A multi-modal sensor network-based full-cycle health monitoring method, system, device and medium for a superimposed station structure

Abstract

The application discloses a kind of based on multi-modal sensing network's superimposed station structure whole cycle health monitoring method, system, device and medium, belong to structural health monitoring technical field, comprising: obtaining the first monitoring data sequence of fiber optic sensor and the second monitoring data sequence of multiple wireless sensors;For each fiber optic sensor, based on spatial position coordinates, calculate the spatial distance between it and each wireless sensor, and the selected K wireless sensor constitutes the check unit of the fiber optic sensor;Get the local environmental disturbance reference corresponding to fiber optic sensor;The first volatility index and local environmental disturbance reference are standardized, calculate abnormal deviation degree, compare abnormal deviation degree with health state threshold value.The present application solves the technical problem that in the existing superimposed station structure health monitoring, environmental disturbance and sensor itself anomaly are mutually coupled, resulting in only relying on single monitoring data is difficult to accurately determine the health status of fiber optic sensor.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, and in particular to a method, system, device and medium for full-cycle health monitoring of composite station structures based on multimodal sensor networks. Background Technology

[0002] Composite station structures typically comprise multiple types of load-bearing units, including composite floor slabs, beams, columns, and their connecting components. During their service life, they are subjected to multiple sources of stress, such as train loads, pedestrian loads, equipment vibrations, temperature changes, and construction and maintenance disturbances, resulting in significant time-varying and spatial variations in structural response. Existing structural health monitoring schemes mostly rely on long-term recording and threshold determination of structural response using single-type sensors (e.g., fiber optic strain monitoring or wireless acceleration monitoring).

[0003] However, single-sensor systems are easily affected by factors such as local environmental disturbances, sensor aging / damage, and changes in installation conditions in engineering applications, making it difficult to distinguish between abnormal monitoring data and the actual state of the structure. For example, when there is strong environmental vibration in a local area, the output fluctuation of the fiber optic sensor may increase significantly; conversely, when the fiber optic sensor itself malfunctions or the coupling conditions change, the output fluctuation may also be abnormal, but it is difficult to determine whether the abnormality is "environmentally caused" or "sensor health problem" based solely on the fiber optic data. Therefore, there is an urgent need for a health monitoring method that can combine sensing information from different modes and construct a verification mechanism through spatial proximity relationships to improve the reliability and interpretability of full-cycle monitoring. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, device, and medium for full-cycle health monitoring of composite station structures based on multimodal sensor networks, in order to solve the technical problem mentioned in the background art that in the existing health monitoring of composite station structures, due to the mutual coupling between environmental disturbances and sensor anomalies, it is difficult to accurately determine the health status of fiber optic sensors by relying on a single monitoring data.

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

[0006] In a first aspect, the present invention provides a method for full-cycle health monitoring of composite station structures based on a multimodal sensor network, comprising: Within the same preset time window ΔT, a first monitoring data sequence from an optical fiber sensor and a second monitoring data sequence from multiple wireless sensors are acquired; wherein the optical fiber sensor and the wireless sensors have corresponding spatial coordinates. For each fiber optic sensor, the spatial distance between it and each wireless sensor is calculated based on the spatial location coordinates, and K wireless sensors with the closest spatial distance are selected from the wireless sensors that meet the preset search radius R; the fiber optic sensor and the selected K wireless sensors constitute the verification unit of the fiber optic sensor. Within the time window ΔT, a wireless quantization index is calculated for the second monitoring data sequence of each wireless sensor in the verification unit, and the wireless quantization index is weighted and fused based on the weight related to spatial distance to obtain the local environmental disturbance benchmark corresponding to the fiber optic sensor. Within the time window ΔT, a first volatility index is calculated for the first monitoring data sequence; and the first volatility index and the local environmental disturbance benchmark are standardized based on the historical baseline value of the first volatility index and the historical baseline value of the local environmental disturbance benchmark, respectively, to obtain a dimensionless first standardized index and a dimensionless benchmark standardized index. Anomaly deviation is calculated based on the ratio between the first standardized index and the benchmark standardized index. The anomaly deviation is then compared with a preset health status threshold to determine the health status of the fiber optic sensor. The health status includes at least one or more of the following: healthy, abnormal, and faulty.

[0007] Preferably, the preset search radius R is a preset positive value, and is selected based on the structural density and / or vibration propagation distance of the structural region where the fiber optic sensor is located; K is a positive integer not less than the preset minimum number of neighbors.

[0008] Preferably, the wireless quantization index is used to characterize the energy or fluctuation intensity of the second monitoring data sequence within the time window ΔT, and the wireless quantization index includes one or more of the following: root mean square value, standard deviation, peak-to-peak value, and energy index.

