Storage yard concrete surface layer monitoring system

Abstract

The invention relates to the field of civil engineering structure health monitoring, and provides a storage yard concrete surface layer monitoring system, which comprises a sensor array, a data acquisition module, a LoRa wireless transmission module and a data analysis platform, and is characterized in that fiber grating strain sensors are buried in the middle of a surface layer, and temperature sensors are arranged 2-4cm below the surface layer and are arranged at intervals of 4-5m along a longitudinal joint; the data is transmitted to the server every 30 minutes through the LoRa; the platform sets a strain threshold value of 400 microstrain and a 24-hour temperature change threshold value of 20 DEG C, and an alarm is triggered through threshold value comparison and model separation strain components. The system optimizes included angle arrangement of sensors to suppress joint displacement noise, transmission reliability is improved in combination with adaptive power adjustment and time slot allocation, long-term monitoring of stress and temperature anomaly is achieved, data support is provided for damage early warning and maintenance, and the risk of structural damage is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of civil engineering structure health monitoring, and more particularly, relates to a storage yard concrete surface layer monitoring system. Background Art

[0002] Yard concrete pavements are widely used in logistics warehouses, port terminals, and other scenarios, primarily bearing the concentrated static loads generated by cargo stacking. For example, a single container in a port container yard can weigh tens of tons and remain static for extended periods, resulting in a significantly different stress distribution within the concrete than that observed on ordinary pavement. These loads are highly persistent and act in a fixed location, easily inducing surface fatigue and leading to localized stress accumulation and cracking risks. However, existing monitoring technologies present the following challenges: 1. Traditional monitoring methods (such as manual inspections or discrete sensors) struggle to capture the gradual strain changes caused by long-term static loads. For example, when containers are stacked for months, tiny cracks may develop in the concrete surface due to creep. However, existing technologies lack the ability to continuously collect data and cannot provide early warning of crack growth.

[0003] 2. Cracks in the concrete surface can directly impact the safety of cargo stacking. For example, if precision instruments or large equipment tilt due to uneven ground, this can cause damage or even casualties. The toppling of high-value goods in a warehouse can lead to significant economic losses. Existing technologies, due to delayed alarms and high false alarm rates, struggle to provide timely support for maintenance decisions.

[0004] The root cause of the above problems is that the long-term and concentrated nature of static loads requires the sensor network to have high sensitivity and spatial resolution. In addition, the coupling effect between concrete creep characteristics and temperature strain further increases the difficulty of data decoupling, resulting in insufficient reliability of damage warning. Summary of the Invention

[0005] An object of the present invention is to address at least the above-mentioned drawbacks and to provide at least the advantages which will be described hereinafter.

[0006] One purpose of the present invention to provide a yard concrete surface layer monitoring system is to achieve real-time monitoring of the health status of the concrete surface layer through a specifically arranged sensor array, periodic data transmission and dynamic threshold judgment, provide data support for structural maintenance, and reduce the risk of structural damage caused by strain accumulation or sudden temperature changes.

[0007] The present invention provides a yard concrete surface layer monitoring system, comprising: Multiple sensor arrays are embedded within the concrete surface layer, each sensor array consisting of a fiber Bragg grating strain sensor and a temperature sensor. Each sensor array is spaced apart along the longitudinal joint direction of the concrete surface layer, and the fiber Bragg grating strain sensor is embedded in the middle of the thickness direction of the concrete surface layer, and the temperature sensor is embedded within a depth of 2-4 centimeters below the surface of the concrete surface layer. A data acquisition module connected to the sensor array, used to acquire real-time strain data of the fiber Bragg grating strain sensor and real-time temperature data of the temperature sensor; A wireless transmission module that communicates with the data acquisition module, the wireless transmission module adopts the LoRa communication protocol and transmits the real-time strain data and real-time temperature data to a remote server at a preset time (e.g., 30 minutes); A data analysis platform deployed on the remote server is configured with a threshold comparison algorithm. The threshold comparison algorithm sets the strain threshold to 400 microstrain and the temperature change threshold to 20 degrees Celsius. When the real-time strain data exceeds the strain threshold, or the change amplitude of the real-time temperature data within 24 hours exceeds the temperature change threshold, the data analysis platform generates an alarm signal and marks the corresponding sensor array position.

