Bridge structure monitoring method and device based on microwave deformation radar

Abstract

The invention provides a bridge structure monitoring method and device based on a microwave deformation radar, and relates to the technical field of bridge structure monitoring, and the method comprises the steps: obtaining the multi-point three-dimensional displacement data of a bridge structure through the microwave deformation radar, carrying out the thermal expansion pseudo displacement compensation and multi-point space smoothing through combining with temperature information, and obtaining a displacement field after environment correction; secondly, extracting a vertical component and separating the vertical component into a static deformation component and a dynamic vibration component by adopting variational mode decomposition; further analyzing and identifying a decoupling region through a time window coherence coefficient, performing recursive quantitative analysis, bispectrum analysis and energy distribution entropy calculation on a dynamic signal of the region, and constructing a high-order damage sensitive feature set; and finally, the dynamic characteristics and the static curvature change are fused to form a comprehensive degradation degree index, the deviation degree is judged according to working condition classification and the mahalanobis distance, and multi-dimensional and cross-working-condition degradation identification and risk early warning of the bridge structure are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of bridge structure monitoring, in particular to a bridge structure monitoring method and device based on microwave deformation radar. BACKGROUND

[0002] As an important infrastructure, the structural health of a bridge is directly related to public safety and traffic efficiency. However, many current bridge monitoring systems mainly rely on periodic physical inspections and simple strain measurements, which have a lag and inaccuracy in detecting potential structural damage. Traditional technologies often cannot assess the health status of a bridge in real time and accurately, especially in dynamic characteristic changes that cannot be captured by simple measurement methods. For example, existing systems cannot effectively combine dynamic parameters such as frequency, amplitude, and phase for comprehensive evaluation, resulting in potential damage often being discovered only after it has developed to a serious stage. This technical limitation increases the cost and risk of bridge maintenance.

[0003] In existing bridge structure health monitoring technologies, a large number of point sensors such as strain gauges, accelerometers, or optical fiber sensors are mainly relied on to achieve the acquisition of local structural responses. However, such methods have low spatial resolution, complex installation, high operation and maintenance costs, and are difficult to achieve high-frequency, non-contact full-field monitoring of the entire bridge structure. In addition, traditional monitoring methods based on a single data source often cannot simultaneously capture static deformation and dynamic vibration characteristics, have limited compensation ability for environmental disturbances such as wind load and temperature effects, are prone to misjudgment or omission, and are difficult to achieve sensitive identification of early structural damage and local degradation.

[0004] Although certain radar or optical technologies have been tried in bridge monitoring in recent years, such as synthetic aperture radar (SAR) or laser scanning methods, these technologies mainly focus on single displacement monitoring and cannot simultaneously acquire multi-modal field information. They also lack dynamic decoupling of deformation fields, systematic extraction and fusion evaluation mechanisms for vibration damage characteristics, especially in complex working conditions. Existing systems cannot construct a classification model of structural responses combined with operating environment information, and lack risk quantification methods based on statistical distribution.

[0005] The above information disclosed in the background section is only intended to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The present application aims to provide a bridge structure monitoring method and device based on microwave deformation radar to solve the problems raised in the background.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical solutions: A bridge structure monitoring method based on microwave deformation radar, the specific steps comprising: Step 1: using microwave deformation radar to perform high-frequency scanning on the bridge surface, generating full-field three-dimensional displacement and its change rate through phase interference, and simultaneously inversely deriving the structure surface thermal field distribution and bridge deck wind speed information from the radar echo signal, performing thermal strain compensation and spatial continuity filtering on the displacement field, and outputting the environment-corrected three-dimensional displacement field data, structure surface temperature field and wind speed field; Step 2: extracting the vertical component from the corrected three-dimensional displacement field, and performing variational modal decomposition thereon to separate the static deformation field and the dynamic vibration field, calculating the space-time coherence coefficient matrix based on the dynamic vibration field, defining the area where the coherence drops by more than 30% as the decoupling area, extracting the Hilbert instantaneous amplitude variance and instantaneous frequency variance, the bispectrum maximum peak value, the recursive quantization index and the vibration energy distribution entropy of the decoupling area, and outputting as the damage sensitive feature set; Step 3: normalizing each component of the damage sensitive feature set according to the historical baseline data of the bridge health state, determining the feature weight by entropy weight method and fusing to generate the dynamic damage index, calculating the curvature change of the whole bridge structure based on the static deformation field, and fusing the dynamic damage index to form the current structure comprehensive degradation degree; Step 4: constructing a historical working condition database within the bridge operation monitoring period, which includes light load and heavy load working condition categories distinguished according to the vibration energy distribution entropy, and weak wind and strong wind working condition categories divided according to the wind speed; for each working condition, the 90% confidence interval of the structure comprehensive degradation degree is calculated, and the corresponding working condition category is matched according to the current load and wind speed state, the Mahalanobis distance of the current comprehensive degradation degree relative to the historical distribution of the working condition is calculated, which is used to realize the structure risk identification.

[0008] Further, the logic of using microwave deformation radar to perform high-frequency scanning on the bridge surface to generate full-field three-dimensional displacement and its change rate is as follows: Determine the key monitoring area of the bridge to be monitored, which includes piers, bridge deck center, abutments, bridge connection areas and bearings, then divide each key monitoring area into several grid units on average, index these grid units with i, and arrange key monitoring points at the center of each grid unit, then the key monitoring point at the center of the i-th grid is the i-th key monitoring point; The microwave deformation radar performs high-frequency phase interference scanning on the key monitoring points of the bridge deck with a continuous wave signal with a frequency of , wherein the wavelength of the radar signal is , and the phase difference between adjacent two echoes is , wherein is the speed of light, According to the principle of phase interference, the radial displacement of the point along the radar line of sight is obtained, and the formula is: ; in, is the i-th key monitoring point at time The phase difference, is the i-th key monitoring point at time Radial displacement; The elevation angle of the i-th key monitoring point relative to the radar is known and azimuth , the radial displacement Projected into the local coordinate system of the bridge deck, the three-axis displacement change components are obtained: ; in, 、 and are the components of the instantaneous displacement change of the bridge in the east-west, north-south and vertical directions respectively; For each key monitoring point i, the three-axis displacement increment is divided by the time interval to obtain the three-axis instantaneous displacement change rate; in the same radar echo, the reflected signal intensity is and spectrum broadening Analysis of the surface temperature field of the structure and bridge deck wind speed field , the structure temperature Using pre-calibrated constants and , the observed Mapped to corresponding temperature : , the bridge deck wind speed By formula Get; Considering the pseudo displacement of bridge material caused by temperature change, compensation is performed for each key monitoring point i as follows: Set is the linear expansion coefficient of the bridge material corresponding to the i-th key monitoring point, is the equivalent observation length of the i-th key monitoring point. At this key monitoring point, the structural expansion and contraction caused by temperature is considered to be expanded as a whole along this length. is the reference temperature, then the point at time The vertical thermal expansion pseudo displacement caused by temperature change is , this value is divided from the measured vertical displacement Deduct from the vertical displacement to get the vertical displacement after thermal compensation. The compensation methods for the east-west and north-south directions are the same. On the predefined grid of the bridge deck, for the grid cell where the i-th key monitoring point is located, its neighborhood cell set is recorded as , calculate the i-th key monitoring point and its neighborhood unit set The spatial distance of the jth key monitoring point , j is the neighborhood unit set The index of the key monitoring point in the table and the definition of Gaussian weights , the thermally compensated displacement of each key monitoring point is smoothed in the three-axis direction in a weighted average manner, and the final output of the environmental correction of each key monitoring point i at time The three-dimensional displacement field data is output, and the inverted structural temperature and wind speed are output simultaneously.