[0009] Preferably, the weights related to spatial distance are determined based on the spatial distance between the fiber optic sensor and each wireless sensor, and the sum of the weights in the verification unit is 1; wherein, the weights decrease as the spatial distance increases.

[0010] Preferably, the first volatility index is used to characterize the degree of volatility of the first monitoring data sequence within the time window ΔT, and the first volatility index includes one or more of standard deviation, root mean square value, peak-to-peak value, and energy index.

[0011] Preferably, the historical baseline value is obtained by dynamically calculating and updating historical data through a sliding time window, and the time span of the sliding time window is a preset duration.

[0012] Preferably, before comparing the abnormal deviation with a preset health status threshold, the method further includes: performing validity detection on the second monitoring data sequence of the wireless sensors in the verification unit within the time window ΔT; when the proportion of the number of wireless sensors that meet the preset invalid conditions to the K wireless sensors exceeds a preset proportion threshold, the verification unit is determined to be in an untrusted state, and the corresponding diagnostic results within the time window ΔT are marked as invalid or unverified.

[0013] Secondly, based on the same inventive concept, the present invention also provides a full-cycle health monitoring system for composite station structures based on a multimodal sensor network, comprising: The data acquisition module is used to acquire the first monitoring data sequence of the fiber optic sensor and the second monitoring data sequence of multiple wireless sensors within the same preset time window ΔT. The verification unit construction module is used to calculate the spatial distance between the fiber optic sensor and the wireless sensor based on the spatial location coordinates, and select the K wireless sensors with the closest spatial distance from the wireless sensors that meet the preset search radius R, and form a verification unit with the corresponding fiber optic sensor. The disturbance reference determination module is used to calculate the wireless quantization index and perform weighted fusion within the time window ΔT to obtain the local environmental disturbance reference. The deviation calculation module is used to calculate the first volatility index and perform standardization within the time window ΔT, and calculate the abnormal deviation based on the standardization result; The health diagnosis module is used to compare the abnormal deviation with a preset health status threshold to determine the health status of the fiber optic sensor. The system is configured to execute the above-described method.

[0014] Thirdly, based on the same inventive concept, the present invention also provides a health monitoring device, including a processor and a memory, wherein the memory stores a computer program, and when the computer program is run on the processor, the processor performs the above-described method.

[0015] Fourthly, based on the same inventive concept, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the above-described method.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention solves the technical problem in existing composite station structure health monitoring where the coupling between environmental disturbances and sensor anomalies makes it difficult to accurately determine the health status of fiber optic sensors by relying solely on single monitoring data.

[0017] This invention constructs verification units based on spatial proximity and utilizes second monitoring data from wireless sensors to form a local environmental disturbance benchmark. This provides an "environmental reference" for judging the output fluctuations of fiber optic sensors, reducing the probability of misjudging sensor health due to environmental disturbances. By introducing historical baselines and standardizing both the fiber optic side fluctuation index and the disturbance benchmark, dimensionless comparable indicators are formed, thereby more stably constructing the abnormal deviation and improving applicability across time periods and operating conditions. By comparing the abnormal deviation with health status thresholds, a graded judgment of health / abnormality / fault is achieved, facilitating the formation of executable inspection and handling strategies by the operation and maintenance side. By detecting the validity of wireless data and marking untrusted verification units, the robustness and reliability of results during long-term operation throughout the entire lifecycle can be further improved. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments are briefly described below. It should be understood that the drawings described herein are for illustrative purposes only and do not constitute a limitation on the scope of protection of the present invention; for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the full-cycle health monitoring method for composite station structures based on multimodal sensor networks provided in this embodiment of the invention; Figure 2 This is a structural block diagram of a multi-modal sensor network-based full-cycle health monitoring system for composite station structures provided in an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] In the following description, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0022] Furthermore, in this application, directional terms such as "upper," "lower," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and may change accordingly depending on the orientation of the components in the accompanying drawings.

[0023] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "coupled" can refer to an electrical connection that enables signal transmission.

[0024] As used herein, “about,” “approximately,” or “approximately” includes the stated value and the average value within an acceptable range of deviation from the given value, wherein the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the given quantity (i.e., the limitations of the measurement system).

[0025] This method utilizes an adaptive virtual redundancy verification mechanism and the collective behavior characteristics of wireless sensor networks to dynamically assess the reliability of individual data from a single node in a fiber optic sensor network in real time. This effectively solves the technical problem of difficulty in quickly and accurately distinguishing between the true structural response and single-point sensor failure in existing technologies, thus achieving the beneficial effects of improving the purity of monitoring data and the efficiency of system operation and maintenance.