[0008] This invention utilizes a specially arranged sensor array, periodic data transmission, and dynamic threshold determination to achieve real-time monitoring of the health of the concrete surface layer, providing data support for structural maintenance and reducing the risk of structural damage caused by strain accumulation or sudden temperature changes. This addresses the problem of existing concrete surfaces in storage yards lacking long-term, real-time monitoring of internal strain and temperature changes, making it difficult to promptly identify concrete cracking risks and abnormal temperature stresses.

[0009] Preferably, the fiber Bragg grating strain sensor is buried within the range of ±10% of 1 / 2 of the thickness H of the concrete surface layer (i.e., the middle position in the thickness direction of the concrete surface layer); each group of sensor arrays is arranged at intervals of 4-5 meters along the longitudinal joint direction, and the distance between two sensors in the group is 0.1-0.6 meters. The connection direction of the fiber Bragg grating strain sensor and the temperature sensor forms an angle of 45° to 90° with the longitudinal joint direction, and the burial axis of the two (i.e., the sensor length direction) is consistent with the connection direction.

[0010] This invention uses fiber Bragg grating (FBG) strain sensors embedded in the middle of the concrete surface layer to reduce baseline interference with strain data caused by bending deformation. A 2-3 cm deep temperature sensor is also used to capture critical surface thermal stress mutation points caused by diurnal temperature differences, improving sensitivity. Furthermore, the sensor's angled placement suppresses strain measurement noise caused by joint displacement. This addresses the coupling error between strain and temperature data in existing monitoring systems, which occurs when the sensor's embedded position does not align with the temperature mutation zone between the middle and surface of the concrete surface layer and fails to eliminate interference from joint displacement on the measurement direction.

[0011] The physical mechanism by which sensor angle placement suppresses joint displacement noise is as follows: The expansion and contraction (e.g., thermal expansion and contraction) of a longitudinal joint primarily manifests as horizontal movement along the length of the joint. When the sensor line forms an angle with the joint (e.g., 60°), the joint displacement decomposes into a component along the line and a component perpendicular to the line. Fiber Bragg grating strain sensors are only sensitive to strain along their axis (i.e., the line), while the perpendicular component has no effect. The magnitude of the joint displacement component along the line is: ΔL·cosθ (ΔL is the total joint displacement, θ is the angle). When θ ≥ 45°, cosθ ≤ 0.707, indicating that the effect of joint displacement on the sensor is weakened. When θ = 90° (the line is perpendicular to the joint), cosθ = 0, indicating that the joint displacement has no direct effect on the sensor data.

[0012] Finite element analysis of the temperature distribution from the concrete surface to a depth of 5 cm revealed that the temperature fluctuation at a depth of 0-2 cm from the surface is 90%-95% of environmental fluctuations, while the fluctuation at a depth of 2-4 cm is 80%-85%, filtering out transient wind chill and sunlight interference. At a depth of 4 cm, the fluctuation is less than 50%, with significant response lag. This indicates that the 2-4 cm depth balances surface thermal stress sensitivity and interference resistance, making it the optimal burial range and enhancing sensitivity.

[0013] Preferably, the data analysis platform is further configured with a strain and temperature vectorized correlation model, the construction method of which includes the following steps: Collect the real-time strain data of the fiber Bragg grating strain sensor and the real-time temperature data of the temperature sensor in the same sensor array, and store them synchronously to the remote server at a fixed period; Based on the thermal expansion characteristics and elastic parameters of concrete, the theoretical strain component caused by pure temperature is calculated through temperature data; Decomposing the measured strain data into a component parallel to the sensor connection direction and a component perpendicular to the connection direction according to the angular position relationship between the fiber Bragg grating strain sensor and the temperature sensor; A correlation equation is established to relate the parallel components to a linear combination of the temperature strain component and the load strain component, where the load strain transfer coefficient is dynamically calibrated according to the sensor angle direction and the thickness of the concrete surface layer; The strain data of the vertical component are processed by time window sliding average to extract the baseline value of the load strain, and the real-time load strain change is calculated by the difference between the parallel component and the temperature strain component.