[0009] Furthermore, the vertical component is extracted from the corrected three-dimensional displacement field, and variational modal decomposition is performed on it to separate the static deformation field and the dynamic vibration field. The logic is as follows: The vertical component of the i-th key monitoring point is decomposed by variational mode decomposition and classified into low-frequency static deformation components according to the size of the center frequency. and high-frequency dynamic vibration components , is the i-th key monitoring point at time The static deformation of is the i-th key monitoring point at time Dynamic vibration; Set time window , extract its vertical dynamic vibration component and calculate it in the time window signal sequence within and , calculate the coherence coefficient matrix to obtain its time window coherence coefficient matrix : ; in, yes arrive The time variable defines the average coherence coefficient of the i-th key monitoring point ,when Compared with the previous moment A decrease of more than 30%, that is When , the grid unit where the i-th key monitoring point is located is identified as the decoupling zone, and the decoupling zone is indexed by p; For the marked decoupling region p, its dynamic vibration component Perform Hilbert transform to extract instantaneous amplitude and instantaneous frequency. The specific steps are as follows: make Express The Hilbert transform result is used to construct the analytical signal: ; in, is the imaginary unit, the modulus of the analytical signal , which is the pth decoupling zone at time The instantaneous amplitude and its phase , the phase derivative is obtained That is the instantaneous frequency of the decoupling zone; Calculate the variance of the instantaneous amplitude and instantaneous frequency within the window and ,when or If the value exceeds 3 standard deviations from the baseline value of the healthy state, the point is considered to have abnormal fluctuations, indicating signs of structural damage.

[0010] Furthermore, for the dynamic vibration signal of the decoupling zone p Implement delay embedding reconstruction, construct its phase space trajectory, and extract the following two recursive quantitative analysis indicators: Certainty Index and stratification , the certainty index indicates the proportion of linear repetition segments in the phase space trajectory, and the stratification degree indicates the proportion of time the phase space trajectory stays in the same state; For the marked decoupling area p, smooth the vertical dynamic vibration signal Perform second-order bispectral analysis to detect high-order nonlinear coupling characteristics. The formula for defining the second-order bispectral function is as follows: ; in, For signal The Fourier transform of represents the complex conjugate, Represents the statistical average of multiple time windows; If there is a frequency pair Make It is significantly higher than the corresponding frequency of the decoupling zone in a healthy state. If the historical baseline value of , it means that the signal has significant non-Gaussian coupling at these two frequencies, and the maximum peak is recorded as: ; And use it as a high-order nonlinear damage indicator; For each marked decoupling zone p, take its vertical dynamic vibration signal Perform bandpass filtering within 5-20Hz to obtain a bandpass signal, divide the bandpass signal into N equal segments within the time window, and calculate the The total energy of the segment is obtained, and then the energy proportion is obtained, and then the energy distribution entropy is defined: wherein is the energy proportion of the segment, measures the dispersion degree of the energy of the region in the selected frequency band on each sub-segment; Finally, the above data extracted from the decoupling area p is composed of the damage sensitive feature set.

[0011] Further, the logic of normalizing each component of the damage sensitive feature set according to the historical baseline data of the bridge health state is as follows: The damage sensitive feature set extracted from the decoupling area p , wherein the oth feature component is subtracted from the mean value of the bridge health state, and divided by the corresponding standard deviation to obtain the dimensionless standard component , o is the index of the feature component in the damage sensitive feature set, and m is the total number of feature components in the damage sensitive feature set; By calculating the information entropy of each feature in the historical sample, the weight of each feature is determined according to the inverse proportion principle of information entropy , the greater the weight, the stronger the feature distinguishes the state difference, and finally the dynamic damage index of the decoupling area p is represented as the weighted sum of each standard feature: ; wherein, is regarded as a comprehensive quantitative value reflecting the dynamic damage signals such as nonlinear characteristics, spectral anomalies, and coupling structure variations of the current region; The static part of the bridge deck deformation field is extracted , and the curvature change from the initial state is calculated: ; wherein, is the time when the bridge deck is built; After normalizing the dynamic damage index and the curvature change , the degradation degree index is obtained by using a linear fusion model: ; wherein, is the normalized dynamic damage index and the curvature change, is a weight coefficient, which is set according to the structural characteristics of the bridge.

[0012] ​Further, a historical working condition database is constructed in a bridge operation monitoring period, the database including light load and heavy load working condition categories distinguished according to vibration energy distribution entropy, and weak wind and strong wind working condition categories divided according to wind speed: Based on external interference and load changes suffered by the bridge structure in the operation process, the working condition of each monitoring time period is divided into the following four categories: if the vibration energy distribution entropy , then the working condition is determined as a light load working condition, otherwise as a heavy load working condition; if the wind speed , then the working condition is determined as a weak wind working condition, otherwise as a strong wind working condition, wherein , are all set threshold values; the two are combined to divide into four typical working condition categories: light load weak wind, light load strong wind, heavy load weak wind, and heavy load strong wind, and is used to index the categories; For each working condition category , the corresponding structural comprehensive degradation degree in the historical period is recorded, the 90% confidence interval and the mean value and the covariance matrix are calculated; For any sampling time t, if the corresponding current working condition category is , then the comprehensive degradation degree is recorded as , and the Mahalanobis distance relative to the historical distribution of the working condition is defined as: ; Wherein, is the Mahalanobis distance, which represents the deviation of the current structure state relative to the same working condition, is the transpose of the matrix; if exceeds the 95% confidence upper limit of the maximum value, it is considered as a potential structure risk event, triggering an alarm.

[0013] The application further provides a bridge structure monitoring device based on microwave deformation radar, which is used to execute the above-mentioned bridge structure monitoring method based on microwave deformation radar, and comprises: A data collection module is used to perform high-frequency scanning on the bridge surface using the microwave deformation radar, generate a full-field three-dimensional displacement and its change rate through phase interference, and simultaneously obtain the structure surface thermal field distribution and the bridge surface wind speed information from the radar echo signal, and perform thermal strain compensation and spatial continuity filtering on the displacement field, and output the environment-corrected three-dimensional displacement field data, the structure surface temperature field and the wind speed field; The data processing module is used for extracting a vertical component from the corrected three-dimensional displacement field, and performing variational mode decomposition on the vertical component to separate a static deformation field and a dynamic vibration field, calculating a space-time coherence coefficient matrix based on the dynamic vibration field, defining a region with a coherence drop of more than 30% as a decoupling region, extracting Hilbert instantaneous amplitude variance and instantaneous frequency variance, a bispectrum maximum peak value, a recursive quantization index and vibration energy distribution entropy from the decoupling region, and outputting the damage sensitive feature set; The comprehensive calculation module is used for normalizing each component of the damage sensitive feature set according to historical baseline data of the bridge health state, determining feature weights by using an entropy weight method and fusing to generate a dynamic damage index, calculating a curvature change amount of the whole bridge structure based on the static deformation field, and fusing the dynamic damage index to form a structure comprehensive degradation degree at a current time. The risk identification module is used for constructing a historical working condition database in a bridge operation monitoring period, the database including light load and heavy load working condition categories distinguished according to vibration energy distribution entropy, and weak wind and strong wind working condition categories divided according to wind speed; 90% confidence intervals of the structure comprehensive degradation degree are calculated for each working condition, and the current load and wind speed state are matched with the corresponding working condition category, the Mahalanobis distance of the current comprehensive degradation degree relative to the historical distribution of the working condition is calculated, and the structure risk identification is realized.