[0026] Example 1: like Figure 1 As shown, this embodiment provides a method for full-cycle health monitoring of composite station structures based on multimodal sensor networks, including the following steps: S100, within the same preset time window ΔT (e.g., 1 second), acquire the first monitoring data sequence of the fiber optic sensor and the second monitoring data sequence of multiple wireless sensors; wherein, the fiber optic sensor and the wireless sensor have corresponding spatial position coordinates.

[0027] In this embodiment, the multimodal sensor network deployed on the composite station structure is activated and begins to collect relevant multimodal data. The fiber optic sensors in the fiber optic sensor network can be fiber grating (FBG) sensor arrays, and the fiber optic demodulator continuously collects and outputs data at a relatively high sampling frequency, such as center reflection wavelength data with a sampling frequency of 100Hz. This wavelength data can be converted into a strain data sequence after temperature compensation and calibration coefficient conversion, serving as the first monitoring data. Wireless sensors can include distributed nodes in the wireless sensor network (e.g., microelectromechanical systems (MEMS) accelerometers or wireless tilt sensors), collecting structural dynamic response data at their location at a relatively low sampling frequency, such as acceleration or tilt data sequences with a sampling frequency of 20Hz. This data is then aggregated to a central gateway via a wireless ad hoc network or star network to form the second monitoring data.

[0028] To ensure that the first monitoring data and the second monitoring data are comparable within the same preset time window ΔT, the system can assign a unified high-precision timestamp to data from different sources. The timestamp can be generated by the hardware clock of the Network Time Protocol (NTP) or the Precision Time Protocol (PTP). When the sampling rates do not match, upsampling processing can be performed on the low-frequency second monitoring data. For example, multiple new data points can be inserted between two original wireless data points spaced 50 milliseconds apart, so that the upsampled second monitoring data matches the 100Hz fiber optic data in terms of time resolution, thereby obtaining an aligned multimodal dataset within the same preset time window ΔT.

[0029] The spatial coordinates of the fiber optic sensor and the wireless sensor can be obtained through pre-mapping and stored in the system database. The spatial coordinates can be the position coordinates (x, y, z) in the three-dimensional coordinate system of the superimposed station structure.

[0030] S200, for each fiber optic sensor, calculate the spatial distance between it and each wireless sensor based on the spatial position coordinates, and select the K wireless sensors with the closest spatial distance from the wireless sensors that meet the preset search radius R; the fiber optic sensor and the selected K wireless sensors form the verification unit of the fiber optic sensor.

[0031] By leveraging spatial proximity, a high-precision fiber optic sensor and a group of wireless sensors are bound together as a functional unit (i.e., a verification unit). The system can load sensor network topology information during initialization. This topology information records the position coordinates (x, y, z) of each sensor (including fiber optic and wireless) in the three-dimensional coordinate system of the composite station structure.

[0032] For each fiber optic sensor The system is based on Using the coordinates as the center, search for wireless sensors within the spatial region corresponding to the preset search radius R, and calculate... The spatial distance (e.g., Euclidean distance) between each wireless sensor; among the wireless sensors that satisfy the preset search radius R, select the K wireless sensors with the closest spatial distance, and then compare them with... Together they form a verification unit The construction process of this verification unit can be executed periodically or triggered by events to adapt to situations such as the addition or removal of wireless nodes or minor adjustments in their positions within the network.

[0033] The preset search radius R is a preset positive value, and is selected based on the structural density and / or vibration propagation distance of the structural region where the fiber optic sensor is located; K is a positive integer not less than the preset minimum number of neighbors.

[0034] The parameter R is used to define the spatial proximity range of the verification unit. Preferably, the value of R is related to the size of the composite station structure components and the expected stress / vibration propagation range: for areas with compact structures and high beam-column density (such as the beam grid structure under the platform slab), the vibration propagation range is small, so a smaller R (e.g., R=3.0m) can be selected to ensure that the wireless sensor in the verification unit can more effectively reflect the local disturbances that are highly correlated with the fiber optic sensor; for open structural areas such as large-span roofs, the vibration propagates further, so R can be appropriately increased (e.g., R=8.0m).

[0035] The parameter K is used to ensure that the local environmental disturbance benchmark obtained by weighted fusion has statistical robustness and to avoid the excessive influence of a single wireless sensor anomaly on the disturbance benchmark. In some implementations, K can be no less than 3 so as to reduce the impact of occasional anomalies through multi-point fusion.

[0036] For example, fiber optic sensors The three-dimensional coordinates are (10.50, 20.30, 4.00) (unit: m). Based on the structural characteristics of this area, the preset verification unit construction parameters are: search radius R = 5.0 m, K = 3; the coordinates of nearby wireless sensors include... (11.10, 20.80, 4.20) (9.80, 19.50, 4.10) (10.20, 21.50, 3.80) (16.20, 22.10, 4.50). Calculated... and The distance is approximately 0.81m, and The distance is approximately 1.11m, and The distance is approximately 1.26m, and The distance is approximately 6.03m; based on this, the following was selected. , , And constitute a verification unit .

