is as follows:

[0073] Under the condition of a bridge in a damage-free state, a co-evolutionary pattern between at least two different orders of measured modal frequencies is determined based on historical monitoring data. This pattern is achieved by performing principal component analysis on the feature vector composed of at least two different orders of measured modal frequencies in the historical monitoring data and extracting the principal component feature directions. Specifically, the historical monitoring data refers to multiple sets of measured modal frequency data of at least two different orders continuously collected and extracted through steps S1 and S2 over a sufficiently long period when the bridge is confirmed to be in a healthy and damage-free state. For example, one set of data can be collected every day over the past year. Each set of data constitutes a feature vector, the dimension of which is equal to the number of selected modal orders, and each element in the vector is the modal frequency of the corresponding order in that set of data. Measured modal frequencies; perform principal component analysis on all historical eigenvectors, calculate the covariance matrix of these vectors, and then solve for the eigenvalues ​​and eigenvectors of the covariance matrix; the obtained eigenvectors are the principal component eigendirections, each principal component eigendirection representing a typical pattern of coordinated change of different orders of modal frequencies in historical data; the magnitude of the eigenvalue reflects the amount of data variance explained by the corresponding principal component eigendirection, and usually the top few principal component eigendirections with a cumulative contribution rate of, for example, 80% to 95%, are selected as the basis vectors representing the coordinated evolution pattern; these selected principal component eigendirections are combined into a projection matrix, which characterizes the quantitative pattern of coordinated change of modal frequencies of different orders caused by environmental factors under non-damaging conditions.

[0074] The measured modal frequencies of at least two different orders obtained in step S2 are projected onto the feature space spanned by the co-evolutionary model to calculate the co-evolutionary component caused by the environment as the environmental influence component. Specifically, the measured modal frequencies at the current time are also constructed into a current feature vector with the same dimension as the historical feature vector. This current feature vector is multiplied by the aforementioned projection matrix, or more specifically, the current feature vector is projected onto each selected principal component feature direction, and its coordinate value in that direction is calculated. These coordinate values ​​represent the degree of matching between the current data and the historical co-evolutionary model, that is, the quantitative reflection of the current environmental state in the historical model. Then, the contribution of environmental factors to the current frequencies of each order is inverted using these coordinate values. Specifically, each principal component feature direction vector is multiplied by its corresponding coordinate value and then summed. The result is the co-evolutionary component caused by the environment. This component is a vector with the same dimension as the current feature vector, called the environmental influence component vector. Each element in the environmental influence component vector is the part of the frequency change caused by the environment in the measured modal frequencies of the corresponding order.

[0075] From the at least two different order measured modal frequencies obtained in step S2, the corresponding environmental influence components are subtracted respectively to obtain the structural modal frequencies of each order after removing the environmental influence. This operation is performed in a vector sense, that is, the current feature vector composed of the current measured modal frequencies is subtracted from the calculated environmental influence component vector. The dimension of the new vector obtained after the subtraction remains unchanged, and each element in the new vector is the corresponding order structural modal frequency after removing the environmental influence. This process mathematically decouples the environmental effect from the structural characteristic response. The structural modal frequencies of each order after removing the environmental influence theoretically mainly reflect the mechanical characteristics of the structure itself, such as stiffness, without including the periodic or trend changes caused by environmental factors such as temperature and humidity. These structural modal frequencies of each order after removing the environmental influence will be used as input data for the long-term trend analysis in the subsequent step S4.

[0076] S4. Perform long-term trend analysis on the structural modal frequencies of each order after removing environmental influences. Decompose the long-term variation trend of the structural modal frequencies of each order after removing environmental influences into the natural time-varying trend component of structural performance and the potential damage trend component, thereby obtaining at least two different potential damage trend components corresponding to each order. The specific implementation is as follows:

[0077] The structural modal frequencies of each order after removing environmental influences obtained in step S3 are smoothed using a long-term time series to highlight long-term trends. The smoothing process uses a moving average method, and the window length is set to effectively filter out random fluctuations caused by accidental data collection or short-term load fluctuations, while retaining the trend components reflecting the slow changes in structural performance. The window length is usually related to the sampling time interval of the data and the period of trend change of interest. For example, when the data is a daily value, the window length of the moving average can be set to 30 to 90 days, which helps to smooth out random fluctuations within the month and highlight the quarterly or annual trends. For each order of structural modal frequencies after removing environmental influences, they are arranged into a sequence in chronological order, and the moving average method is applied to calculate a new, smoother sequence. This smoothed sequence is the long-term trend of the modal frequencies of that order.