[0014] Compared with the prior art, the present application has the following advantages: The present application significantly improves the performance of the prior art in the aspects of structure state perception dimension, abnormality recognition sensitivity and working condition disturbance suppression capability by constructing a bridge structure health evaluation method that fuses microwave deformation radar monitoring, structure dynamic response feature extraction and typical working condition classification modeling.

[0015] The present application solves the problems of traditional single feature vibration indexes being easily disturbed by noise and having insufficient state discrimination by using multi-scale decomposition and entropy weight modeling on radar inversion displacement time series data, constructing a dynamic damage sensitive feature set, and realizing weighted fusion of damage indexes by using inverse weight information entropy. In order to solve the problem of high misjudgment risk caused by the great influence of wind load, traffic load and other external disturbances on structure response, the present application proposes a typical working condition recognition and Mahalanobis distance deviation amount analysis mechanism, which first divides the structure response working condition into four types of light load-heavy load and weak wind-strong wind combinations based on vibration energy entropy and wind speed characteristics, and then constructs a historical degradation degree distribution model according to the working condition.

[0016] The application realizes consistent state evaluation across working conditions by calculating Mahalanobis distance of the current structure state relative to the mean value in the working condition, which has the advantages of dimension independence and variance sensitivity, can effectively distinguish response abnormalities caused by environmental disturbance and intrinsic degradation of the structure, significantly reduces the misjudgment rate, and improves the reliability of early damage identification, and the application solves the core problems of single information dimension and lack of quantitative risk evaluation mechanism in existing structure health monitoring methods by constructing a multi-modal index system and a statistical driving risk criterion. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is a whole method flowchart of the application; Figure 2 It is a whole device structure schematic diagram of the application; Figure 3 It is a contrast diagram of original vertical displacement and vertical displacement after temperature compensation of the application; Figure 4 It is a structure degradation degree-dynamic loss index curve diagram of the application; Figure 5 It is a static curvature change-structure degradation degree curve diagram of the application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the application clearer and more apparent, the application will be further described in detail below in combination with specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the application should be understood as the usual meaning understood by those skilled in the art to which the application belongs. The "first", "second" and similar words used in the application do not represent any order, quantity or importance, but are only used to distinguish different components. "Include" or "contain" and similar words mean that the elements or objects before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connected" or "connected" and similar words are not limited to physical or mechanical connection, but can include electrical connection, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent relative positional relationship, when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] EMBODIMENT: Please refer to Figures 1-5 The application provides a technical scheme: A bridge structure monitoring method based on microwave deformation radar, the specific steps comprising: Step 1: High-frequency scanning of the bridge surface using microwave deformation radar, generating full-field three-dimensional displacement and its rate of change through phase interference, and simultaneously inverting the structural surface thermal field distribution and bridge deck wind speed information from the radar echo signal, compensating for thermal strain and spatial continuity filtering of the displacement field, and outputting the environmental corrected three-dimensional displacement field data, structural surface temperature field, and wind speed field; Key locations on the bridge surface to be monitored include bridge piers, which bear and distribute the vertical and horizontal loads of the bridge, and the displacement detection of the bridge piers can help identify changes in structural stress due to foundation settlement, load changes, or other factors; The bridge deck center, which is the most concentrated place of stress on the bridge, especially for suspension bridges and cable-stayed bridges, monitoring the displacement of the bridge deck center can reveal the impact of traffic loads and environmental factors (such as wind loads) on the bridge; Abutments, which connect the bridge to the bank slope and bear the horizontal load of the bridge deck, changes in the displacement of the abutments can indicate changes in the structural stress of the bridge due to temperature changes or uneven foundation settlement; Bridge connection points, which allow the bridge to freely stretch and contract with temperature changes, monitoring these points can help identify structural deformation or wear at the connection points due to thermal expansion and contraction; Bearings, which bear the weight and deformation of the bridge body, monitoring the displacement of the bearings can reveal changes in the function of the bearings due to load changes or material fatigue; The logic for using microwave deformation radar to perform high-frequency scanning of the bridge surface to generate full-field three-dimensional displacement and its rate of change is as follows: Determine the key locations on the bridge surface to be monitored, divide them into several key monitoring areas, and then divide each key monitoring area into several grid cells, index these grid cells with i, and place key monitoring points at the center of each grid cell, then the key monitoring point at the center of the ith grid is the ith key monitoring point; The microwave deformation radar uses a continuous wave signal with a frequency of to perform high-frequency phase interference scanning of the key monitoring points on the bridge deck, where the radar signal wavelength , and is the speed of light, is the sampling time, the radar continuously scans the same point, and records the phase difference between adjacent two echoes , according to the principle of phase interference, the radial displacement of the point along the radar line-of-sight direction is obtained, the formula is: ; where, is the phase difference of the ith key monitoring point at time , and is the radial displacement of the ith key monitoring point at time ; Reflects the distance change between the target and the radar. The larger the distance change, the more severe the structural vibration or deformation. Reflects the instantaneous displacement of the structure along the radar direction. The larger the displacement, the greater the structural deformation amplitude. The formula is essentially a phase-displacement conversion, and the coefficient Determined by the principle of electromagnetic wave interference, the phase difference increases , corresponding to the displacement change ; The elevation angle of the i-th key monitoring point relative to the radar is known and azimuth , the radial displacement Projected into the local coordinate system of the bridge deck, the three-axis displacement change components are obtained: ; in, 、 and are the components of the instantaneous displacement change of the bridge in the east-west, north-south and vertical directions respectively; Pitch angle is the angle between the radar line of sight and the horizontal plane, azimuth is the angle between the projection of the radar line of sight on the horizontal plane and due north, Reflects the degree of tilt of the radar line of sight. The larger the value, the greater the difference in point height and determines the weight of the vertical component; azimuth Reflects the horizontal direction of the radar line of sight. The larger the value, the greater the difference in the point's azimuth and determines the direction of the horizontal component. Reflects the vertical deformation of the bridge. The larger the value, the lower the vertical stiffness of the bridge or the increase in load. It is proportional to ; Reflects the east-west horizontal displacement of the bridge. A larger value indicates lateral structural instability or eccentric load. The radial displacement Decomposed into the local coordinate system of the bridge, the pitch angle Control the vertical component ratio ( , all are horizontal displacements), azimuth Assign horizontal displacement directions; For each key monitoring point i, the three-axis displacement increment is divided by the time interval to obtain the three-axis instantaneous displacement change rate; in the same radar echo, the reflected signal intensity is and spectrum broadening Analysis of the surface temperature field of the structure and bridge deck wind speed field , the structure temperature Using pre-calibrated constants and , the observed Mapped to corresponding temperature : , the bridge deck wind speed is obtained by formula ; reflects the electromagnetic wave reflection ability of the material surface, and the greater the value, the greater the change in the dielectric constant of the material or the change in the surface state, reflects the real-time temperature of the structure surface, and the greater the value, the higher the ambient temperature or the stronger the solar radiation; the dielectric constant of the material changes with temperature, affecting the radar echo intensity, and the coefficient and need to be calibrated for different materials (steel / concrete); reflects the degree of dispersion of the echo signal frequency, and the greater the value, the more intense the wind-induced turbulence or the greater the wind speed, and reflects the real-time wind speed of the bridge deck, and the greater the value, the greater the wind load; moving particles, such as wind-borne water droplets, cause the echo frequency to broaden, and the broadening amount is proportional to the wind speed , and the coefficient is determined by the radar wavelength; Considering the thermal expansion pseudo-displacement of the bridge material caused by temperature changes, the compensation for each key monitoring point i is as follows: set as the linear expansion coefficient of the bridge material corresponding to the i-th key monitoring point, as the equivalent observation length of the i-th key monitoring point, at which the temperature-induced structure expansion is considered to be developed along this length as a whole, as the reference temperature, then the point at time The vertical thermal expansion pseudo-displacement caused by temperature changes is , which is deducted from the measured vertical displacement , i.e. the vertical displacement after thermal compensation, and the compensation processing method in the east-west and north-south directions is the same; reflects the temperature change amplitude, and the greater the value, the greater the thermal expansion effect, which is proportional to the pseudo-displacement amount, reflects the thermal deformation sensitivity, and the greater the value, the greater the material thermal expansion or the structure size, which amplifies the temperature effect; when the temperature rises, the material expansion causes the vertical displacement measurement value to be positively offset (false uplift), and the vertical thermal expansion pseudo-displacement needs to be deducted to obtain the true deformation; Wherein, the parameters of the vertical displacement after temperature compensation are shown in Table 1.