[0037] S300, within the time window ΔT, for the verification unit The second monitoring data sequences of each wireless sensor are used to calculate wireless quantization indicators, and these indicators are weighted and fused based on weights related to spatial distance to obtain the local environmental disturbance benchmark corresponding to the fiber optic sensor. The local environmental disturbance reference This is used to characterize the level of physical disturbance in the local area where the fiber optic sensor is located within the time window ΔT.

[0038] For the verification unit Each wireless sensor inside The system extracts its second monitoring data sequence within the time window ΔT and calculates a wireless quantization index to characterize the energy or fluctuation intensity of the second monitoring data sequence within the time window ΔT. The wireless quantization index can be a single-valued or low-dimensional index, used to compress the second monitoring data sequence of the wireless sensor within the time window ΔT into a fusionable representation.

[0039] In some implementations, the wireless quantification index includes one or more of the following: root mean square value, standard deviation, peak-to-peak value, and energy index; preferably, the root mean square value (RMS) can be used as the wireless quantification index, and the system calculates the corresponding RMS index for the second monitoring data sequence of K wireless sensors in the verification unit.

[0040] The system uses fiber optic sensors With each wireless sensor in the verification unit The spatial distance between them determines the weights associated with the spatial distance. And make the sum of all weights in the verification unit equal to 1; wherein the weights decrease as the spatial distance increases, so as to give higher weights to the wireless sensors that are closer to the central fiber optic sensor.

[0041] In this embodiment, the wireless quantization index is used to compress the second monitoring data sequence of the wireless sensor within a time window ΔT into a fusionable single-value or low-dimensional index to characterize the intensity of the local disturbance sensed by the wireless node. Preferably, the root mean square (RMS) value can be used as the wireless quantization index; in other optional embodiments, standard deviation, peak-to-peak value, or energy index can also be used, or multiple indices can be combined to enhance the sensitivity to different disturbance patterns.

[0042] In some implementations, the weights can be determined using a Gaussian decay function, for example: ; in, Represents the verification unit The distance weight corresponding to the j-th wireless sensor; For the j-th wireless sensor and fiber optic sensor Spatial distance, The distance scale parameter is used to control the weight decay rate; the system can normalize the above weights so that each verification unit... Internal weights The sum is 1. Compared to the simple reciprocal of the distance, this method can handle distance changes more smoothly and give higher weight to nearby sensors.

[0043] The system performs weighted fusion of K wireless quantization indicators based on the aforementioned weights to obtain a local environmental disturbance benchmark. For example, local environmental disturbance reference. It can be a weighted average of various wireless quantization indicators, thus obtaining a single scalar value that can represent the comprehensive disturbance level of the local area within the time window ΔT.

[0044] In practical engineering applications, the weights related to spatial distance can be determined not only using a Gaussian decay function, but also using other distance-related methods such as the inverse of distance. The weights are then normalized so that the sum of all weights within the verification unit is 1. For ease of understanding, this embodiment uses the following inverse distance normalization method to demonstrate the weight calculation.

[0045] For ease of understanding, let's take the verification unit as an example. For example, the set calculation time window for Within this time window, the system extracted data from three wireless accelerometers. The acceleration data sequence (second monitoring data); assuming a normal operating condition (e.g., a train passing in the distance), the acceleration signals collected by these three sensors in... Fluctuating within a range, of which Acceleration due to gravity (unit) At this point, the system calculates the root mean square value of each of the three data sequences to obtain their respective quantitative indicators: , , The system then bases its decisions on their relationship with... The weights for the distances (0.81m, 1.11m, and 1.26m respectively) are calculated as follows:

[0046]

[0047]

[0048] in Ultimately, the local environmental disturbance benchmark. The result is obtained through weighted average calculation: ; This value This means that in During this period of time, The expected normal vibration level at the sensor location.

[0049] S400, within the time window ΔT, calculate the first volatility index for the first monitoring data sequence; and standardize the first volatility index and the local environmental disturbance benchmark based on the historical baseline value of the first volatility index and the historical baseline value of the local environmental disturbance benchmark, respectively, to obtain the dimensionless first standardized index and the dimensionless benchmark standardized index. The first volatility index is used to characterize the degree of volatility of the first monitoring data sequence within the time window ΔT, and can reflect the severity of the fluctuation in the output data of the fiber optic sensor within this time window. In some embodiments, the first volatility index includes one or more of the following: standard deviation, root mean square value, peak-to-peak value, and energy index.

