A multi-element comprehensive quantitative detection method for pipe gallery structure health state

Abstract

The present application belongs to the technical field of structural health monitoring, and relates to a kind of for pipe gallery structural health state Multivariate comprehensive quantitative detection method, this method: through the temperature data of pipe gallery target section and structural deformation data are time alignment;Calculation introduces the spatial temperature difference modulus of pipe gallery orthotropic heat conduction weight and temperature change rate, constructs heat load gradient vector sequence, to define thermal action period and establish analysis window;In window, with the gradient vector sequence of local topological denoising processing as independent variable, construct structure response hysteresis curve;The closed region of curve is integrated and normalized by adaptive horizontal and vertical range envelope rectangle, generate the thermal response hysteresis index of quantification characterization health state;Finally, compared with the elastic reference threshold, the degree of nonlinear damage is determined.The present application solves the problem that the existing conventional linear delay monitoring method cannot accurately identify the internal microscopic non-elastic physical friction energy consumption of pipe gallery.

Description

Technical Field

[0001] This invention belongs to the technical field of structural health monitoring and relates to a multi-dimensional comprehensive quantitative detection method for the structural health status of utility tunnels. Background Technology

[0002] As an important component of urban infrastructure, underground utility tunnels are in long-term service in complex underground environments. They are subject to the combined effects of multiple factors such as temperature changes, soil pressure, groundwater erosion, and internal pipeline loads. Over time, they may suffer cumulative damage such as material deterioration, microcrack propagation, or joint deformation. To ensure the safe and stable operation of utility tunnels, it is necessary to continuously monitor and evaluate their structural health status.

[0003] In existing technologies, structural health monitoring of utility tunnels typically relies on static or quasi-static measurement methods. This involves deploying sensors such as strain gauges or displacement gauges to collect structural deformation data and comparing the measured values ​​with design limits to determine structural safety. Regarding the handling of temperature effects, the industry practice generally treats temperature changes as interfering factors. Monitoring systems often employ temperature compensation algorithms or remove temperature-correlated components during data preprocessing to attempt to isolate the structural response caused by other loads. Alternatively, they may establish a linear regression model of structural strain and temperature to isolate temperature-related strain components. Furthermore, in the field of bridges and other large-span reinforced concrete structures, some existing technologies attempt to calculate the lag time of the structural response by finding the time difference between the temperature time series peak and the deformation time series peak, and use this lag time as a basis for evaluating the structural condition.

[0004] However, the aforementioned approach of treating temperature effects as interference and eliminating them has limitations. This method presupposes that the structure is always in a linear elastic response state under temperature, meaning that the loading and unloading paths coincide and there is no energy dissipation. For structures with existing internal micro-damage, the opening and closing of cracks under temperature reciprocation will generate frictional energy dissipation, causing its thermally induced deformation to exhibit nonlinear hysteresis characteristics. Traditional methods are unable to capture this early damage information; the approach of only calculating the kinematic constant of time (hysteresis time difference) has fundamental limitations. For structures with existing micro-crack damage, the friction at the downstream interface under thermal reciprocation not only manifests as time lag but also as an irreversible energy dissipation in the thermodynamic process. Relying solely on traditional hysteresis time analysis cannot quantify the energy dissipated due to friction from a physical perspective. At the same time, the uniform surrounding thermal environment of bridges is completely different from the orthotropic thermal boundary conditions of underground utility tunnels, which have sunlight shining from the top, constant temperature at the bottom, and heating from cables on the side walls. Simply applying single-dimensional temperature data hysteresis analysis to utility tunnels will lead to misjudgments. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a multivariate comprehensive quantitative detection method for the health status of utility tunnel structures.

[0006] A multivariate comprehensive quantitative detection method for the structural health status of utility tunnels includes the following steps: S1. Collect temperature data and structural deformation data of the target section of the pipe gallery, and use a time synchronization mechanism to align the temperature data and structural deformation data to construct the original monitoring dataset; S2. Extract multi-point temperature data from the original monitoring dataset, calculate the vertical and horizontal temperature differences of the target cross-section of the utility tunnel, and perform structural heat transfer weighting calculations to obtain the orthogonal anisotropic spatial temperature difference modulus, which characterizes the degree of uneven temperature distribution within the target cross-section of the utility tunnel; calculate the temperature change rate of the temperature data in the time dimension; fuse the orthogonal anisotropic spatial temperature difference modulus and the temperature change rate at the same moment to generate a heat load gradient vector sequence; and combine the temperature extreme points and the inflection points of the change rate in the sequence as a time-domain phase reference. S3. Divide the complete thermal action cycle using the time-domain phase reference, extract the structural response data segment corresponding to the thermal action cycle, and establish a thermal excitation-structural response analysis window. S4. In the thermal excitation-structural response analysis window, the temperature change rate component in the thermal load gradient vector sequence is used to identify the heating loading path and the cooling unloading path, and the corresponding temperature data is used as the horizontal axis and the structural response data is used as the vertical axis to construct the structural response hysteresis curve. S5. Perform local noise reduction on the structural response hysteresis curve, and then obtain the hysteresis loop area value that characterizes the inelastic friction energy dissipation of the crack through area integration; calculate the normalized reference area based on the range of the denoised structural response hysteresis curve in the horizontal and vertical directions; finally, calculate the ratio of the hysteresis loop area value to the normalized reference area to generate the dimensionless thermal response hysteresis index. S6. Compare the thermally induced hysteresis index with the elastic benchmark threshold obtained from historical data statistics based on the health status of the utility tunnel structure. Determine the degree of nonlinear damage to the utility tunnel structure based on the comparison results and output the health status assessment results.

[0007] A further aspect of the present invention, step S1, includes the following steps: Temperature sensor arrays and deformation sensor arrays are deployed on the same cross section of the target section of the utility tunnel, and the collected temperature data and structural deformation data are timestamped using a unified clock source. The temperature sensor array and deformation sensor array are activated synchronously according to the preset sampling frequency to acquire data frames with timestamps. The difference between the timestamps of the temperature data frame and the structural deformation data frame is calculated. Data with timestamp differences within the preset time alignment tolerance range are extracted and combined to generate the original monitoring dataset.

[0008] A further aspect of this invention involves obtaining the orthotropic spatial temperature difference modulus, characterizing the degree of temperature non-uniformity within the target cross-section of the pipe gallery, through a weighted calculation based on structural heat transfer weights. This includes the following steps: Based on the soil depth gradient of the utility tunnel and the heat dissipation power of the internal pipelines, a first structural heat transfer weight is assigned to the vertical temperature difference. Assigning heat transfer weight to the second structure based on the horizontal temperature difference. ; In this process, the reference thermal resistance constant and the reference heat flux density constant are normalized and introduced to set the first structural heat transfer weight. With the second structure heat transfer weight This is a dimensionless penalty coefficient; the specific setting rules are as follows: The value is the difference between the first thermal resistance from the top plate of the pipe gallery to the ground surface and the second thermal resistance from the bottom plate to the deep constant temperature soil layer, and the ratio of the reference thermal resistance constant. The value is the ratio of the heat flux density gradient from the asymmetric high-voltage cable side to the non-cable side inside the pipe gallery to the reference heat flux density constant.