[0021] Table 1

[0022] The original vertical displacement gradually increases from 10 for sample 1 to 17 for sample 15, indicating that the material's deformation properties change at different temperatures as the sample advances. While the change in the material's linear expansion coefficient is not significant, it increases slightly to 0.000011 for samples 6 to 10, likely reflecting the effect of changing material properties on thermal expansion. When analyzing the relationship between current temperature and vertical displacement after temperature compensation, the current temperature shows an increasing trend with increasing sample number, gradually increasing from 20°C to 37°C. The corresponding vertical displacement after temperature compensation increases significantly to 12.47 for sample number 6 and reaches 14.35 for sample number 10. This indicates that as the current temperature increases, the vertical displacement after temperature compensation also increases, reflecting the influence of temperature on material deformation. The reference temperature remains unchanged at 20°C, further emphasizing that the change in current temperature is a key factor affecting the vertical displacement after temperature compensation. From sample numbers 9 to 15, although the original vertical displacement continues to increase, the increase in the vertical displacement after temperature compensation is significantly reduced. This may be because at higher temperatures, the thermal expansion effect of the material begins to stabilize; the relationship between the vertical displacement after temperature compensation and the original vertical displacement, the linear expansion coefficient of the material, the current temperature and the reference temperature shows that the thermal expansion effect plays an important role in the deformation behavior of the material. As the sample number increases, the increase in current temperature directly leads to an increase in the vertical displacement after temperature compensation, emphasizing the impact of temperature on structural health assessment. In practical applications, these data can provide an important basis for material selection and engineering design to optimize the stability and safety of the structure.

[0023] Equivalent observation length , refers to the distance along which the displacement of the structure caused by thermal expansion is considered as an overall change by the radar at the monitoring point. Its physical meaning and acquisition methods are usually two kinds: structural drawings or design parameter method and experimental calibration method: For structural drawings or design parameter methods, if the i-th key monitoring point is located at a beam section, slab thickness, or expansion joint section, the actual length of this section is usually used as the reference length for thermal expansion—that is, the nominal length of the entire material stretching or contracting when the temperature changes. For example, if the monitoring point is located in the center of the bridge deck, then The thickness of the plate or a small section in the span can be used, such as 1 m or 2 m, depending on monitoring needs. If located on the surface of a vertical steel beam, the length of the corresponding section on the beam span can be used. The advantage of this is that a clear and engineering-understandable expansion length can be obtained directly from the structural blueprint of the bridge or on-site measurement. For the experimental calibration method, a known temperature difference calibration test is performed on a monitoring point under field or laboratory conditions: the temperature is artificially changed, and the displacement converted from the corresponding phase change of the radar is measured, and then the displacement can be obtained. , the equivalent length of the whole inflation behavior is concentrated on, so that the theoretical compensation formula is consistent with the measured value ; On the predefined grid of the bridge deck, for the grid unit where the ith key monitoring point is located, the set of its neighborhood units is denoted as , the spatial distance between the ith key monitoring point and the jth key monitoring point in the set of its neighborhood units is calculated , j is the index of the key monitoring point in the set of neighborhood units , and a Gaussian type weight is defined , the three-dimensional displacement field data of the environmental corrected key monitoring point i at time is outputted, and the structure temperature and wind speed obtained by inversion are also outputted; reflects the proximity of spatial position, and the greater the value, the weaker the neighborhood correlation, reflects the contribution weight of neighborhood points, and the greater the value, the greater the influence of nearby points; Filtering is to eliminate sudden changes caused by radar measurement noise or local interference and to retain the true structure deformation trend. The Gaussian weight ensures that nearby points have greater influence, which conforms to the spatial continuity principle of engineering structure deformation; Firstly, the bridge deck is pre-divided into regular grids, such as rectangular grids, triangular grids, etc. For regular rectangular grids, units that share a side or a vertex with the ith key monitoring point are all considered as neighborhood, i.e. eight-connected or four-connected manner. Four-connected shares four directions (up, down, left, right) of the side, and eight-connected includes four units in diagonal direction in addition to four-connected; First, the coordinates of the center point of each unit are calculated , and a spatial radius threshold R is set. All key monitoring points j that satisfy are included in the neighborhood unit set to complete the preliminary screening. Then, in the screening results, the distances between each unit and all other units are sorted from small to large, and the first K nearest units are selected to form the neighborhood, ensuring that the neighborhood is neither too large in sparse areas nor too many in dense areas; According to the characteristics of bridge structure, such as beam span, support position or stress concentration area, the neighborhood range of each unit is adaptively adjusted. For example, units near the support are spatially dense, but their structural response is relatively independent, so R or K can be appropriately reduced. The neighborhood of the middle region can be increased to capture the overall synergistic effect. According to the density of the monitoring grid, the span of the bridge, the monitoring target and the calculation resources, appropriate neighborhood determination strategies are selected to ensure the accuracy and efficiency of the coherence analysis and smoothing filtering.

[0024] Step 2: Extract the vertical component from the corrected three-dimensional displacement field, and perform variational mode decomposition on it to separate the static deformation field and the dynamic vibration field, calculate the space-time coherence coefficient matrix based on the dynamic vibration field, define the area where the coherence drops by more than 30% as the decoupling area, extract the Hilbert instantaneous amplitude variance and instantaneous frequency variance, the maximum peak value of bispectrum, the recursive quantization index and the vibration energy distribution entropy of the decoupling area, and output as the damage sensitive feature set; The logic for extracting the vertical component from the corrected three-dimensional displacement field and performing variational mode decomposition on it to separate the static deformation field and the dynamic vibration field is as follows: The vertical component of the i-th key monitoring point is decomposed by variational mode decomposition, and is classified according to the size of the center frequency, which is divided into low-frequency static deformation component and high-frequency dynamic vibration component , is the static deformation of the i-th key monitoring point at time , is the dynamic vibration of the i-th key monitoring point at time ; reflects the slow deformation of the bridge, such as creep and foundation settlement, and the larger the value, the more severe the long-term deformation accumulation, and is positively correlated with the load history and environmental temperature; reflects the instantaneous vibration of the bridge, such as wind vibration and vehicle vibration, and the larger the value, the stronger the dynamic response, and is positively correlated with the decay of structural stiffness or load excitation; variational mode decomposition separates signals according to center frequency, and static component has a center frequency close to 0Hz, and dynamic component is concentrated near the structure's fundamental frequency, such as 0.5-10Hz; set the time window , extract its vertical dynamic vibration component to calculate the signal sequence and in the time window , calculate the coherence coefficient matrix to obtain its time window coherence coefficient matrix : ; wherein, is the time variable from to , the average coherence coefficient of the i-th key monitoring point is defined as , when compared with the previous time drops by more than 30%, i.e. , the grid element identification of the i-th key monitoring point is judged as a decoupling area, and p is used to index the decoupling area; the processing method in the east-west and north-south directions is the same; reflects the local dynamic response strength of the structure, and the larger the value, the greater the vibration energy, Reflects the similarity of two-point vibration waveforms, the larger the value, the stronger the synchronization of structural dynamic response, proportional to the integral value of signal product; Reflects the overall synergy of the monitoring point and its neighborhood, the larger the value, the stronger the local dynamic coupling of the structure, Reflects the sudden drop in coherence, the larger the value, the greater the local stiffness mutation of the structure, Reflects the loss of absolute synergy, the larger the value, the more complete the decoupling of the structure; For the decoupling area p marked, the dynamic vibration component is subjected to Hilbert transform to extract the instantaneous amplitude and instantaneous frequency, the specific steps are as follows: Let denote the Hilbert transform result of , and then construct the analytic signal: ; Where, is the imaginary unit, , the modulus of the analytic signal , that is, the instantaneous amplitude of the pth decoupling area at time , the phase , and the derivative of the phase is , which is the instantaneous frequency of the decoupling area; Hilbert transform converts real signals into 90° phase-shifted virtual parts to form analytic signals on the complex plane, which is used to avoid negative frequency interference and accurately extract instantaneous characteristics; Reflects the instantaneous intensity of vibration energy, the larger the value, the greater the impact load or resonance, Reflects the instantaneous value of the main frequency, the larger the value, the greater the change in effective stiffness of the structure; Calculate the variance of the instantaneous amplitude and the instantaneous frequency of the variance and , when or If it exceeds 3 times the standard deviation of the baseline value of the healthy state, the point is considered to have abnormal fluctuations, indicating signs of structural damage; ; Damage condition: or ; Reflects the degree of amplitude fluctuation, the larger the value, the greater the nonlinear vibration, Reflects the degree of frequency fluctuation, the larger the value, the more significant the stiffness time-varying nature; , is the mean and standard deviation of the historical data of the healthy state.