[0050] In this embodiment, the system extracts data from the fiber optic sensor within the same preset time window ΔT. The first monitoring data sequence (e.g., strain data sequence) is used to calculate the first volatility index. Preferably, the standard deviation of the first monitoring data sequence can be used as the first volatility indicator. .

[0051] Due to the first volatility indicator The dimensions (e.g., strain μɛ) may be related to the local environmental disturbance reference. Since the units (e.g., acceleration g) are different, directly performing ratio calculations may lack clear physical meaning. Therefore, in this embodiment, before calculating the abnormal deviation, the system introduces historical baseline values ​​to standardize the two respectively.

[0052] The historical baseline value is obtained by dynamically calculating and updating historical data through a sliding time window, the time span of which is a preset duration. In this embodiment, to adapt to the influence of slowly changing factors such as seasonal variations, passenger flow changes, and equipment status changes during the long-term service of the composite station structure, the system maintains historical baseline values ​​for both the first volatility index and the local environmental disturbance benchmark. For example, it maintains the historical average value of the first volatility index separately. and the historical average value of local environmental disturbance benchmarks It is continuously updated over time by using a sliding time window (e.g., the past 24 hours) to ensure that the standardized processing is adaptive.

[0053] Obtain the first volatility indicator within the current time window ΔT. and local environmental disturbance benchmarks Then, the system uses the corresponding historical baseline values ​​to... and Standardization is performed to obtain a dimensionless first standardized index and a dimensionless benchmark standardized index; the standardization process is used to eliminate dimensional differences, so that subsequent deviation calculations based on ratio relationships are comparable.

[0054] For ease of understanding, For example: Within the time window t = [10.0s, 11.0s], assume that the system obtains the historical average value of the first volatility index through long-term monitoring. με, and the historical average value of the local environmental disturbance benchmark. Under normal operating conditions, The strain data fluctuates smoothly around its mean, and the first volatility index within the current time window is calculated. με; simultaneously, the local environmental disturbance reference is obtained according to the aforementioned steps. The system is based on the corresponding historical baseline values ​​for each... and After standardization, the dimensionless first standardized index is obtained. Dimensionless benchmark standardized index .

[0055] S500, calculate the anomaly deviation based on the ratio between the first standardized index and the benchmark standardized index, and then... The health status of the fiber optic sensor is determined by comparing it with a preset health status threshold; wherein the health status includes at least one or more of healthy, abnormal, and faulty.

[0056] The system calculates the anomaly deviation based on the ratio between the first standardized index and the benchmark standardized index. This is used to characterize the degree of deviation of the fluctuation of the fiber optic sensor relative to the level of disturbance in its local environment. Preferably, the abnormal deviation can be determined in the following form: ; in, This represents the first volatility indicator after standardization. This represents the standardized local environmental disturbance reference. To prevent extremely small positive numbers with a denominator of zero (e.g., 10) 6 ).when When the value is close to 1, it indicates that the fluctuation of the fiber optic sensor matches the level of disturbance in its environment, and its behavior is normal; when... A value significantly greater than 1 indicates that the fluctuation of the fiber optic sensor far exceeds the level of environmental disturbance, and there may be a malfunction.

[0057] The system will determine the abnormal deviation. The health status of the fiber optic sensor is determined by comparing it with a preset health status threshold. This health status threshold typically includes at least two levels: an alarm threshold and a... With fault threshold Both can be calibrated experimentally or set empirically based on the on-site noise level, structural characteristics, and monitoring requirements. Diagnostic logic, for example: like ≤ If so, the health status is determined to be healthy; like < ≤ If so, the health status is determined to be abnormal; like > If the health status is determined to be faulty, the subsequent handling process will be triggered.

[0058] For ease of understanding, we will still use For example: Under the aforementioned normal operating conditions, the calculation yields... ≈1.034, this value is close to 1 and satisfies ≤ Therefore, the system determines The health status is healthy. For example, suppose at t=10.5s, Due to localized failure of the adhesive layer, the reading abruptly changed from -250 με to -500 με. This caused a sharp increase in the standard deviation of the strain data within the current time window, resulting in the calculated... με; however, the actual physical environment of the station has not changed, therefore the local environmental disturbance reference... Still After normalization, , The abnormal deviation at this time If this value is much greater than 1, it indicates that the behavior of the sensor is seriously mismatched with its environment, and so on.

[0059] In this embodiment, before comparing the abnormal deviation with a preset health status threshold, the method further includes: performing validity detection on the second monitoring data sequence of the wireless sensors in the verification unit within the time window ΔT; when the proportion of the number of wireless sensors that meet the preset invalid conditions to the K wireless sensors exceeds a preset proportion threshold, the verification unit is determined to be in an untrusted state, and the corresponding diagnostic results within the time window ΔT are marked as invalid or unverified.