[0078] A natural time-varying trend model of structural performance is established based on the long-term performance evolution law of bridge structural materials. This model describes the law of slow decay of the overall stiffness of the bridge over time due to the inherent material properties such as concrete shrinkage and creep, and relaxation of steel components, under no-damage conditions. The natural time-varying trend model of structural performance is usually approximated by linear functions or exponential decay functions. For example, a linear model indicates that the stiffness decreases slowly and linearly with the service time, and the square of the corresponding modal frequency also decreases approximately linearly. The specific form of the model and the initial range of parameters are determined based on the long-term performance experimental data of the materials used in the bridge design, the long-term observation experience of similar bridges, or the results of finite element time-varying analysis. For example, for a concrete bridge, its natural time-varying trend model can be initially set as a linear model, that is, the square value of the frequency has a negative linear relationship with time, and the slope parameter in the model can be initially estimated based on the material creep coefficient.

[0079] The long-term variation trend of structural modal frequencies of each order after smoothing is fitted using a natural time-varying trend model of structural performance, thus separating the natural time-varying trend components of structural performance. The fitting process is carried out independently for each order of smoothed long-term variation trend sequence, with time as the independent variable and the smoothed modal frequency value as the dependent variable. The least squares method is used to match the selected natural time-varying trend model of structural performance with the data sequence, and the undetermined parameters in the model are optimized to minimize the overall error between the model calculation value and the actual smoothed sequence value. The optimal model obtained through fitting is the mathematical expression describing the natural time-varying trend component of structural performance of that order. For any given time point, the value obtained by substituting that time point into this optimal model is the natural time-varying trend component of structural performance of that order at that moment.

[0080] Subtracting the corresponding natural time-varying trend component of structural performance from the long-term variation trend of the smoothed structural modal frequencies of each order, the remaining part is identified as the potential damage trend component corresponding to each order, thus obtaining at least two potential damage trend components of different orders. Mathematically, this operation is to subtract two sequences point by point, that is, for each time point, the smoothed modal frequency value at that point is subtracted from the value of the natural time-varying trend component of structural performance calculated by the fitting model at that point. The residual sequence obtained after subtraction represents, in a physical sense, the frequency variation part that cannot be explained by the natural time-varying law of the material. This part of the variation is attributed to possible structural damage and is therefore defined as the potential damage trend component of that order. Through the above process, a corresponding potential damage trend component time series is generated for each analyzed order. These sequences constitute the at least two potential damage trend components of different orders required for the subsequent step S5.

[0081] S5. Analyze the differences in variation characteristics between at least two different orders of potential damage trend components. Specifically, this is implemented as follows:

[0082] The rate of change of at least two different orders of potential damage trend components obtained in step S4 is calculated respectively. The method of calculating the rate of change is to calculate the first difference of the time series of each potential damage trend component and then divide it by the corresponding time interval to obtain a rate of change sequence, that is, the instantaneous rate of change at each time point. The time interval depends on the time resolution of data acquisition. For example, if the data is one value per day, the time interval is 1 day. The purpose of calculating the rate of change is to convert the absolute change of each order of potential damage trend component into a relative rate of change so as to facilitate standardized comparison between different orders.

[0083] Based on the rate of change, a correlation coefficient matrix describing the synchronicity of changes in potential damage trend components of different orders is constructed. The construction process first organizes the rate of change sequence of each order into a column vector, and combines the column vectors of all orders to form a rate of change data matrix. Then, the correlation coefficient matrix of this rate of change data matrix is ​​calculated. The Pearson correlation coefficient formula is used to calculate the covariance between any two rate of change sequences divided by the product of their respective standard deviations. The resulting correlation coefficient matrix is ​​a symmetric matrix, with the number of rows and columns equal to the number of orders analyzed. Each element in the matrix has a value between -1 and +1, representing the degree of linear correlation between the corresponding two rate of change sequences. This correlation coefficient matrix is ​​the quantitative indicator describing the synchronicity of changes in potential damage trend components of different orders.