[0009] A further aspect of the present invention, step S3, includes the following steps: Based on the time-domain phase reference, a time period containing at least one complete heating and cooling process is extracted from the heat load gradient vector sequence and defined as the thermal action period. Structural deformation data with timestamps falling within the thermal action period are retrieved from the original monitoring dataset to form structural response data segments; By using the thermal load gradient vector sequence within the thermal action cycle as the input excitation and the structural response data segment as the output response, a mapping relationship between the two in the time dimension is established, forming a thermal excitation-structural response analysis window.

[0010] A further aspect of the present invention, step S4, includes the following steps: Establish a two-dimensional coordinate system, with the average temperature value corresponding to the heat load gradient vector sequence within the thermal action cycle as the abscissa and the amplitude of the corresponding structural response data segment as the ordinate. Based on the positive and negative characteristics of the rate of temperature change in the heat load gradient vector sequence, the first response trajectory under the heating loading path and the second response trajectory under the cooling unloading path are depicted respectively. The structural response hysteresis curve is constructed by using the graph formed by the first response trajectory and the second response trajectory closing in a two-dimensional coordinate system.

[0011] A further aspect of this invention involves performing local noise reduction on the structural response hysteresis curve, and then obtaining the hysteresis loop area value characterizing the inelastic frictional energy dissipation of the crack through area integration calculation, including the following steps: A topological intersection detection algorithm is used to identify self-intersecting local disordered closed loops caused by high-frequency environmental noise in the structural response hysteresis curve, and these local disordered closed loops are removed from the main hysteresis curve. The identification of the self-intersecting locally disordered closed loop is specifically as follows: traverse the vertex set of the hysteresis curve in chronological order, and use the line segment intersection algorithm to determine whether there is an intersection point between non-adjacent line segments; if there is an intersection point, define the intersecting polygon region as a pseudo energy dissipation region triggered by high-frequency noise, and remove the area of ​​the pseudo energy dissipation region when performing polygon vertex coordinate integration to solve the area value of the hysteresis loop.

[0012] A further aspect of the present invention, step S6, includes the following steps: If the thermal response hysteresis index is less than or equal to the elastic benchmark threshold, the pipe gallery structure is determined to be in a linear elastic healthy state. If the thermal response hysteresis index is greater than the elastic benchmark threshold, it is determined that there is a nonlinear hysteresis effect inside the pipe gallery structure. When a nonlinear hysteresis effect is determined to exist, a health status assessment result containing nonlinear damage conclusions is output.

[0013] In a further embodiment of the present invention, the temperature sensor array consists of at least two platinum resistance thermometers covering the top plate, bottom plate and side wall of the target section of the pipe gallery, and the deformation sensor array consists of at least two vibrating wire strain gauges co-located with the platinum resistance thermometers; the preset time alignment tolerance is set based on the maximum allowable phase error of the system.

[0014] A further aspect of this invention is that the temperature change rate of the temperature data over time is specifically calculated by: calculating the first-order backward difference of the average temperature of all measuring points at adjacent sampling times.

[0015] In a further embodiment of the present invention, step S6 further includes the following steps: Obtain a set of nonlinear grade division thresholds based on a multiple relationship of elastic benchmark thresholds; The thermal response hysteresis index is compared with the nonlinear level classification threshold at multiple levels to determine the damage level of the pipe gallery structure. The degree of injury will be included in the health status assessment results.

[0016] In summary, the present invention has the following beneficial technical effects: 1. A data extraction method that reflects the unique physical characteristics of utility tunnels: This invention does not simply use average temperature changes, but introduces first and second structural heat transfer weighting coefficients that consider the thermal resistance of the soil covering the top and bottom of the utility tunnel and the asymmetry of internal cable heating, generating an orthotropic spatial temperature difference modulus. Compared to the comparison scheme in the bridge field that only uses simple air temperature, this invention can filter out complex underground environments and internal disturbances, and extract the characteristic thermal loads that truly affect the utility tunnel structure.

[0017] 2. A leap from kinematic time difference to thermodynamic energy dissipation: Compared to existing techniques that only calculate the lag time of temperature and deformation (single-time kinematic dimension), this invention proposes to calculate the energy dissipation by integrating the closed area after removing high-frequency pseudo-noise from its own crossover, and then performing dimensionless normalization under the rectangular envelope of the horizontal and vertical range differences. This method corresponds to the inelastic internal friction energy dissipation rate of the system's thermal cycle at the physical level, enabling quantitative measurement of energy dissipation from micro-damage.

[0018] 3. By using the topological intersection noise reduction method, the pseudo-energy dissipation zone caused by environmental noise and transient vibration of heavy-duty vehicles was objectively extracted and identified from the response hysteresis curve of the target structure, which improved the anti-interference ability of the algorithm in actual operation. The dimensionless normalization exponent reduced the numerical oscillation interference caused by the difference in cross-sectional size of the pipe gallery and the difference in absolute temperature difference of the foundation in different seasons, and realized the portability of the evaluation index across time and structure. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings are used to provide a further understanding of the present invention.

[0020] Figure 1 A flowchart illustrating an embodiment of this application is disclosed.

[0021] Figure 2 Structural schematic diagrams of embodiments of this application are disclosed.

[0022] Figure 3 A schematic diagram of the structural response hysteresis curve in an embodiment of this application is disclosed.

[0023] Figure 4 The principle of topological intersection noise reduction and a partial schematic diagram of the pseudo-energy dissipation region triggered by high-frequency noise in the embodiments of this application are disclosed. Detailed Implementation

[0024] The following is in conjunction with the appendix Figure 1 - Figure 4 A preferred description of the present invention is provided below.

[0025] See attached document Figure 1This invention proposes a multivariate comprehensive quantitative detection method for the health status of utility tunnel structures, comprising the following steps: S1. Collect temperature data and structural deformation data of the target section of the pipe gallery, and use a time synchronization mechanism to align the temperature data and structural deformation data to construct the original monitoring dataset; S2. Extract multi-point temperature data from the original monitoring dataset, calculate the vertical and horizontal temperature differences of the target cross-section of the utility tunnel, and perform structural heat transfer weighting calculations to obtain the orthogonal anisotropic spatial temperature difference modulus, which characterizes the degree of uneven temperature distribution within the target cross-section of the utility tunnel; calculate the temperature change rate of the temperature data in the time dimension; fuse the orthogonal anisotropic spatial temperature difference modulus and the temperature change rate at the same moment to generate a heat load gradient vector sequence; and combine the temperature extreme points and the inflection points of the change rate in the sequence as a time-domain phase reference. S3. Divide the complete thermal action cycle using the time-domain phase reference, extract the structural response data segment corresponding to the thermal action cycle, and establish a thermal excitation-structural response analysis window. S4. In the thermal excitation-structural response analysis window, the temperature change rate component in the thermal load gradient vector sequence is used to identify the heating loading path and the cooling unloading path, and the corresponding temperature data is used as the horizontal axis and the structural response data is used as the vertical axis to construct the structural response hysteresis curve. S5. Perform local noise reduction on the structural response hysteresis curve, and then obtain the hysteresis loop area value that characterizes the inelastic friction energy dissipation of the crack through area integration; calculate the normalized reference area based on the range of the denoised structural response hysteresis curve in the horizontal and vertical directions; finally, calculate the ratio of the hysteresis loop area value to the normalized reference area to generate the dimensionless thermal response hysteresis index. S6. Compare the thermally induced hysteresis index with the elastic benchmark threshold obtained from historical data statistics based on the health status of the utility tunnel structure. Determine the degree of nonlinear damage to the utility tunnel structure based on the comparison results and output the health status assessment results.