[0060] The validity detection is used to avoid distortion of the local environmental reference caused by "abnormal wireless data," which could lead to false health judgments of the fiber optic sensor. Preset invalidation conditions may include: interruption of the second monitoring data, prolonged constant value, exceeding the measurement range, or abnormal data noise, etc.

[0061] For example, in calculating local environmental disturbance references Previously, the system first checked the verification unit. The system checks whether the data sequences of the K wireless sensors within the current time window ΔT are valid. When more than a preset proportion (e.g., 50%) of the wireless sensors experience data interruption or constant values ​​within the time window, the system determines that the verification unit is in an "untrusted" state and marks the corresponding diagnostic results within the time window ΔT as "invalid or unverified" to avoid erroneous diagnosis based on unreliable disturbance benchmarks.

[0062] Second embodiment: This embodiment provides a full-cycle health monitoring system for composite station structures based on a multimodal sensor network. The system is configured to acquire multimodal monitoring data, construct verification units, determine local environmental disturbance benchmarks, calculate anomaly deviations, and determine health status. It can diagnose the health status of fiber optic sensors throughout the entire operation cycle of the composite station structure. The system includes: The data acquisition module is used to acquire the first monitoring data sequence of the fiber optic sensor and the second monitoring data sequence of multiple wireless sensors within the same preset time window ΔT. In this embodiment, the data acquisition module is used to collect multi-source monitoring data from a multimodal sensor network deployed on the composite station structure. The fiber optic sensor can be a fiber grating (FBG) sensor array, and at least one FBG demodulator continuously collects data at a relatively high sampling frequency; for example, it collects center reflection wavelength data, which can be converted into a strain data sequence through temperature compensation and calibration coefficients, serving as the first monitoring data. The wireless sensor may include a wireless accelerometer, a wireless tilt sensor, etc., which collects structural dynamic response data (such as vibration, tilt, etc.) at a relatively low sampling frequency, serving as the second monitoring data, and can be aggregated to a central gateway via a wireless ad hoc network or a star network.

[0063] To ensure that the first and second monitoring data sequences fall within the same preset time window ΔT and are comparable, the data acquisition module can assign a unified high-precision timestamp to data from different sources. This timestamp can be generated by a hardware clock using Network Time Protocol (NTP) or Precision Time Protocol (PTP). For data streams with different sampling rates, the data acquisition module can also perform time alignment processing. For example, it can upsample / interpolate the low-frequency second monitoring data to map it to a time axis consistent with the high-frequency first monitoring data; or it can insert multiple new data points between two original wireless data points spaced 50 milliseconds apart, so that the upsampled second monitoring data matches the fiber optic data in terms of time resolution, thereby forming a multimodal dataset aligned within the same preset time window ΔT.

[0064] The verification unit construction module is used to calculate the spatial distance between the fiber optic sensor and the wireless sensor based on the spatial location coordinates, and select the K wireless sensors with the closest spatial distance from the wireless sensors that meet the preset search radius R, and form a verification unit with the corresponding fiber optic sensor. In this embodiment, the verification unit construction module is used to construct a corresponding verification unit for each fiber optic sensor based on the physical deployment location of the sensors. During initialization, the system can load pre-mapped sensor network topology information stored in a database. This topology information records the spatial coordinates (x, y, z) of each fiber optic sensor and wireless sensor in the three-dimensional coordinate system of the overlapping station structure. During system operation, for each fiber optic sensor... Based on the spatial coordinates, the spatial distance (e.g., Euclidean distance) between the sensor and each wireless sensor is calculated, and the K wireless sensors with the closest spatial distance are selected from the wireless sensors that satisfy the preset search radius R. Together they form a verification unit The verification unit construction process can be executed periodically or triggered by events to adapt to situations such as the addition or removal of wireless nodes or fine-tuning of their positions.

[0065] In some implementations, the preset search radius R is a preset positive value and can be selected based on the structural density of the area where the fiber optic sensor is located and / or the vibration propagation distance; K is a positive integer not less than the preset minimum number of neighbors to ensure the statistical robustness of wireless information fusion within the verification unit and reduce the impact of single-point anomalies.