[0025] For the dynamic vibration signal of the decoupling area p Implement delay embedding reconstruction, construct its phase space trajectory, and extract the following two recursive quantitative analysis indicators: Certainty Index and stratification , the certainty index indicates the proportion of linear repetition segments in the phase space trajectory, and the stratification degree indicates the proportion of time the phase space trajectory stays in the same state; The purpose of reconstruction is to convert the one-dimensional vibration signal into a high-dimensional phase space trajectory to reveal the hidden characteristics of the dynamic system. The trajectory of the healthy structure is regular and orderly, while the trajectory of the damaged structure is divergent and chaotic. The deterministic measure is the proportion of segments on the diagonal line (recovering to the same or similar state) on the phase space trajectory, which can also be understood as the regularity of the signal's recurrence: ; in, Is the length of The number of repeated segments of the phase space trajectory, is the minimum length of a line segment to be counted, is the maximum length of the line segment that is counted; the sum of the total lengths of all consecutive repeated line segments in the numerator reflects the cumulative time that the system repeats the state along the diagonal in the phase space, and the sum of the lengths of all repeated points (U-turn trajectories) in the denominator, including isolated points, represents the time of all repeated behaviors; The higher it is, the more and longer the segments of the signal that repeat along the diagonal direction in the phase space are, and the more predictable and linear the system trajectory is. In the early stage of damage initiation or crack propagation, due to the enhanced nonlinear effect, will decrease significantly; If there are more and longer diagonal structures in the trajectory, that is, the similar state is maintained for a long time, Pair length A substantial increase, Increase; when the system becomes chaotic or cracks trigger mutations, the phase space trajectory tends to spread out, with short segments Rising, long segments decreasing, reduce; The stratification measure is the proportion of points in the phase space trajectory that stay in the same state (same row / column), also known as the vertical line percentage: ; in, The length of the vertical line is The numerator is the total length of time the system stays in a state, and the denominator is the total length of time of all repeated points appearing in the recursive graph. The higher, the longer the system hovers or stays between certain states, and the more viscous or lagging the dynamic response is; when cracks initiate or locally relax, vertical vibration may hover in certain modes, vertical line segments increase, leading to rise when trajectories repeatedly appear in the same phase space region, vertical line segments are more and longer, and increase, increase, the system changes rapidly or the degree of chaos intensifies, vertical aggregation decreases, decreases; For the marked decoupling area p, smooth the vertical dynamic vibration signal Do second-order bispectrum analysis to detect high-order nonlinear coupling characteristics, and the formula for defining the second-order bispectrum function is: ; Among them, is the Fourier transform of the signal , the complex conjugate is represented by , and the statistical average of multiple time windows is represented by ; Measures the third-order phase coupling between the frequency pair and its sum frequency If the signal is a pure linear Gaussian process, the should theoretically be zero. The larger the peak value at the frequency pair, the more significant the non-Gaussian, nonlinear interaction between the three components of the system vibration; If there is a frequency pair such that is significantly higher than the historical baseline value of the corresponding frequency pair in the healthy state of the region, it indicates that the signal has produced significant non-Gaussian coupling at these two frequencies. Record the maximum peak value as ; And use it as a high-order nonlinear damage indicator; The higher, the stronger the nonlinear coupling produced by crack initiation or relaxation, and it is a sensitive feature of high-order damage. When nonlinear coupling intensifies, the of a certain frequency pair or group of frequency pairs significantly increases, and the peak value rises. When the structure is healthy, there is no specific phase locking of each frequency component, and the bispectrum peak tends to zero. For each marked decoupling area p, take its vertical dynamic vibration signal Band-pass filter within 5-20Hz to obtain the band-pass signal, divide the band-pass signal into N segments within the time window, calculate the instantaneous total energy of the first segment, and then obtain the energy proportion, and then define the energy distribution entropy: where is the first Segment energy proportion, Measure the dispersion degree of the energy in the selected frequency band in each sub-segment of the region; Measure the dispersion / uniformity of energy in N sub-segments, if energy is concentrated in a few segments, such as crack resonance, Distribution skewness, Smaller; The higher, the vibration energy dispersion, the system may enter the multi-modal coupling or broadband excitation state, Significant changes often correspond to energy redistribution before damage or crack opening change; When the energy of some sub-segments Increases or decreases sharply, so that Become uneven, Down; energy is evenly distributed in more sub-segments, Up; Finally, the above data extracted from the decoupling area p form the damage sensitive feature set.