[0026] In one embodiment of the present invention, step S1 includes the following steps: Temperature sensor arrays and deformation sensor arrays are deployed on the same cross section of the target section of the utility tunnel, and the collected temperature data and structural deformation data are timestamped using a unified clock source. The temperature sensor array and deformation sensor array are activated synchronously according to a preset sampling frequency to obtain data frames with timestamps. The difference between the timestamps of the temperature data frames and the structural deformation data frames is calculated. Data with timestamp differences within a preset time alignment tolerance range are extracted and combined to generate the original monitoring dataset.

[0027] Specifically, the data processing server reads the sensor deployment information for the designated target cross-section of the utility tunnel from a pre-configured equipment asset list. It should be noted that the target cross-section refers to a representative structural cross-section selected along the longitudinal direction of the utility tunnel, identified by a unique segment code within the system. This information defines the spatial layout and unique device identifiers of a temperature sensor array and a deformation sensor array. The temperature sensor array consists of at least two PT100 platinum resistance thermometers, positioned to cover the top, bottom, and left and right sidewalls of the target cross-section, capturing temperature field data reflecting the temperature values ​​of the internal air and structural surfaces at different locations within the target cross-section at a specific time. The deformation sensor array consists of at least two vibrating wire strain gauges deployed alongside the thermometers, measuring structural deformation data—i.e., strain or displacement data—that spatially corresponds to the temperature field data acquisition points.

[0028] At startup, the data processing server and the distributed data acquisition units connected to each sensor periodically synchronize with the same GPS-synchronized upper-level time server via the Network Time Protocol (NTP) to establish a unified clock source. It should be understood that this time synchronization mechanism is a hierarchical time service architecture based on the Network Time Protocol (NTP), which ensures that the deviation between the clocks of all devices involved in data acquisition and processing and the international standard time (UTC) is controlled within milliseconds, thereby guaranteeing that the timestamps of subsequently acquired data have microsecond-level synchronization accuracy.

[0029] Subsequently, the data processing server broadcasts a data acquisition trigger command to all relevant data acquisition units according to a preset sampling frequency. The preset sampling frequency is set based on the temperature change rate of the environment in which the utility tunnel is located, typically within the range of 0.01 Hz to 1 Hz. The specific setting is based on the fact that concrete structures have significant thermal inertia and exhibit a lag in their temperature response. According to the heat conduction equation and the Nyquist sampling theorem, the sampling frequency must be greater than twice the structural thermal response cutoff frequency. For conventional utility tunnels with wall thicknesses of 30 cm to 50 cm, their thermal response time constant is usually on the order of minutes. Therefore, setting a frequency range of 0.01 Hz to 1 Hz not only allows for complete recording of the entire thermal effect caused by sunlight or sudden fires but also effectively avoids data redundancy and storage pressure due to oversampling. For example, in normal monitoring mode, the sampling frequency can be set to 0.1 Hz; however, when an abnormally high temperature change rate is detected, it can be adaptively adjusted to 1 Hz. Upon receiving the command, each data acquisition unit simultaneously excites its connected sensors, converting the acquired analog signals, such as resistance values ​​or frequency signals, into digital quantities. Each data acquisition unit encapsulates the digital quantity with the current high-precision timestamp and its own device identifier into a data frame, and publishes it to the designated data bus topic via the MQTT protocol.

[0030] The data alignment service subscribes to all temperature and deformation data topics related to the target cross-section of the utility tunnel and stores the received data frames into temporary, timestamp-sorted memory cache queues. The service continuously retrieves data from the head of both queues. When the absolute value of the difference between the timestamps of the temperature and deformation data frames is found to be less than the preset time alignment tolerance, the temperature and structural deformation data within these two data frames are extracted and combined into a multi-dimensional data vector containing a single timestamp and all sensor measurements at that moment. The time alignment tolerance is a threshold used to determine whether data from different sources belong to the same acquisition time. Its value is set based on the processing latency of the data acquisition unit and network transmission jitter, for example, it can be set from 5 ms to 20 ms. The setting logic of this tolerance value follows the principle of the system's maximum permissible phase error. If the ambient temperature inside the utility tunnel changes at a maximum rate of 0.1℃ / s, to ensure that the correlation analysis error between temperature and strain data is less than 0.1%, the time synchronization error needs to be controlled within 0.001℃ / (0.1℃ / s) = 10 ms. Therefore, for high-precision monitoring requirements, this embodiment preferably sets the time alignment tolerance to 10 ms. This multidimensional data vector is inserted as a record into the data table of the time-series database. As the acquisition process continues, this data table constitutes the original monitoring dataset. This dataset is a structured data table with time as the primary key. Each record contains a unique timestamp and complete data measured by all temperature and deformation sensors on the target cross-section of the pipe gallery at that timestamp.

[0031] For example, consider a specific engineering scenario used to create the initial monitoring dataset. First, the system administrator designates section DK50+200 of the pipe rack as the target section. Four temperature sensors are installed on this section, with identifiers as follows: , , , And four strain sensors in the same location, identified as , , , The preset sampling frequency is set to 0.1 Hz, meaning a sample is collected every 10 seconds. At system time 1640995200.000 seconds, the data processing server synchronizes its clock with all data acquisition units via the NTP protocol and then issues a data acquisition command. Upon receiving the command, each data acquisition unit acquires sensor readings almost simultaneously. For example, with the sensor... The connected acquisition unit recorded a reading of 28.5 degrees Celsius at 1640995200.002 seconds and compared it with the sensor identifier. Released after packaging. Simultaneously, with sensors... The connected acquisition unit recorded and published a reading of 55.2 microstrain at 1640995200.004 seconds. Subsequently, the data alignment service received all data frames from the eight sensors, all with timestamps falling between 1640995200.002 and 1640995200.008 seconds. Since the set time alignment tolerance was 10 ms, all these data frames were determined to belong to the same sampling time. These data were then integrated into a data record, which is {timestamp: 1640995200}. 28.5 26.1, 27.3 27.4 55.2, 48.9 51.5, :51.8}. Finally, this data record is written as a row to the database table, constituting the data point of the original monitoring dataset at that moment. This process is repeated every 10 seconds, thereby continuously generating the original monitoring dataset.

[0032] In one embodiment of the present invention, step S2 includes the following steps: Based on the soil depth gradient of the utility tunnel and the heat dissipation power of the internal pipelines, a first structural heat transfer weight is assigned to the vertical temperature difference. Assigning heat transfer weight to the second structure based on the horizontal temperature difference. ; In this process, the reference thermal resistance constant and the reference heat flux density constant are normalized and introduced to set the first structural heat transfer weight. With the second structure heat transfer weight This is a dimensionless penalty coefficient; the specific setting rules are as follows: The value is the difference between the first thermal resistance from the top plate of the pipe gallery to the ground surface and the second thermal resistance from the bottom plate to the deep constant temperature soil layer, and the ratio of the reference thermal resistance constant. The value is the ratio of the heat flux density gradient from the asymmetric high-voltage cable side to the non-cable side inside the pipe gallery to the reference heat flux density constant. This ensures the calculated... and For a purely numerical constant, the thermal modulus of orthotropic space is... The calculation formula is as follows: The dimensions after calculation are restored to the temperature dimension.