[0084] Based on the correlation coefficient matrix, the relative rates of change among the potential damage trend components of each order are quantified. This quantification process is achieved by assigning prior weight coefficients to each order regarding damage sensitivity, calculating the damage sensitivity ranking index for each order in conjunction with the correlation coefficient matrix, and comparing the magnitudes of the damage sensitivity ranking indices. The prior weight coefficients are pre-set based on bridge structural dynamics theory and finite element analysis results. They reflect the sensitivity of different order modal frequencies to damage at different locations in the structure. For example, lower order frequencies are more sensitive to changes in overall stiffness, while higher order frequencies may be more sensitive to local damage. Therefore, different priors can be assigned to different orders. Weighting coefficients, for example, the first-order weighting coefficient is 0.5, the second-order weighting coefficient is 0.3, and the third-order weighting coefficient is 0.2. The damage sensitivity ranking index is calculated by summing the absolute values ​​of the elements in each row of the correlation coefficient matrix to obtain the correlation coefficient sum of the corresponding order. Then, this correlation coefficient sum is multiplied by the prior weighting coefficient of that order, and the product is the damage sensitivity ranking index of that order. By comparing the size of the damage sensitivity ranking index of each order, we can judge the relative speed of change of its potential damage trend components. The larger the index, the stronger the indicative power of that order for damage, and the faster its rate of change may be in a relative sense.

[0085] By integrating the relative relationship between synchronicity and rate of change, a holistic characterization of the differences in change characteristics between at least two different orders of potential damage trend components is formed. The integration process is accomplished by combining the synchronicity pattern of change revealed by the correlation coefficient matrix with the relative ranking of change rates revealed by the damage sensitivity ranking index. For example, it can be observed whether orders with faster change rates have high synchronicity, or whether orders with large differences in change rates have low synchronicity. This integrated analysis forms a comprehensive judgment that describes whether the potential damage trend components of different orders are coordinated or separated in their change behavior, as well as the relative intensity of their respective changes, thus constituting a complete description of the differences in change characteristics. This holistic characterization will serve as an important input basis for damage identification and assessment in step S6.

[0086] S6. Based on the differences in variation characteristics between at least two different orders of potential damage trend components, identify and assess the structural damage of the bridge, specifically as follows:

[0087] The variation characteristics difference between at least two different orders of potential damage trend components formed in step S5 is compared with a preset damage characteristic difference threshold. The preset damage characteristic difference threshold is a reference benchmark set based on the upper limit of the normal fluctuation range obtained from long-term monitoring data of a large number of bridge structures under healthy conditions, or by simulating the theoretical change of the difference between potential damage trend components of each order under typical damage conditions using a bridge finite element model. For example, the damage characteristic difference threshold can be set to three times the standard deviation of the long-term average value of the indicator by calculating the standard deviation of the variation characteristic difference index in historical health data. The comparison process is to calculate the quantitative index of variation characteristic difference obtained in the current monitoring period and compare the value of the index with the preset damage characteristic difference threshold.

[0088] When the difference in change characteristics exceeds a preset threshold for damage characteristics, the bridge structure is judged to be damaged. The logic behind this judgment is that the difference in change characteristics exceeding the threshold indicates that at least two different orders of structural dynamic characteristics have undergone uncoordinated changes that exceed the range of normal environmental fluctuations and natural aging of materials. This uncoordinated change is considered a strong indication of structural damage. If the difference in change characteristics does not exceed the threshold, the current bridge structure is judged to be in a normal state, and no damage alarm needs to be issued.

[0089] Based on the distribution pattern of the variation characteristics across different orders, and combined with the sensitivity of the modal modes corresponding to each order's modal frequencies to the damage location, the structural regions where damage may occur are inferred. The analysis of the distribution pattern involves observing which orders the variation characteristics are more significant, and whether the variation characteristics of these orders are positively or negatively correlated. The modal modes corresponding to each order's modal frequencies and their sensitivity to the damage location are obtained in advance through the analysis of the bridge's finite element model. This model describes the maximum displacement or strain energy concentration areas of different order modes. When damage occurs in a specific area, the frequency or trend components of the modes sensitive to strain energy changes in that area will show more significant changes. The inference process involves matching the observed distribution pattern, such as only the high-order potential damage trend components showing significant changes, with the sensitivity characteristic spectrum obtained from the finite element analysis. If the pattern matches the characteristic pattern generated when damage occurs at a certain location simulated in the model, then it is inferred that the damage may occur in the corresponding structural region, for example, it is inferred that the damage may occur in the mid-span region of the main beam or at the bottom of the pier.