[0026] Step 3: Normalize each component of the damage sensitive feature set according to the historical baseline data of the bridge health state, determine the feature weight using the entropy weight method and fusion to generate a dynamic damage index, calculate the curvature change of the whole bridge structure based on the static deformation field, and fuse the dynamic damage index to form the current time. Structure comprehensive degradation degree; The logic of normalizing each component of the damage sensitive feature set according to the historical baseline data of the bridge health state is as follows: The damage sensitive feature set extracted from the decoupling area p , the oth feature component Subtract its mean value under the bridge health state, and divide by the corresponding standard deviation to get the dimensionless standard component , o is the index of the feature component in the damage sensitive feature set, and m is the total number of feature components in the damage sensitive feature set; Reflects the original intensity of the damage feature, the larger the value, the more severe the damage, Reflects the standard deviation multiple of the deviation from the health baseline, the larger the value, the more significantly higher the current value of the feature than the health mean, the more serious the deviation, suggesting that the damage signal corresponding to the feature is more significant. It indicates the deviation of the feature at this time relative to the health state (historical baseline), and the unit becomes several standard deviations; when the original feature Increases, the normalized value Linearly increases; By calculating the information entropy of each feature in the historical sample, and determining the weight of each feature according to the inverse proportion principle of information entropy​ The greater the weight, the stronger the feature's ability to distinguish state differences. The dynamic damage indicator of the final decoupling area p is expressed as the weighted sum of each standard feature: ; Wherein, is regarded as a comprehensive quantitative value reflecting the dynamic damage signals of the current area, such as nonlinear characteristics, spectral anomalies, and coupling structure variations. It comprehensively reflects the overall intensity of various dynamic damage signals, including nonlinear coupling, high-order spectral characteristics, and energy distribution anomalies. Reflects the overall damage intensity of the area. The greater the value, the more severe the dynamic damage, and the more likely the structure is to be unhealthy, such as crack development and relaxation instability. When there is an anomaly, the greater the value, the more severe the damage, When the state is better than the historical baseline; The greater the weight or its corresponding weight , the higher the ; , Wherein is the number of healthy state historical samples, is the proportion of feature o in the kth sample, is the information entropy of feature o, which is used to measure the degree of dispersion, and is the weight of feature o; Reflects the dispersion of feature data. The greater the value, the more uniform and disordered the data, Reflects the importance of the feature. The greater the value, the more discriminative it is in the historical samples, and the greater its contribution to the comprehensive indicator. The smaller the fluctuation of the feature in the healthy samples, i.e. , the more significant the change when damaged, i.e. the greater the weight ; The greater the information entropy, the greater the fluctuation of the feature itself in the healthy state, and the weaker its ability to distinguish between health and damage, so it should be given a lower weight. The smaller the information entropy, the stronger the difference, and the higher the weight given. The static part of the bridge deck deformation field , the spatial second-order Laplacian is extracted, and the curvature change from the initial state is calculated: ; Wherein, is the time when the bridge deck is built; Reflects the local curvature of the displacement field, Reflects the change in relative curvature; the current static deformation increases, increases positively; if the deformation recovery approaches the initial state, then approaches 0; Curvature The curvature or the degree of concave-convex of the deformation field, and the difference with the initial state Then reflects the evolution of permanent deformation due to creep, settlement or material degradation, the greater the more the curvature of the deformation rises in this area, the structure may appear uneven down, local settlement or larger creep; The dynamic damage index And the curvature change After normalization, the degradation degree index is obtained by using a linear fusion model : ; Where, The normalized dynamic damage index and the curvature change, The weight coefficient is set according to the structural characteristics of the bridge; this formula quantifies the dual degradation effects of dynamic abnormalities and static deformation of the structure; Reflects the regional degradation degree, the greater the value, the lower the structural safety, the worse the overall health of the structure in the current working condition, both dynamic vibration abnormalities and static deformation accumulation, Reflects the dynamic damage contribution weight, which needs to be increased for dynamic sensitive structures ; Dynamic index Captures short-term damage such as impact and crack propagation, and static curvature Reflects cumulative damage such as creep and corrosion; the two complement each other to avoid missed detection; When dynamic abnormalities dominate, Greater, The increase of Significantly increases, when static deformation dominates, Smaller, The increase of Promotes Rise more significantly.

[0027] Wherein the calculation of the structure degradation degree part parameters is shown in Table 2.

[0028] Table 2

[0029] Through the analysis of the data in the first 15 rows of the table, it can be observed that there is a certain correlation between the dynamic damage index and the static curvature change, for example, the dynamic damage index and the static curvature change increase with the increase of the sample number. Specifically, as the sample number increases from 1 to 15, the dynamic damage index increases from 0.1 to 0.8, and the static curvature change increases from 0.2 to 0.9. This shows that the damage degree gradually increases under the influence of continuous dynamic load, accompanied by changes in static curvature; When analyzing the impact of the fusion weight on the structural degradation degree, the numerical change in the fusion weight was small as the sample number increased, from 0.1 to 0.55. Although the change in weight was relatively stable, it significantly affected the calculation of the final structural degradation degree. For example, the fusion weight of sample number 4 was 0.15, corresponding to a structural degradation degree of 0.26, while the fusion weight of sample number 15 was 0.55, and the structural degradation degree increased to 0.73. This shows that in the comprehensive evaluation of dynamic damage indicators and static curvature changes, the setting of the fusion weight plays a key role in the final result.

[0030] The weighted combination of dynamic damage index and static curvature change influences the assessment of structural health. For example, in sample number 10, the dynamic damage index is 0.55, the static curvature change is 0.65, and the fusion weight is 0.35, resulting in a calculated degradation degree of 0.53. This combination demonstrates that even under high dynamic damage conditions, the contribution of static curvature change cannot be ignored, and its results are crucial for assessing the structural health. In summary, the relationship between dynamic damage index, static curvature change, and fusion weight reflects the complexity of structural health assessment.

[0031] Step 4: Construct a historical operating condition database during the bridge operation monitoring cycle. This database includes light-load and heavy-load operating condition categories based on the entropy of the vibration energy distribution, as well as weak-wind and strong-wind operating condition categories based on wind speed. For each operating condition, calculate the 90% confidence interval of the structural comprehensive degradation degree. Then, match the corresponding operating condition category based on the current load and wind speed status. Calculate the Mahalanobis distance of the current comprehensive degradation degree relative to the historical distribution of the operating condition to identify structural risks. A historical operating condition database is constructed during the bridge operation monitoring period. The database includes light load and heavy load operating condition categories based on the entropy of the vibration energy distribution, as well as weak wind and strong wind operating condition categories based on wind speed: Based on the external interference and load changes of the bridge structure during operation, the working conditions of each monitoring period are divided into the following four categories: , it is determined to be a light load condition, otherwise it is a heavy load condition; if the wind speed , is a weak wind condition, otherwise it is a strong wind condition, where 、 The two combinations are divided into 4 typical working conditions: light load weak wind, light load strong wind, heavy load weak wind, heavy load strong wind, and Index these categories; Reflects the uniformity of the time distribution of vibration energy. The larger the value, the more random the energy dispersion. Reflects the critical value of load intensity, usually taken from historical the median; Reflects the intensity of the environmental wind load, the greater the value indicates the enhancement of wind-induced vibration, Reflects the critical value of wind vibration influence, usually the upper limit of 6-level wind; From below The threshold value, the classification is switched to strong wind, and below the threshold value is weak wind; Vibration energy distribution entropy The lower, the more concentrated the energy in a few frequency bands, often corresponding to local modal response under small load, and the higher the entropy, the more dispersed the energy, indicating the superposition effect of multi-modal vibration caused by live load such as vehicle; When From below the threshold Increase and cross the threshold, switch from light load classification to heavy load; as long as Keep below All belong to light load; For each type of working condition , record the corresponding structure comprehensive degradation degree in its history period, calculate its 90% confidence interval And the mean And the covariance matrix ; The upper and lower 5% samples outside the interval are considered as extreme values, corresponding to the rare or known relatively serious structure state in history; Represent the normal degradation level under the typical working condition, Then describes the fluctuation range and mutual correlation of the index under this working condition; For any sampling time t, if its corresponding current working condition category is , its comprehensive degradation degree is recorded as , and its Mahalanobis distance relative to the historical distribution of the working condition is defined as: ; Where, The Mahalanobis distance, which represents the deviation of the current structure state relative to the same working condition, has the advantages of dimensionless and variance sensitivity, Is the transpose of the matrix; if Exceeds the 95% upper limit of the maximum value in history, it is considered as a potential structure risk event, triggering an alarm; Reflects the absolute deviation from the benchmark, the greater the value indicates the higher the damage possibility, Reflects the historical fluctuation correction, the greater the value indicates the suppression of false positives in high fluctuation working conditions, Is a column vector, indicating the deviation of the current value from the historical mean of the category, Transpose this column vector to a row vector; Is the current comprehensive degradation degree And the distribution center of the historical working condition of the same category The normalized Euclidean distance between The larger the value, the more serious the deviation of the current structural state from the typical health center in the multidimensional index space. Comparison Increase or decrease, exceeding its The standard deviation in Increase; if Exactly equal to ,but , indicating that it is consistent with the historical typical state; when Exceeding the upper limit of the historical 95% range means that the current state has exceeded the previous 95% historical normal fluctuation limit, which may indicate structural risks such as crack expansion and changes in support constraints.