[0033] Specifically, this step is executed automatically by the data processing server. Its purpose is to transform the raw, discrete multi-point temperature measurement data into structured features that characterize the driving forces of thermal action and their dynamic changes. During execution, the data processing server first establishes a first-in-first-out (FIFO) data processing pipeline and extracts multi-point temperature data records from the raw monitoring dataset time-stamp by time. For each timestamp... The data processing server reads the temperature values ​​distributed across the top slab, bottom slab, and side walls of the target section of the utility tunnel from this record. , , , .

[0034] First, to quantify the spatial non-uniformity of the temperature field, it is not sufficient to rely solely on the simple root mean square. Considering the anisotropic thermodynamic environment of the underground utility tunnel structure—for example, the shallowly covered roof is significantly affected by sudden temperature changes, while the deeply buried floor has substantial thermal inertia, and a heating main cable may be laid on one side—a weighting coefficient is introduced. The data processing server pre-reads the first structural heat transfer weight, a dimensionless dimension, from the configuration dictionary. (For example, a value of 1.5 is used because the vertical temperature gradient is more sensitive to roof settlement) and the heat transfer weight of the second structure. (For example, a value of 0.8 is used to adjust for lateral ventilation and cable heat dissipation.) The server calculates the temperature difference along the vertical direction. Temperature difference along the horizontal direction After weighting and harmonizing, the thermodynamic modulus of orthotropic space in single scalar form is calculated, and the formula is: .because and The use of dimensionless coefficients ensures that the spatial temperature modulus results retain the original Celsius dimension. This weighting process enhances the model's ability to amplify and capture localized degradation features.

[0035] Secondly, in order to quantify the rate of change of the temperature field over time, the server calculates the arithmetic mean of the temperatures at all measuring points at the current time t. And retrieve the previous moment from the FIFO data processing pipeline. average temperature value The rate of temperature change was calculated using the first-order backward difference method. The rate of temperature change is also a scalar quantity; a positive value indicates heating, a negative value indicates cooling, and the absolute value represents the rate of heating or cooling.

[0036] Subsequently, the data processing server will calculate the spatial temperature difference modulus at the same time t. and rate of temperature change The two components of the two-dimensional vector are merged into the heat load gradient vector. As the data stream is continuously processed, a series of time-ordered heat load gradient vectors are stored sequentially, thus constructing a heat load gradient vector sequence. This sequence is a time series set composed of heat load gradient vectors arranged in time order over a complete monitoring period, which can fully describe the dynamic process of thermal action from both spatial non-uniformity and temporal rate of change dimensions.

[0037] Finally, the server analyzes the data within a preset period, such as 24 hours, because the temperature changes in shallow-buried pipe galleries affected by solar radiation exhibit a clear diurnal periodicity; the generated auxiliary data sequence, namely the average temperature time series... Time series of temperature change A peak detection algorithm is applied. The timestamps of extreme points are determined by searching for local maxima and minima in the average temperature time series. These extreme points mark the moments when the average temperature reaches its highest or lowest, signifying the end of the heating phase and the beginning of the cooling phase, or vice versa. Simultaneously, the timestamps of inflection points in the rate of change are determined by searching for local maxima and minima in the rate of change time series. These inflection points are the moments when the heating or cooling rate reaches its maximum, typically corresponding to the periods of most intense external heat input or dissipation. This identified set of specific timestamps is tagged and stored by the system as a time-domain phase reference to precisely define the start and end times of the heat loading and heat unloading phases.

[0038] Furthermore, to avoid misjudgments of extreme points, such as false peaks, caused by random noise from the sensor, the peak detection algorithm includes a minimum peak protrusion threshold. This threshold is set to 3-5 times the sensor's measurement accuracy. For example, if the measurement accuracy of the PT100 temperature sensor is ±0.1℃, the minimum peak protrusion threshold is set to 0.3℃ to 0.5℃. Only when the difference between a local maximum and its nearest local minimum exceeds this threshold is a valid temperature extreme point recognized, thus ensuring that the identified phase reference corresponds to the actual physical heating or cooling process.

[0039] For ease of understanding, the above calculation process follows the formula below: At any given moment Space temperature modulus Temperature change rate and heat load gradient vector The calculation follows the formula below.

[0040] Calculate the average temperature :

[0041] Calculate the rate of temperature change :

[0042] The formula for calculating the thermal modulus of orthotropic space is:

[0043] Constructing the heat load gradient vector :

[0044] in, , , , They are time points Temperature data of the top plate, bottom plate, left side wall, and right side wall obtained from the original monitoring dataset, in degrees Celsius. The time interval between adjacent data sampling points is determined by the preset sampling frequency set in S1, and the unit is seconds. , In the formula, At the current sampling time, This represents the vertical temperature difference of the target cross-section of the utility tunnel at the current moment. This represents the horizontal temperature difference at the target cross-section of the utility tunnel at the current moment. As the first structural heat transfer weight, This is the heat transfer weight of the second structure; for The arithmetic mean of the temperatures at all measuring points at any given time; for Rate of temperature change over time; for The thermal modulus of orthogonal anisotropic space at time; for The heat load gradient vector generated at each moment; As the first structural heat transfer weight, This is the heat transfer weight of the second structure; For example, following the example of step S1, the data processing server has obtained the time. Temperature data for 1,640,995,200 seconds { 28.5 26.1, 27.3 27.4, all in degrees Celsius. The previous sampling time was retrieved from the original monitoring dataset. It is 1640995190 seconds, that is The temperature data from seconds ago is { 28.4 26.0 27.2, :27.3}. First, calculate Average temperature at time Degrees Celsius, and calculate Average temperature at time Degrees Celsius. Next, calculate. Rate of temperature change at time Degrees Celsius per second. Then, calculations were performed. Spatial temperature modulus at time Degrees Celsius. Finally, generated... Heat load gradient vector at time t This calculation process is repeated continuously at each sampling point, forming a heat load gradient vector sequence. After 24 hours of continuous monitoring, the average temperature sequence within this time period is analyzed, identifying the highest daily temperature of 38.6 degrees Celsius at 14:35:10, which is recorded as one of the extreme points. Simultaneously, analysis of the temperature change rate sequence reveals the highest temperature rise rate at 10:15:20, at 0.005 degrees Celsius / second, which is recorded as one of the inflection points of the change rate. These identified timestamps collectively constitute a time-domain phase reference, serving as the basis for subsequent division of the thermal cycle.

[0045] In one embodiment of the present invention, step S3 includes the following steps: Based on the time-domain phase reference, a time period containing at least one complete heating and cooling process is extracted from the heat load gradient vector sequence and defined as the thermal action cycle. Structural deformation data with timestamps falling within the thermal action cycle are retrieved from the original monitoring dataset to form a structural response data segment. The heat load gradient vector sequence within the thermal action cycle is used as the input excitation, and the structural response data segment is used as the output response. A mapping relationship between the two in the time dimension is established to form a thermal excitation-structural response analysis window.

[0046] Specifically, this step is executed by the data processing server, aiming to integrate the disordered monitoring data stream into analytical units with clear physical meaning. First, the data processing server analyzes the periodicity of the heat load gradient vector sequence, based on the time-domain phase reference established in step S2. The server locates two consecutive phase points representing the same physical state by searching the time-domain phase reference set, such as two consecutive maxima of average temperature. The time span between these two maxima is automatically extracted by the system and defined as a complete thermal cycle. It should be noted that this cycle setting ensures the completeness and comparability of the analysis, guaranteeing that the selected time period necessarily includes a complete heating and cooling process, for example, from one daily maximum temperature to the next daily maximum temperature.