[0066] The disturbance reference determination module is used to calculate the wireless quantization index and perform weighted fusion within the time window ΔT to obtain the local environmental disturbance reference. In this embodiment, the disturbance reference determination module is used to determine the disturbance reference based on the verification unit. The second monitoring data from the internal wireless sensor is used to construct a local environmental disturbance benchmark that can characterize the level of physical disturbance in the local area where the fiber optic sensor is located. Specifically, the disturbance reference determination module, within the same preset time window ΔT, targets the verification unit. For each wireless sensor, a second monitoring data sequence is used to calculate a wireless quantization index. The wireless quantization index is used to characterize the energy or fluctuation intensity of the second monitoring data sequence within the time window ΔT. It may include one or more of the root mean square value, standard deviation, peak-to-peak value, and energy index. Preferably, the root mean square value (RMS) may be used as the wireless quantization index.

[0067] Furthermore, the disturbance reference determination module performs weighted fusion of the wireless quantization indicators based on weights related to spatial distance to obtain a local environmental disturbance reference. The weights can be based on the central fiber optic sensor. The spatial distance between each wireless sensor within the verification unit is determined, ensuring that the sum of all weights within the verification unit is 1, and that the weights decrease as the spatial distance increases, thus assigning a weight to sensors closer to the verification unit. The wireless sensors receive higher weights. In some implementations, these weights can be determined using a Gaussian decay function, for example: ; in, Represents the verification unit The distance weight corresponding to the j-th wireless sensor in the middle, This indicates that the j-th wireless sensor and the central fiber optic sensor Spatial distance between them The distance scale parameter is used to control the weight decay rate; the weights can be further normalized to make the verification unit... The sum of all weights is 1. By weighting and fusing the K wireless quantization indicators, a local environmental disturbance benchmark that can represent the comprehensive disturbance level of the local area within the time window ΔT is obtained. .

[0068] To facilitate engineering implementation, the disturbance reference determination module can also adopt other weight determination methods related to distance, and normalize the weights to ensure that the sum of the weights is 1.

[0069] The deviation calculation module is used to calculate the first volatility index and perform standardization within the time window ΔT, and calculate the abnormal deviation based on the standardization result; In this embodiment, the deviation calculation module is used to calculate the local environmental disturbance reference based on the first monitoring data from the fiber optic sensor and the disturbance reference determination module output by the disturbance reference determination module. Calculate the abnormal deviation. Specifically, the deviation calculation module calculates the first volatility index for the first monitoring data sequence of the fiber optic sensor within the same preset time window ΔT. This is used to characterize the volatility of the first monitoring data sequence within the time window ΔT. The first volatility index may include one or more of the following: standard deviation, root mean square value, peak-to-peak value, and energy index; preferably, standard deviation may be used as the first volatility index. .

[0070] Due to the first volatility indicator The dimensions may differ from the local environmental disturbance reference. Since the dimensions are different, the deviation calculation module further introduces historical baseline values ​​to standardize both, thereby obtaining a dimensionless first standardized index and a dimensionless benchmark standardized index. The historical baseline values ​​are obtained by dynamically calculating and updating historical data through a sliding time window. The time span of the sliding time window is a preset duration (e.g., the past 24 hours) to adapt to seasonal changes or long-term trends in structural response. The deviation calculation module calculates the abnormal deviation based on the ratio between the standardized first standardized index and the benchmark standardized index. This characterizes the degree of deviation of the output fluctuation of the fiber optic sensor from the level of disturbance in its local environment.

[0071] The health diagnosis module is used to compare the abnormal deviation with a preset health status threshold to determine the health status of the fiber optic sensor. In this embodiment, the health diagnosis module is used to determine the degree of abnormal deviation. The health status of the fiber optic sensor is compared with a preset health status threshold to output the health status of the sensor. The health status threshold may include at least two levels: an alarm threshold and a... With fault threshold Diagnostic logic, for example: when ≤ When the health status is determined to be healthy; when < ≤ When the health status is determined to be abnormal; > If the health status is determined to be faulty, subsequent handling procedures can be triggered.

[0072] Data correction and fusion module: This module processes fiber optic sensor data based on the health status output by the health diagnosis module. For example, it directly transmits data that is determined to be healthy to the upper-layer application, marks data that is determined to be abnormal for subsequent manual review, and performs removal / deletion operations on data that is determined to be faulty. Optionally, the data correction and fusion module can also use wireless sensor data within the same verification unit to generate correction values ​​through interpolation or substitution models to ensure the continuity of the data stream.

[0073] Early warning and visualization module: used to display health diagnosis results and corrected and fused structural monitoring data in the form of charts or 3D models; when a fault state is identified, an early warning mechanism is triggered, and maintenance personnel can be notified via SMS, email or system pop-up, while indicating the fault sensor number and its physical location.

[0074] In this embodiment, the system is configured to execute the method described in the first embodiment; the above modules work together to enable the system to construct a verification unit based on multimodal data and determine the local environmental disturbance benchmark within the same preset time window ΔT, and then calculate the abnormal deviation degree and output the health status, so as to realize the full-cycle health monitoring of the fiber optic sensor.