[0032] The present invention further provides a bridge structure monitoring device based on a microwaveform variable radar, wherein the device is used to perform the above-mentioned bridge structure monitoring method based on a microwaveform variable radar, comprising: The data collection module is used to perform high-frequency scanning of the bridge surface using a micro-waveform radar, generate full-field three-dimensional displacement and its rate of change through phase interferometry, and simultaneously invert the structure surface thermal field distribution and bridge deck wind speed information from the radar echo signal. The displacement field is then subjected to thermal strain compensation and spatial continuity filtering, and the environmentally corrected three-dimensional displacement field data, structure surface temperature field, and wind speed field are output; The data processing module is used to extract the vertical component from the corrected three-dimensional displacement field and perform variational modal decomposition on it to separate the static deformation field and the dynamic vibration field. The spatiotemporal coherence coefficient matrix is ​​calculated based on the dynamic vibration field. The region where the coherence drops by more than 30% is defined as the decoupling zone. The Hilbert instantaneous amplitude variance and instantaneous frequency variance, the maximum peak of the bispectrum, the recursive quantization index, and the vibration energy distribution entropy are extracted from the decoupling zone, and the output is a damage-sensitive feature set. The comprehensive calculation module is used to normalize the components of the damage-sensitive feature set based on the historical baseline data of the bridge health status, determine the feature weights using the entropy weight method, and fuse them to generate a dynamic damage index. The curvature change of the entire bridge structure is calculated based on the static deformation field, and is integrated with the dynamic damage index to form the comprehensive structural degradation degree at the current moment; The risk identification module is used to build a historical operating condition database during the bridge operation monitoring cycle. The database includes light-load and heavy-load operating condition categories distinguished by the entropy of the vibration energy distribution, as well as weak-wind and strong-wind operating condition categories divided by wind speed. The 90% confidence interval of the comprehensive structural degradation degree is calculated for each operating condition, and the corresponding operating condition category is matched in combination with the current load and wind speed status. The Mahalanobis distance of the current comprehensive degradation degree relative to the historical distribution of the operating condition is calculated to achieve structural risk identification.

[0033] The above formulas are all dimensionless values calculated, the formula is obtained by collecting a large amount of data to simulate the most recent real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0034] The above embodiments can be implemented wholly or partially by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented wholly or partially in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0035] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0036] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for monitoring a bridge structure based on microwave deformation radar, characterized by, The specific steps include: Step 1: High-frequency scanning of the bridge surface is performed using a microwave deformation radar, a full-field three-dimensional displacement and its rate of change are generated by phase interference, and the structural surface thermal field distribution and bridge deck wind speed information are obtained by inversion from the radar echo signal, the displacement field is subjected to thermal strain compensation and spatial continuity filtering, and the three-dimensional displacement field data after environmental correction, the structural surface temperature field and the wind speed field are output; Step 2: The vertical component is extracted from the corrected three-dimensional displacement field, and variational modal decomposition is performed thereon to separate the static deformation field and the dynamic vibration field, the time-space coherence coefficient matrix is calculated based on the dynamic vibration field, the area where the coherence drops by more than 30% is defined as the decoupling area, the Hilbert instantaneous amplitude variance and instantaneous frequency variance, the bispectrum maximum peak value, the recursive quantization index and the vibration energy distribution entropy are extracted from the decoupling area, and the output is a damage sensitive feature set; Step 3: Each component of the damage sensitive feature set is normalized according to the historical baseline data of the bridge health state, the entropy weight method is used to determine the feature weight and fusion to generate a dynamic damage index, the curvature change of the whole bridge structure is calculated based on the static deformation field, and the dynamic damage index is fused to form the current structural comprehensive degradation degree; Step 4: A historical working condition database is constructed within the bridge operation monitoring period, which includes light load and heavy load working condition categories distinguished according to vibration energy distribution entropy, and weak wind and strong wind working condition categories divided according to wind speed; the 90% confidence interval of the structural comprehensive degradation degree is calculated for each working condition, and the corresponding working condition category is matched according to the current load and wind speed state, the Mahalanobis distance of the current comprehensive degradation degree relative to the historical distribution of the working condition is calculated, which is used to realize the structural risk identification.

2. The microwave deformation radar-based bridge structure monitoring method according to claim 1, characterized in that: The logic for using a microwave deformation radar to perform high-frequency scanning of the bridge surface to generate a full-field three-dimensional displacement and its rate of change is as follows: Determine the key monitoring area of the bridge to be monitored, which includes piers, bridge deck centers, abutments, bridge connection areas and bearings, then divide each key monitoring area into a plurality of grid units, index these grid units with i, and arrange key monitoring points at the center of each grid unit, then the key monitoring point at the center of the i-th grid is the i-th key monitoring point; The microwave deformation radar scans the key monitoring points on the bridge deck with a continuous wave signal with a frequency of , wherein the radar signal wavelength , wherein is the speed of light, is the sampling time, the radar continuously scans the same point, and the phase difference between the adjacent two echoes is recorded, and according to the phase interference principle, the radial displacement of the point along the radar line-of-sight direction is obtained, and the formula is: ； wherein, is the phase difference of the i-th key monitoring point at time is the radial displacement of the i-th key monitoring point at time is the radial displacement of the i-th key monitoring point at time​ The pitch angle of the i-th key monitoring point relative to the radar and the azimuth angle The radial displacement is projected into the local coordinate system of the bridge deck to obtain three-axis displacement variation components: ； wherein, , and are respectively the instantaneous displacement variation components of the bridge in the east-west, north-south and vertical directions. For each key monitoring point i, the three-axis displacement increment is divided by the time interval to obtain the three-axis instantaneous displacement change rate; in the same radar echo, the reflected signal intensity is and spectrum broadening Analysis of the surface temperature field of the structure and bridge deck wind speed field , the structure temperature Using pre-calibrated constants and , the observed Mapped to corresponding temperature : , the bridge deck wind speed By formula Get; Considering the thermal expansion pseudo-displacement of the bridge material caused by temperature change, the compensation for each key monitoring point i is as follows: set αi as the linear expansion coefficient of the bridge material corresponding to the ith key monitoring point, Li as the equivalent observation length of the ith key monitoring point, at which the temperature-induced structural expansion is considered to be developed along the length, T0 as the reference temperature, then the point at time The vertical thermal expansion pseudo-displacement caused by temperature change is Subtract this value from the measured vertical displacement , that is, the vertical displacement after thermal compensation, and the compensation processing method in the east-west and north-south directions is the same; On the predefined grid of the bridge deck, for the grid cell where the ith key monitoring point is located, the set of its neighborhood cells is denoted as , the spatial distance between the ith key monitoring point and the jth key monitoring point in the set of its neighborhood cells is calculated , j is the index of the key monitoring point in the set of neighborhood cells , and a Gaussian-type weight is defined. In a weighted average manner, the thermal-compensated displacement of each key monitoring point is smoothed in the three-axis direction, and the final output is the three-dimensional displacement field data of the environmental-corrected key monitoring point i at time , while the structure temperature and wind speed obtained by inversion are also output.