[0047] In practical applications, an effective thermal amplitude threshold is introduced to eliminate invalid or weak thermal fluctuation cycles. The temperature range within the selected cycle is calculated, i.e., the difference between the highest and lowest temperatures. If this range is less than the preset effective thermal amplitude threshold, for example, 3°C, the thermal excitation within that cycle is deemed insufficient to elicit an observable hysteresis response in the structure, and the cycle is automatically discarded without further analysis. This threshold is set based on the thermal expansion coefficient of the structural material and the signal-to-noise ratio of the sensor, ensuring that the analyzed data segment has a sufficient signal-to-noise ratio.

[0048] Next, the server uses the start and end timestamps of the determined thermal cycle to perform a time-range query on the time-series database storing the original monitoring dataset. This query filters out all data records whose timestamps fall within the thermal cycle from the original monitoring dataset, and specifically extracts the structural deformation data portion from these records, such as the reading sequence of all strain sensors. This extracted subset of structural deformation data, which perfectly matches the thermal cycle in time, constitutes the structural response data segment. This data segment can be the strain or displacement data of a key measuring point, or it can be a weighted average of data from multiple measuring points to represent the overall response of the cross-section.

[0049] Finally, to establish a direct correlation between input and output, the data processing server constructs a new data structure in memory: the thermal excitation-structural response analysis window. The server iterates through each sampling timestamp within the thermal cycle. At each moment, the thermal load gradient vector at that time is used as the input excitation data, and the structural response amplitude with the same timestamp extracted from the structural response data segment is used as the output response data. These two are stored as a data pair. By performing this pairing operation on all timestamps within the cycle, a structured dataset consisting of a large number of time-synchronized data pairs is ultimately formed. This dataset is the thermal excitation-structural response analysis window, whose internal elements are one-to-one corresponding data pairs. This establishes a clear and unique mapping relationship between thermal excitation and structural response in the time dimension, providing a direct data foundation for subsequent nonlinear behavior analysis.

[0050] For example, following the example of step S2, the data processing server has identified the first daily maximum temperature point as being located at The highest temperature of the day was recorded at 14:30:10 on the first afternoon, and monitoring continued, identifying the second highest temperature point of the day as being located at... The second, 14:30:10 the following afternoon. Based on this... to The time span between these points is defined as a complete thermal cycle. Subsequently, a query is sent to the database storing the original monitoring dataset. After executing the query, the database returns a time series containing 8640 data points; this series is the structural response data segment, containing data from the top plate strain sensors. All readings during the thermal irradiation cycle. Finally, begin constructing the thermal excitation-structural response analysis window. For example, in At this moment, the heat load gradient vector is retrieved from the results generated in step S2. Meanwhile, the strain value at that moment was retrieved from the structural response data segment that was just acquired. Micro-strain. Generate data pairs: {excitation: [2.941, 0.01], response: 55.2}. Repeat this pairing process for each sampling moment within the thermal irradiation cycle, ultimately generating a set containing 8640 such data pairs. This set is defined as the thermal excitation-structural response analysis window and is passed to the next step for processing.

[0051] In one embodiment of the present invention, step S4 includes the following steps: A two-dimensional coordinate system is established, with the average temperature value corresponding to the heat load gradient vector sequence within the thermal action cycle as the abscissa and the amplitude of the corresponding structural response data segment as the ordinate. Based on the positive and negative characteristics of the temperature change rate in the heat load gradient vector sequence, the first response trajectory under the heating loading path and the second response trajectory under the cooling unloading path are plotted respectively. The structural response hysteresis curve is constructed using the graph formed by the closure of the first and second response trajectories in the two-dimensional coordinate system.

[0052] Specifically, this step is executed by the data processing server, whose core task is to convert the time-series data within the thermal excitation-structural response analysis window into geometric figures characterizing the nonlinear behavior of the structure. First, the data processing server initializes a two-dimensional Cartesian coordinate system in memory. The horizontal axis of this coordinate system is defined as a measure of the absolute state of the apparent thermal load of the pipe gallery environment; to maintain the classical thermophysical accuracy corresponding to structural strain, the average temperature value at each moment within the thermal cycle is used as its value. The vertical axis of this coordinate system is defined as a measure of the structural response, i.e., the amplitude at the corresponding moment in the structural response data segment.

[0053] The server sequentially traverses each data pair in the thermal excitation-structural response analysis window. To avoid misjudgments due to lag caused by relying solely on the average temperature value, for any given data pair, the server extracts the temperature rate of change component of its included heat load gradient vector and obtains the average temperature value for that moment from the metadata associated with that vector. Based on the sign of the temperature rate of change component, the server assigns the data points to two different trajectory data sets.

[0054] When the temperature change rate component exceeds the preset positive zero drift threshold, it indicates that the structure is in the heating loading stage. The preset positive zero drift threshold and the corresponding negative zero drift threshold are small values ​​set to avoid misclassification of data points due to data noise when the temperature reaches an extreme point. Specifically, the positive zero drift threshold is set as follows: The negative zero drift threshold is set to .in The value is based on the standard deviation of the temperature data under steady-state conditions. 2 times, that is For example, if preliminary testing shows that the standard deviation of the background noise in the temperature data is 0.005℃ / s, then... The set value is 0.01℃ / s. This means that only when the rate of temperature change actually exceeds the background noise level, i.e., >0.01℃ / s, is the structure confirmed to have entered a clear heating loading path; similarly, when the rate of change is <-0.01℃ / s, the cooling unloading path is confirmed. Small fluctuations in data points between -0.01 and +0.01℃ are considered transitional states and are either discarded or returned to the previous state to ensure the smoothness of the hysteresis curve trajectory. During this stage, the data points formed by the average temperature value at that moment and the structural response amplitude (…) S) is stored in the first trajectory data set. The heating loading path corresponding to this set refers to a series of system states corresponding to the time intervals in which the temperature change rate component of the heat load gradient vector is positive within the thermal action cycle.

[0055] Conversely, when the temperature change rate component is less than the preset negative zero drift threshold, it indicates that the structure is in the cooling and unloading phase. The cooling and unloading path corresponding to this phase refers to a series of system states corresponding to the time intervals during which the temperature change rate component of the heat load gradient vector is negative within the thermal cycle. At this time, the corresponding data points are stored in the second trajectory data set.

[0056] After traversing all data points in the thermal excitation-structural response analysis window, the server connects all points in the first trajectory data set in chronological order to draw the first response trajectory. Similarly, it connects all points in the second trajectory data set in chronological order to draw the second response trajectory. Finally, the graph formed by the first and second response trajectories closing together in the coordinate plane is constructed as the structural response hysteresis curve and stored as an ordered vertex set as the output of this step. It should be understood that this structural response hysteresis curve is a closed curve formed by connecting the two trajectories end to end. For an ideal linear elastic structure, its loading and unloading paths should completely overlap, making it impossible to form a closed curve. Therefore, the existence of this hysteresis curve itself characterizes the nonlinearity and energy dissipation characteristics of the structure's response under thermal action.