[0090] The development stage and severity of damage are assessed by classifying the differences in change characteristics based on their duration and rate of change. Duration refers to the length of time the difference in change characteristics continuously exceeds a preset threshold, and rate of change refers to the average slope of the difference index as it increases over time. The classification criteria are pre-defined; for example, the development stage can be divided into initial, development, and severe stages. The initial stage corresponds to a short duration and low rate of change; the development stage corresponds to a moderate duration and an increasing rate of change; and the severe stage corresponds to a long duration and a high rate of change. Severity assessment can combine the absolute magnitude, duration, and rate of change of the difference in change characteristics for a comprehensive score. For example, a severity level from 1 to 5 can be set, where level 1 represents minor abnormalities requiring attention, and level 5 represents severe damage requiring immediate repair. Through this classification assessment, a qualitative judgment and quantitative rating of the bridge structure's damage status are output, thus completing the transformation from data to diagnostic results.

[0091] Example 2: Figure 2 A schematic diagram of a bridge structure damage rapid diagnosis system based on modal frequency identification is provided. This system includes the following modules:

[0092] The signal acquisition module is used to acquire vibration response signals of the bridge under operating environment excitation;

[0093] The frequency extraction module is used to extract at least two measured modal frequencies of different orders of the bridge from the vibration response signal.

[0094] The frequency stripping module is used to decouple at least two different orders of measured modal frequencies based on the pre-established cooperative evolution mode of measured modal frequencies of each order under a non-damaging state, separate the environmental influence component, and obtain the structural modal frequencies of each order after removing the environmental influence.

[0095] The component generation module is used to perform long-term trend analysis on the structural modal frequencies of each order after removing the environmental influence. It decomposes the long-term change trend of the structural modal frequencies of each order after removing the environmental influence into the natural time-varying trend component of structural performance and the potential damage trend component, thereby obtaining at least two different potential damage trend components corresponding to each order.

[0096] The difference analysis module is used to analyze the differences in the variation characteristics between at least two different orders of potential damage trend components.

[0097] The damage identification module is used to identify and assess structural damage to bridges based on the differences in variation characteristics between at least two different orders of potential damage trend components.

[0098] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0099] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0100] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0101] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0102] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0104] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A bridge structure damage rapid diagnosis method based on modal frequency identification, characterized in that, The method comprises the following steps: S1, collecting vibration response signals of the bridge under operating environment excitation; S2, extracting at least two different order measured modal frequencies of the bridge from the vibration response signals; S3, based on the pre-established cooperative evolution mode of each order measured modal frequency in the undamaged state, decoupling processing is performed on the at least two different order measured modal frequencies to separate the environmental influence component, and each order structural modal frequency after stripping the environmental influence is obtained; Under the undamaged state of the bridge, the cooperative evolution mode between the at least two different order measured modal frequencies is determined based on the historical monitoring data, which is realized by performing principal component analysis on the feature vector composed of the at least two different order measured modal frequencies in the historical monitoring data, and extracting the principal component feature direction; S4, long-term trend analysis is performed on each order structural modal frequency after stripping the environmental influence, the long-term change trend of each order structural modal frequency after stripping the environmental influence is decomposed into a structural performance natural time-varying trend component and a potential damage trend component, thereby obtaining at least two different order potential damage trend components corresponding to each order, comprising: Smooth processing of the long-term time series of each order structural modal frequency after stripping the environmental influence to highlight the long-term change trend; Establishing a structural performance natural time-varying trend model based on the long-term performance evolution law of the bridge structure material; Fitting the long-term change trend of each order structural modal frequency after smoothing using the structural performance natural time-varying trend model to separate the structural performance natural time-varying trend component; Subtracting the corresponding structural performance natural time-varying trend component from the long-term change trend of each order structural modal frequency after smoothing, and the remaining part is identified as the potential damage trend component corresponding to each order, thereby obtaining at least two different order potential damage trend components; S5, analyzing the change characteristic difference between the at least two different order potential damage trend components; S6, based on the change characteristic difference between the at least two different order potential damage trend components, identifying and evaluating the structural damage of the bridge.