3. The microwave deformation radar-based bridge structure monitoring method according to claim 2, characterized in that: The logic for extracting the vertical component from the corrected three-dimensional displacement field and performing variational modal decomposition thereon to separate the static deformation field and the dynamic vibration field is as follows: The vertical component of the i th key monitoring point is decomposed by using variational mode decomposition, and is classified into a low-frequency static deformation component and a high-frequency dynamic vibration component according to the size of the central frequency . is the static deformation of the i th key monitoring point at time , is the dynamic vibration of the i th key monitoring point at time .​ Setting time window , extracting its vertical dynamic vibration component to calculate the signal sequence in the time window and , calculating the coherence coefficient matrix to obtain its time window coherence coefficient matrix :​ ； Wherein, is to Time variable, define the average coherence factor of the ith key monitoring point When Compared with the last time Fall by more than 30%, that is When, judge the grid unit mark where the ith key monitoring point is located as a decoupling area, and index the decoupling area with p. For the marked decoupling zone p, the dynamic vibration component The Hilbert transform is performed to extract the instantaneous amplitude and instantaneous frequency, and the specific steps are as follows: Let denote the Hilbert transform of and construct the analytic signal as ； wherein, is the imaginary unit, the modulus of the analytic signal is the instantaneous amplitude of the p-th decoupling zone at time with phase The derivative of the phase is the instantaneous frequency of the decoupling zone. The variance of the instantaneous amplitude and the variance of the instantaneous frequency within the computation window and When or If it exceeds the baseline value of the health state by 3 standard deviations, the point is considered to have abnormal fluctuations, indicating signs of structural damage.

4. The microwave deformation radar-based bridge structure monitoring method according to claim 3, characterized in that: Dynamic vibration signals for decoupling region p The delay-embedding reconstruction is implemented, the phase space trajectory is constructed, and the following two recursive quantification analysis indexes are extracted: deterministic index and layering degree The deterministic index represents the proportion of linear repeating segments in the phase space trajectory, and the layering degree represents the proportion of time length that the phase space trajectory stays in the same state. For the marked decoupling region p, its smooth vertical dynamic vibration signal Second-order bispectrum analysis is performed to detect high-order nonlinear coupling characteristics. The formula for defining the second-order bispectrum function is: ； wherein is a signal is the Fourier transform of the signal denotes complex conjugation denotes a statistical average over a plurality of time windows If there is a pair of frequencies such that is significantly higher than the historical baseline value of the pair of frequencies in the healthy state, it indicates that the signal has produced a significant non-Gaussian coupling at these two frequencies, and the maximum peak is recorded as: ； And it is used as a high-order nonlinear damage index; For each marked decoupling area p, take its vertical dynamic vibration signal Band-pass filter within 5-20Hz to obtain a band-pass signal, divide the band-pass signal into N segments in time window, calculate the instantaneous total energy of the first segment, and then obtain the energy proportion, and then define the energy distribution entropy: , wherein is the energy proportion of the first segment, measures the dispersion degree of the energy of the area in the selected frequency band on each sub-segment; Finally, the above data extracted from the decoupling area p form a damage sensitive feature set.

5. The microwave deformation radar-based bridge structure monitoring method according to claim 4, characterized in that: The logic for normalizing each component of the damage sensitive feature set according to the historical baseline data of the bridge health state is as follows: A set of damage sensitive features extracted for decoupling zone p where the oth feature component is subtracted from its mean value at the bridge health state and divided by the corresponding standard deviation to obtain a dimensionless standard component , o is the index of the feature component in the set of damage sensitive features, m is the number of feature components in the set of damage sensitive features The total number of the bridge health state historical baseline data is determined according to the historical baseline data of the bridge health state, and the total number of the bridge health state historical baseline data is determined according to the historical baseline data of the bridge health state. The information entropy of each feature is calculated by counting the discrete degree of each feature in the historical samples, and the weight of each feature is determined according to the inverse proportion principle of information entropy The greater the weight is, the stronger the distinguishing ability of the feature to the state difference is. Finally, the dynamic damage index of the decoupling area p is expressed as the weighted sum of each standard feature. ； wherein, is regarded as a comprehensive quantitative value reflecting the dynamic damage signals of the nonlinear characteristics, spectral anomalies, coupling structure variations, etc. of the current region; from the static part of the bridge deck deformation field , extracting the spatial second order laplacian and calculating the curvature change from the initial state : ； wherein t0is the time of construction of the bridge deck; The dynamic damage index is calculated by the following equation: The curvature change amount is calculated by the following equation: After the normalization processing, the deterioration degree index is obtained by using a linear fusion model : ； wherein, is the normalized dynamic damage index and the curvature change, is the weight coefficient, which is set according to the structural characteristics of the bridge.

6. The microwave deformation radar-based bridge structure monitoring method according to claim 4, characterized in that: A historical working condition database is constructed within the bridge operation monitoring period, which includes light load and heavy load working condition categories distinguished according to vibration energy distribution entropy, and weak wind and strong wind working condition categories divided according to wind speed: Based on the external interference and load changes of the bridge structure in the operation process, the working conditions of each monitoring time period are divided into the following four categories: if the vibration energy distribution entropy , it is determined as light load working condition, otherwise as heavy load; if the wind speed , it is weak wind working condition, otherwise as strong wind, wherein , are all set threshold values; the combination of the two is divided into four typical working condition categories: light load weak wind, light load strong wind, heavy load weak wind, heavy load strong wind, and is used to index these categories; For each type of working condition , record the corresponding comprehensive structural degradation degree in the historical period and calculate its 90% confidence interval and the mean and the covariance matrix ; For any sampling time t, if its corresponding current working condition category is , then its comprehensive deterioration degree is recorded as , and its Mahalanobis distance relative to the working condition history distribution is defined as: ； wherein, is the Mahalanobis distance, which represents the deviation of the current structural state from the same type of working condition, is the transpose of the matrix; if If it exceeds the 95% upper limit of the historical maximum value, it is considered a potential structural risk event, triggering an alarm.

7. A microwave deformation radar based bridge structure monitoring apparatus, characterized by: The device is used to perform the microwave deformation radar-based bridge structure monitoring method of any one of claims 1-6, comprising: The data collection module is used for high-frequency scanning of a bridge surface by using a microwave deformation radar, generating a full-field three-dimensional displacement and a change rate thereof by phase interference, simultaneously inversing a structure surface thermal field distribution and a bridge surface wind speed information from a radar echo signal, compensating a thermal strain of the displacement field, and filtering spatial continuity, and outputting an environmental corrected three-dimensional displacement field data, a structure surface temperature field and a wind speed field; The data processing module is used for extracting a vertical component from the corrected three-dimensional displacement field, and implementing variational mode decomposition on the vertical component to separate a static deformation field and a dynamic vibration field, calculating a space-time coherence coefficient matrix based on the dynamic vibration field, defining a region with a coherence drop of more than 30% as a decoupling region, extracting a Hilbert instantaneous amplitude variance and an instantaneous frequency variance, a bispectrum maximum peak value, a recursive quantization index and a vibration energy distribution entropy from the decoupling region, and outputting as a damage sensitive feature set; The comprehensive calculation module is used for normalizing each component of the damage sensitive feature set according to historical baseline data of a bridge health state, determining a feature weight by using an entropy weight method and fusing to generate a dynamic damage index, calculating a curvature change amount of the whole bridge structure based on the static deformation field, and fusing the dynamic damage index to form a structure comprehensive degradation degree at a current time; The risk identification module is used for constructing a historical working condition database in a bridge operation monitoring period, the database including a light load and a heavy load working condition category distinguished according to the vibration energy distribution entropy, and a weak wind and a strong wind working condition category divided according to a wind speed; for each working condition, a 90% confidence interval of the structure comprehensive degradation degree is calculated, and a corresponding working condition category is matched according to a current load and a wind speed state, a Mahalanobis distance of the current comprehensive degradation degree relative to a historical distribution of the working condition is calculated, and is used for realizing structure risk identification.

Images