[0057] For example, following the example from step S3, the data processing server has obtained a thermal excitation-structural response analysis window containing 8640 data pairs. The server establishes a coordinate system with average temperature as the horizontal axis (in degrees Celsius) and top plate strain as the vertical axis (in degrees Celsius). The server begins to iterate through the data. Assume that at some point in the morning... The data pair read out shows an average temperature of 30.0 degrees Celsius, a strain value of 65.4 microstrain, and a temperature change rate component in the corresponding heat load gradient vector of +0.004 degrees Celsius / second. Since this rate of change is greater than zero, the data point (30.0, 65.4) is assigned to the data set of the first response trajectory. Later, at some point in the afternoon… The average temperature is also 30.0 degrees Celsius, but the strain value is 69.8 microstrains, and the corresponding temperature change rate component is -0.003 degrees Celsius / second. Since this change rate is less than zero, the data point (30.0, 69.8) is assigned to the data set of the second response trajectory. After the server completes this classification operation on all 8640 data points, it connects the points in the first trajectory data set to form a loading curve extending from the lowest temperature point to the highest temperature point. Subsequently, it connects the points in the second trajectory data set to form an unloading curve returning from the highest temperature point to the lowest temperature point. Since the strain value of 69.8 microstrain during unloading is greater than the strain value of 65.4 microstrain during loading at the same temperature of 30.0 degrees Celsius, these two curves do not overlap, thus forming a clearly visible closed loop on the coordinate plane. This loop is the structural response hysteresis curve generated in this embodiment. See [link to relevant documentation]. Figure 3 .

[0058] In one embodiment of the present invention, step S5 includes the following steps: A topological intersection detection algorithm is used to identify self-intersecting local disordered closed loops caused by high-frequency environmental noise in the structural response hysteresis curve, and these local disordered closed loops are removed from the main hysteresis curve. The identification of the self-intersecting locally disordered closed loop is specifically as follows: traverse the vertex set of the hysteresis curve in chronological order, and use the line segment intersection algorithm to determine whether there is an intersection point between non-adjacent line segments; if there is an intersection point, define the intersecting polygon region as a pseudo energy dissipation region triggered by high-frequency noise, and remove the area of ​​the pseudo energy dissipation region when performing polygon vertex coordinate integration to solve the area value of the hysteresis loop.

[0059] Specifically, this step is initiated immediately by the data processing server after obtaining the structural response hysteresis curve, aiming to transform the geometry into a standardized numerical index. First, the data processing server takes the structural response hysteresis curve, stored as an ordered vertex set, as input and calls the built-in numerical integration module. Before calling the shoelace formula, the data processing server performs a topology denoising check on the trajectory formed by all sequentially arranged vertices. High-frequency environmental noise, such as sudden vibrations caused by heavy trucks passing over the road surface above the utility tunnel, often causes a slight foldback in the structural response at a certain moment, forming a self-intersecting figure-eight pseudo-energy-dissipating "small loop." The system uses a vector cross product algorithm to determine whether any non-adjacent line segments cross or intersect; if they intersect, the coordinate product within the segment is truncated and its small area is deducted. After this "surgical truncation and denoising," the "main hysteresis loop" is then processed by the built-in numerical integration module. This module uses the shoelace formula to calculate the area of ​​the closed polygon defined by the ordered vertex set to obtain the original value of the size of the closed region enclosed by the quantized hysteresis curve, i.e., the hysteresis loop area value. See [link to relevant documentation]. Figure 4 From a physical perspective, the area of ​​a hysteresis loop is proportional to the energy dissipated by the structure during a complete thermal cycle due to inelastic behaviors such as the opening and closing friction of internal microcracks or the slippage of connecting components. For an ideal linear elastic structure, this value, i.e., the area of ​​the hysteresis loop, theoretically approaches zero when thermal conduction hysteresis is ignored. However, in practical engineering, due to the thermal inertia of concrete, there is a natural phase difference between temperature conduction and strain response. This means that even a healthy linear elastic structure will exhibit a hysteresis loop with a certain basic area in its temperature-strain curve. Therefore, the criterion for judging structural damage is not simply whether the area is zero, but whether the area exceeds the normal range determined by thermal inertia. When cracks or cumulative damage occur inside the structure, i.e., when the energy dissipation mechanism is enhanced, the area of ​​the hysteresis loop will significantly increase and deviate from the initial state. This is the theoretical basis for setting the elastic reference threshold in the subsequent steps of this invention, accurately identifying nonlinear changes caused by structural damage by subtracting the basic influence of thermal hysteresis.

[0060] To eliminate the impact of differences in thermal amplitude and structural response magnitude under specific operating conditions, the server performs normalization on the area value. This process first iterates through all the vertex coordinates of the structural response hysteresis curve, determining the maximum and minimum values ​​of the abscissa (mean temperature) and the ordinate (structural response amplitude). The server uses these four boundary values ​​to calculate the area of ​​the smallest rectangular region that completely encloses the hysteresis curve, and uses this area as the normalization benchmark. By dividing the calculated hysteresis loop area by this normalized benchmark area, a dimensionless value between 0 and 1 is generated, formally defined as the thermally induced response hysteresis index. The technical advantage of this normalization method is that it transforms absolute energy dissipation values ​​into relative dissipation rates. Even if the absolute values ​​of the hysteresis loop areas produced by two pipe racks of different sizes or under different seasonal temperature differences are different, as long as their nonlinear damage levels are similar, their normalized indices should be close, thus achieving spatial portability and temporal comparability of the index.

[0061] Finally, the data processing server establishes a connection with the historical status database and associates the thermally induced response hysteresis index generated in this calculation with the end timestamp of the thermal cycle that generated the index and the unique identifier of the pipe gallery cross-section, forming a new status record. The server performs a database write operation, inserting the status record into the specified time-series data table, thereby providing continuous data support for subsequent structural health degradation trend analysis. It should be noted that the thermally induced response hysteresis index is the core indicator output by this method for quantifying the structural health status of the pipe gallery, and its value is positively correlated with the degree of nonlinear damage within the structure.

[0062] For ease of understanding, the formulas involved in the above calculations are as follows: If the structural response hysteresis curve consists of n vertices arranged in sequence... The hysteresis loop area value is then determined. The calculation formula is:

[0063] Among them, vertex Defined as relative to the starting point The same applies to ensure the curve is closed.

[0064] The reference area used for normalization The calculation formula is:

[0065] The final generated thermal response hysteresis index The calculation formula is:

[0066] In the formula: This represents the total number of vertices contained in the structural response hysteresis curve. Vertex index ( ); and The first The x-coordinate (average temperature value) and y-coordinate (structural response amplitude) of each vertex; to ensure curve closure calculation, the vertices in the formula... Default value and starting point same. and These represent the maximum and minimum temperatures across all vertex data. and These represent the maximum and minimum values ​​of the structural response amplitude across all vertex data.

[0067] For example, the data processing server obtained a structural response hysteresis curve consisting of 8640 vertices. First, the server applied the shoelace formula to all 8640 vertices, and through iterative calculation, obtained the area of ​​the hysteresis loop. Celsius and micro-strain. Next, the server iterates through all vertex data to perform normalization, determining the temperature range as follows: Celsius Celsius, strain range micro strain to Micro-strain. Based on this, the normalized reference area is calculated. Celsius and micro-strain. Subsequently, the thermally induced hysteresis index was calculated. Finally, the generated thermal hysteresis index of 0.0265, together with the end timestamp of this thermal cycle (1641040210 seconds) and the pipe gallery cross-section identifier DK50+200, were packaged into a record and inserted into the database table to complete the quantitative assessment of this health status and accumulate new data points for long-term trend analysis.