[0075] Third embodiment: This embodiment provides a health monitoring device, including a processor and a memory. The memory stores a computer program, and when the computer program is run on the processor, it causes the processor to execute the method described in the first embodiment.

[0076] Fourth embodiment: This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the method described in the first embodiment.

[0077] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0078] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented, in whole or in part, as a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium such as a solid-state drive (SSD).

[0079] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for full-cycle health monitoring of a composite station structure based on a multi-modal sensor network, characterized in that, include: Within the same preset time window ΔT, a first monitoring data sequence from an optical fiber sensor and a second monitoring data sequence from multiple wireless sensors are acquired; wherein the optical fiber sensor and the wireless sensors have corresponding spatial coordinates. For each fiber optic sensor, the spatial distance between it and each wireless sensor is calculated based on the spatial location coordinates, and K wireless sensors with the closest spatial distance are selected from the wireless sensors that meet the preset search radius R; the fiber optic sensor and the selected K wireless sensors constitute the verification unit of the fiber optic sensor. Within the time window ΔT, a wireless quantization index is calculated for the second monitoring data sequence of each wireless sensor in the verification unit, and the wireless quantization index is weighted and fused based on the weight related to spatial distance to obtain the local environmental disturbance benchmark corresponding to the fiber optic sensor. Within the time window ΔT, a first volatility index is calculated for the first monitoring data sequence; and the first volatility index and the local environmental disturbance benchmark are standardized based on the historical baseline value of the first volatility index and the historical baseline value of the local environmental disturbance benchmark, respectively, to obtain a dimensionless first standardized index and a dimensionless benchmark standardized index. Anomaly deviation is calculated based on the ratio between the first standardized index and the benchmark standardized index. The anomaly deviation is then compared with a preset health status threshold to determine the health status of the fiber optic sensor. The health status includes at least one or more of the following: healthy, abnormal, and faulty.

2. The method according to claim 1, characterized in that: The preset search radius R is a preset positive value, and is selected based on the structural density and / or vibration propagation distance of the structural region where the fiber optic sensor is located; K is a positive integer not less than the preset minimum number of neighbors.

3. The method according to claim 1, characterized in that: The wireless quantization index is used to characterize the energy or fluctuation intensity of the second monitoring data sequence within the time window ΔT. The wireless quantization index includes one or more of the following: root mean square value, standard deviation, peak-to-peak value, and energy index.

4. The method according to claim 1, characterized in that: The weights related to spatial distance are determined based on the spatial distance between the fiber optic sensor and each wireless sensor, and the sum of the weights in the verification unit is 1; wherein, the weights decrease as the spatial distance increases.

5. The method according to claim 1, characterized in that: The first volatility index is used to characterize the degree of volatility of the first monitoring data sequence within the time window ΔT. The first volatility index includes one or more of the following: standard deviation, root mean square value, peak-to-peak value, and energy index.

6. The method according to claim 1, characterized in that: The historical baseline value is obtained by dynamically calculating and updating historical data through a sliding time window, and the time span of the sliding time window is a preset duration.

7. The method according to claim 1, characterized in that: Before comparing the abnormal deviation with a preset health status threshold, the method further includes: performing validity detection on the second monitoring data sequence of the wireless sensors in the verification unit within the time window ΔT; when the proportion of the number of wireless sensors that meet the preset invalid conditions to the K wireless sensors exceeds a preset proportion threshold, the verification unit is determined to be in an untrusted state, and the corresponding diagnostic results within the time window ΔT are marked as invalid or unverified.

8. A full-cycle health monitoring system for composite station structures based on multimodal sensor networks, characterized in that, include: The data acquisition module is used to acquire the first monitoring data sequence of the fiber optic sensor and the second monitoring data sequence of multiple wireless sensors within the same preset time window ΔT. The verification unit construction module is used to calculate the spatial distance between the fiber optic sensor and the wireless sensor based on the spatial location coordinates, and select the K wireless sensors with the closest spatial distance from the wireless sensors that meet the preset search radius R, and form a verification unit with the corresponding fiber optic sensor. The disturbance reference determination module is used to calculate the wireless quantization index and perform weighted fusion within the time window ΔT to obtain the local environmental disturbance reference. The deviation calculation module is used to calculate the first volatility index and perform standardization within the time window ΔT, and calculate the abnormal deviation based on the standardization result; The health diagnosis module is used to compare the abnormal deviation with a preset health status threshold to determine the health status of the fiber optic sensor. The system is configured to perform the method according to any one of claims 1 to 7.

9. A health monitoring device, characterized in that, The method includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is run on the processor, it causes the processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, is used to implement the method according to any one of claims 1 to 7.

Images