2. The method according to claim 1, characterized in that, S1 comprises: Multiple acceleration sensors are arranged at key positions of the bridge structure to form a sensing network; A unified sampling frequency and time reference are set, and each acceleration sensor in the sensing network is triggered synchronously to synchronously collect multi-channel vibration response signals generated by the bridge structure under operating environment excitation; During the collection process, the amplitude range of each channel vibration response signal is monitored and recorded in real time to ensure that the signal is within the effective range of the sensor and the collection device.

3. The method of claim 1, wherein S2 Comprise: Pretreatment is performed on the multi-channel vibration response signals to suppress noise; Frequency domain transformation is performed on the pretreated multi-channel vibration response signals to obtain the frequency spectrum characteristics of the multi-channel vibration response signals; Based on the frequency spectrum characteristics of the multi-channel vibration response signals, multiple significant peaks corresponding to the overall vibration of the bridge structure in the frequency spectrum are identified, and the frequency value corresponding to each significant peak is determined as a measured modal frequency; According to the known modal order sequence of the bridge structure, at least two different order measured modal frequencies are selected from the identified multiple measured modal frequencies.

4. The method of claim 1, wherein S3 Comprise: In the undamaged state of the bridge, the coordinated evolution pattern between the measured modal frequencies of at least two different orders is determined based on historical monitoring data; The measured modal frequencies of at least two different orders are projected into the characteristic space formed by the coordinated evolution pattern to calculate the coordinated change component caused by the environment as the environmental influence component; The environmental influence component is subtracted from the measured modal frequencies of at least two different orders respectively to obtain the structural modal frequencies of each order after stripping the environmental influence.

5. The method of claim 1, wherein S5 It includes: The change rates of the potential damage trend components of at least two different orders are calculated respectively; A correlation coefficient matrix describing the change synchronization of the potential damage trend components of different orders is constructed based on the change rates; The relative speed relationship between the change rates of the potential damage trend components of different orders is quantified according to the correlation coefficient matrix; The overall characterization of the change characteristic difference between the potential damage trend components of at least two different orders is formed by comprehensively considering the change synchronization and the relative speed relationship.

6. The method of claim 5, wherein the method further comprises: The relative speed relationship between the change rates of the potential damage trend components of different orders is quantified according to the correlation coefficient matrix by setting prior weight coefficients for the damage sensitivity of each order, calculating the damage sensitivity ranking index of each order based on the correlation coefficient matrix, and comparing the damage sensitivity ranking index.

7. The method of claim 1, wherein S6 It includes: The change characteristic difference between the potential damage trend components of at least two different orders is compared with the preset damage characteristic difference threshold; When the change characteristic difference exceeds the preset damage characteristic difference threshold, it is determined that the bridge structure is damaged; The structure region where damage may occur is inferred according to the distribution pattern of the change characteristic difference among different orders and the sensitive characteristics of the modal shapes corresponding to the modal frequencies of each order to damage location; The development stage and severity of the damage are classified and evaluated according to the duration and change rate of the change characteristic difference.

8. A bridge structure damage rapid diagnosis system based on modal frequency identification, used to implement the bridge structure damage rapid diagnosis method based on modal frequency identification in any one of claims 1-7, characterized in that, It includes the following modules: A signal acquisition module for acquiring vibration response signals of the bridge under operational environmental excitation; A frequency extraction module for extracting at least two different orders of measured modal frequencies of the bridge from the vibration response signals; A frequency stripping module for decoupling at least two different orders of measured modal frequencies based on the coordinated evolution pattern of each order of measured modal frequencies in the undamaged state to separate the environmental influence component and obtain the structural modal frequencies of each order after stripping the environmental influence; A component generation module for long-term trend analysis of the structural modal frequencies of each order after stripping the environmental influence, decomposing the long-term change trend of the structural modal frequencies of each order after stripping the environmental influence into a structural performance natural time-varying trend component and a potential damage trend component, thereby obtaining at least two different orders of potential damage trend components corresponding to each order; A difference analysis module for analyzing the change characteristic difference between the potential damage trend components of at least two different orders; A damage identification module for identifying and evaluating the structural damage of the bridge based on the change characteristic difference between the potential damage trend components of at least two different orders.

Images