[0068] In one embodiment of the present invention, step S6 includes the following steps: If the thermal hysteresis index is less than or equal to the elastic baseline threshold, the pipe gallery structure is determined to be in a linear elastic healthy state; if the thermal hysteresis index is greater than the elastic baseline threshold, the pipe gallery structure is determined to have a nonlinear hysteresis effect; when a nonlinear hysteresis effect is determined to exist, a health status assessment result containing nonlinear damage conclusions is output.

[0069] Specifically, this step, as the decision-output stage of the entire detection method, is seamlessly executed by the data processing server after calculating the thermal hysteresis index. First, the data processing server retrieves the elastic reference threshold corresponding to the currently monitored target cross-section of the utility tunnel from its internal configuration file or dedicated structural characteristic database. This elastic reference threshold... The settings are not based on subjective experience, but rather derived through statistical analysis of specific benchmark calibration periods. The specific setup steps are as follows: Determine the reference period: Select the first 30 days after the completion and acceptance of the pipe gallery structure, or 30 consecutive thermal cycles after confirming that the structure is in a healthy state, as the reference period.

[0070] Baseline data acquisition: Calculate the thermal hysteresis index generated in each cycle within the baseline period to form a baseline sample set. .

[0071] Statistical distribution analysis: Calculate the arithmetic mean of the sample set. and standard deviation Due to measurement noise and the inherent micro-inelasticity of materials, the hysteresis index of healthy structures follows a normal distribution.

[0072] Threshold calculation: Based on statistical principles, a flexible benchmark threshold is set. This threshold represents the upper limit of the hysteresis index fluctuation of a healthy structure at a 99.7% confidence level.

[0073] For example, if the average hysteresis index of a certain cross section in the base period Standard deviation The elasticity benchmark threshold is then set to Any observation exceeding 0.011 is considered a statistically significant anomaly.

[0074] The server directly compares the value of the thermal hysteresis index generated in step S5 with the retrieved elastic reference threshold. If the comparison result shows that the thermal hysteresis index is less than or equal to the elastic reference threshold, the server determines that the current structural state is a linear elastic healthy state, indicating that no significant energy dissipation caused by damage was detected during the thermal action cycle.

[0075] If the comparison result shows that the thermally induced hysteresis index is greater than the elastic baseline threshold, the server determines that a nonlinear hysteresis effect caused by factors such as microcrack opening and closing friction or loosening of connections exists within the structure. In this case, the server will further compare this index sequentially with a set of preset nonlinear level thresholds used to classify the degree of damage to determine the specific level of damage. The nonlinear level thresholds are based on the elastic baseline threshold. The severity of the damage is quantified by using a multiple relationship to classify the damage: Level I, slight nonlinearity / early damage: when the exponential... satisfy At this time, microcracks are initiating in the structure, and it is recommended to increase the monitoring frequency.

[0076] Level II, moderate nonlinear / extended damage: when exponential satisfy At this point, the crack may have already expanded or developed into a macroscopic crack, and it is recommended to arrange an on-site inspection.

[0077] Level III, Severe Nonlinearity / Dangerous State: When the exponent... satisfy At this point, the structural stiffness has significantly degraded, posing a safety hazard. Immediate reinforcement measures are recommended.

[0078] This grading method based on benchmark multiples can adapt to the structural characteristics of different utility tunnels and avoid misjudgment of fixed values ​​in different scenarios.

[0079] The data processing server encapsulates the conclusions of this assessment, including the status description and damage level, along with key monitoring data such as timestamps, cross-sectional identifiers, and the thermal hysteresis index itself, into a structured health status assessment result. This result is the final output of this method, and its content should at least include monitoring location, assessment time, quantitative indicators, status conclusions, and recommended measures, aiming to provide maintenance personnel with clear, quantitative, and actionable decision support information. This result is generated in JSON or XML format and pushed to the upper-layer application platform via the system message queue for visualization on the monitoring center's user interface or to trigger corresponding early warning notification processes.

[0080] For example, the data processing server has calculated that the thermal hysteresis index of the DK50+200 section of the utility tunnel is 0.0265 at timestamp 1641040210. First, the server reads the preset elasticity benchmark threshold for this type of utility tunnel from its configuration library. Assuming this threshold is 0.011, calculated based on historical health data of this section, as mentioned above... The server then calculates the index 0.0265 and compares it with the threshold 0.011. Since 0.0265 is greater than 0.011, the server determines that the cross-section of the utility tunnel has exhibited a nonlinear hysteresis effect. Subsequently, the server retrieves the thresholds for classifying nonlinear damage levels, assuming that according to the aforementioned multiple rule: Level I interval is (0.011, 0.033], Level II interval is (0.033, 0.066], and Level III interval is >0.066. The server matches the index 0.0265 with these intervals and finds that it falls within the Level I interval (0.011 < 0.0265 ≤ 0.033).

[0081] Finally, the data processing server generates a health status assessment result in JSON format, for example: {"location_id": "DK50+200", "timestamp": 1641040210, "index_value": 0.0265, "status_description": "Nonlinear damage", "damage_level": "Level I: Slight nonlinearity", "suggestion": "Status noteworthy, recommended to be included in the next cycle inspection focus"}.

[0082] The assessment results were sent to the operation and maintenance monitoring platform. The health indicator light on the corresponding section of the platform's dashboard changed from green to yellow, and the detailed information of this assessment was recorded.

[0083] See appendix Figure 2 The present invention also proposes a multi-dimensional comprehensive quantitative detection system for the health status of utility tunnel structures, comprising the following modules: The data acquisition and alignment module is used to collect temperature data and structural deformation data of the target cross section of the utility tunnel, and to align the temperature data and structural deformation data using a time synchronization mechanism to construct the original monitoring dataset.

[0084] The heat load feature extraction module calculates the heat load gradient vector sequence based on the temperature data in the original monitoring dataset and identifies the time-domain phase reference used to define the heat loading and heat unloading stages.

[0085] The analysis window construction module uses a time-domain phase reference to divide the complete thermal action cycle, extracts the structural response data segments corresponding to the thermal action cycle, and establishes a thermal excitation-structural response analysis window.

[0086] The hysteresis behavior characterization module, within the thermal excitation-structural response analysis window, uses the temperature change rate component in the thermal load gradient vector sequence to identify the heating loading path and the cooling unloading path, and constructs a structural response hysteresis curve with the corresponding temperature data as the abscissa and the structural response data as the ordinate.

[0087] The state index quantification module calculates the closed region of the structural response hysteresis curve and generates a thermal response hysteresis index that quantitatively characterizes the health status of the utility tunnel structure.

[0088] The health status assessment module compares the thermal response hysteresis index with the elastic benchmark threshold derived from historical data on the health status of the utility tunnel structure. Based on the comparison results, it determines the degree of nonlinear damage to the utility tunnel structure and outputs the health status assessment results.

[0089] Each of the modules can be implemented in whole or in part through software, hardware, or a combination thereof. It supports hardware embedded in or independent of the processor in the computer device, and also supports software stored in the memory of the computer device, so that the processor can call and execute the operations corresponding to each of the above modules.

[0090] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A multivariate comprehensive quantitative detection method for the structural health status of utility tunnels, characterized in that, Includes the following steps: S1. Collect temperature data and structural deformation data of the target section of the pipe gallery, and use a time synchronization mechanism to align the temperature data and structural deformation data to construct the original monitoring dataset; S2. Extract multi-point temperature data from the original monitoring dataset, calculate the vertical and horizontal temperature differences of the target section of the pipe gallery, and obtain the orthogonal anisotropic spatial temperature difference modulus that characterizes the degree of uneven temperature distribution within the target section of the pipe gallery after weighted calculation by structural heat transfer weight. Calculate the rate of temperature change of temperature data in the time dimension; fuse the temperature difference modulus of the orthogonal space at the same moment with the rate of temperature change to generate a heat load gradient vector sequence; and combine the temperature extreme points in the average temperature time series with the rate of change in the time series as a time-domain phase reference. S3. Divide the complete thermal action cycle using the time-domain phase reference, extract the structural response data segment corresponding to the thermal action cycle, and establish a thermal excitation-structural response analysis window. S4. In the thermal excitation-structural response analysis window, the temperature change rate component in the thermal load gradient vector sequence is used to identify the heating loading path and the cooling unloading path, and the corresponding temperature data is used as the horizontal axis and the structural response data is used as the vertical axis to construct the structural response hysteresis curve. S5. Perform local noise reduction on the structural response hysteresis curve, and then obtain the hysteresis loop area value that characterizes the inelastic friction energy dissipation of the crack through area integration; calculate the normalized reference area based on the range of the denoised structural response hysteresis curve in the horizontal and vertical directions; finally, calculate the ratio of the hysteresis loop area value to the normalized reference area to generate the dimensionless thermal response hysteresis index. S6. Compare the thermally induced hysteresis index with the elastic benchmark threshold obtained from historical data statistics based on the health status of the utility tunnel structure. Determine the degree of nonlinear damage to the utility tunnel structure based on the comparison results and output the health status assessment results.

2. The multivariate comprehensive quantitative detection method for the health status of utility tunnel structures according to claim 1, characterized in that, The process of constructing the original monitoring dataset includes the following steps: Temperature sensor arrays and deformation sensor arrays are deployed on the same cross section of the target section of the utility tunnel, and the collected temperature data and structural deformation data are timestamped using a unified clock source. The temperature sensor array and deformation sensor array are activated synchronously according to the preset sampling frequency to acquire data frames with timestamps. The difference between the timestamps of the temperature data frame and the structural deformation data frame is calculated. Data with timestamp differences within the preset time alignment tolerance range are extracted and combined to generate the original monitoring dataset.

3. The multivariate comprehensive quantitative detection method for the health status of utility tunnel structures according to claim 1, characterized in that, After structural heat transfer weighting calculation, the orthotropic spatial temperature difference modulus, which characterizes the degree of temperature non-uniformity within the target cross-section of the pipe gallery, is obtained, including the following steps: Based on the soil depth gradient of the utility tunnel and the heat dissipation power of the internal pipelines, a first structural heat transfer weight is assigned to the vertical temperature difference. Assigning heat transfer weight to the second structure based on the horizontal temperature difference. ; In this process, the reference thermal resistance constant and the reference heat flux density constant are normalized and introduced to set the first structural heat transfer weight. With the second structure heat transfer weight This is a dimensionless penalty coefficient; the specific setting rules are as follows: The value is the difference between the first thermal resistance from the top plate of the pipe gallery to the ground surface and the second thermal resistance from the bottom plate to the deep constant temperature soil layer, and the ratio of the reference thermal resistance constant. The value is the ratio of the heat flux density gradient from the asymmetric high-voltage cable side to the non-cable side inside the pipe gallery to the reference heat flux density constant.

4. The multivariate comprehensive quantitative detection method for the structural health status of utility tunnels according to claim 1, characterized in that, Step S3 includes the following steps: Based on the time-domain phase reference, a time period containing at least one complete heating and cooling process is extracted from the heat load gradient vector sequence and defined as the thermal action period. Structural deformation data with timestamps falling within the thermal action period are retrieved from the original monitoring dataset to form structural response data segments; By using the thermal load gradient vector sequence within the thermal action cycle as the input excitation and the structural response data segment as the output response, a mapping relationship between the two in the time dimension is established, forming a thermal excitation-structural response analysis window.

5. The multivariate comprehensive quantitative detection method for the health status of utility tunnel structures according to claim 1, characterized in that, Step S4 includes the following steps: Establish a two-dimensional coordinate system, with the average temperature value corresponding to the heat load gradient vector sequence within the thermal action cycle as the abscissa and the amplitude of the corresponding structural response data segment as the ordinate. Based on the positive and negative characteristics of the rate of temperature change in the heat load gradient vector sequence, the first response trajectory under the heating loading path and the second response trajectory under the cooling unloading path are depicted respectively. The structural response hysteresis curve is constructed by using the graph formed by the first response trajectory and the second response trajectory closing in a two-dimensional coordinate system.

6. The multivariate comprehensive quantitative detection method for the structural health status of utility tunnels according to claim 1, characterized in that, The structural response hysteresis curve is locally denoised, and then the hysteresis loop area value, which characterizes the inelastic frictional energy dissipation of the crack, is obtained by area integration. The steps include: A topological intersection detection algorithm is used to identify self-intersecting local disordered closed loops caused by high-frequency environmental noise in the structural response hysteresis curve, and these local disordered closed loops are removed from the main hysteresis curve. The identification of the self-intersecting locally disordered closed loop is specifically as follows: traverse the vertex set of the hysteresis curve in chronological order, and use the line segment intersection algorithm to determine whether there is an intersection point between non-adjacent line segments; if there is an intersection point, define the intersecting polygon region as a pseudo energy dissipation region triggered by high-frequency noise, and remove the area of ​​the pseudo energy dissipation region when performing polygon vertex coordinate integration to solve the area value of the hysteresis loop.

7. The multivariate comprehensive quantitative detection method for the structural health status of utility tunnels according to claim 1, characterized in that, Step S6 includes the following steps: If the thermal response hysteresis index is less than or equal to the elastic benchmark threshold, the pipe gallery structure is determined to be in a linear elastic healthy state. If the thermal response hysteresis index is greater than the elastic benchmark threshold, it is determined that there is a nonlinear hysteresis effect inside the pipe gallery structure. When a nonlinear hysteresis effect is determined to exist, a health status assessment result containing nonlinear damage conclusions is output.

8. The multivariate comprehensive quantitative detection method for the structural health status of utility tunnels according to claim 2, characterized in that, The temperature sensor array consists of at least two platinum resistance thermometers covering the top plate, bottom plate and side wall of the target section of the pipe gallery, and the deformation sensor array consists of at least two vibrating wire strain gauges that are co-located with the platinum resistance thermometers. The preset time alignment tolerance is based on the system's maximum permissible phase error setting.

9. The multivariate comprehensive quantitative detection method for the health status of utility tunnel structures according to claim 1, characterized in that, The calculation of the rate of temperature change over time is specifically performed by calculating the first backward difference of the average temperature of all measuring points at adjacent sampling times.

10. A multivariate comprehensive quantitative detection method for the structural health status of utility tunnels according to claim 7, characterized in that, Step S6 also includes the following steps: Obtain a set of nonlinear grade division thresholds based on a multiple relationship of elastic benchmark thresholds; The thermal response hysteresis index is compared with the nonlinear level classification threshold at multiple levels to determine the damage level of the pipe gallery structure. The degree of injury will be included in the health status assessment results.

Images