[0014] This invention leverages thermodynamic theory to isolate the influence of temperature on strain, avoiding false alarms caused by ambient temperature fluctuations. It also utilizes spatial directional decomposition to highlight the mechanical characteristics of load strain, suppressing the impact of joint displacement noise on the data. It also dynamically calibrates the transfer coefficient to adapt to the actual needs of different engineering parameters, improving the model's applicability. This addresses the problem that existing monitoring systems fail to physically separate temperature strain from load strain, resulting in alarm signals unable to distinguish between structural damage and normal thermal expansion and contraction, reducing the relevance of maintenance decisions.

[0015] Preferably, when the data analysis platform generates an alarm signal, it converts the real-time load strain variation Δε output by the strain and temperature vector correlation model into L (t) as input data for the threshold comparison algorithm; When Δε L (t) when the strain threshold of 400 microstrain is exceeded, it is determined to be a load abnormality and an alarm signal is generated; When the real-time temperature data changes by more than 20 degrees Celsius within 24 hours, and Δε L (t) When the Pearson correlation coefficient with the temperature variation amplitude meets one of the following conditions, it is determined to be a temperature stress anomaly and an alarm signal is generated: (1) |r|>0.7 and the signs are consistent in two consecutive sliding time windows; (2) The temperature variation amplitude ΔT>25℃ and |r|>0.7 within one sliding time window.

[0016] This invention triggers alarms based on load-strain data separated by a vectorized model, avoiding false alarms caused by temperature fluctuations. It also incorporates a temperature-load correlation coefficient to identify the synergistic effects of sudden temperature changes and load-strain (such as the superimposed stresses caused by heavy load operations during cold waves). It is compatible with existing threshold parameters (400με, 20°C), eliminating the need for system recalibration and maintaining manageable implementation costs. This addresses the problem that existing threshold alarms rely directly on raw strain data without stripping out the temperature-strain component, leading to thermal expansion during hot seasons being easily misinterpreted as structural damage and failing to identify the coupled effects of temperature and load-strain.

[0017] Add a data preprocessing module to the data analysis platform and convert Δε L (t) is synchronized with the temperature data to input the threshold comparison algorithm, without adding new hardware, through Δε L (t)>400με directly reflects the risk of structural damage; temperature changes exceeding 20℃ and with Δε L(t) is strongly correlated (|r|>0.7), suggesting that sudden temperature change induces additional stress.

[0018] The calculation of the correlation coefficient is achieved by the following steps: 1. Set a 24-hour sliding time window in the data analysis platform and extract the temperature change amplitude ΔT(t) and real-time load strain change Δε within the window L (t) time series data, where: ΔT(t) represents the difference between the real-time temperature data of the temperature sensor at time point t and the temperature data 24 hours ago; Δε L (t) represents the real-time load strain change output by the strain and temperature vectorization correlation model; The sliding time window is defined as 48 data points collected periodically within a continuous 24-hour period; 2. Calculate ΔT(t) and Δε based on the Pearson correlation coefficient formula L The linear correlation of (t) is calculated as follows: in: n is the total number of data points in the sliding time window, n=48; ΔT i is the temperature change amplitude corresponding to the i-th data point; Δε Li is the real-time load strain change corresponding to the i-th data point; For all ΔT in the window i The arithmetic mean of For all Δε in the window Li The arithmetic mean of 3. When the following conditions are met, it is determined to be abnormal temperature stress: |r|>0.7 (indicating strong linear correlation); If the correlation coefficients in two consecutive sliding time windows have the same sign (both positive or both negative), or if the temperature change amplitude ΔT exceeds 25°C, only one window needs to meet |r|>0.7 to trigger an alarm.

[0019] Preferably, the method for calculating the theoretical strain component caused purely by temperature comprises the following steps: Obtain the thermal expansion coefficient α and elastic modulus E of concrete; Calculate the difference ΔT(t) between the current temperature T(t) and the reference temperature T0 based on the real-time temperature data collected by the temperature sensor, where the reference temperature T0 is the initial curing temperature of the concrete surface layer when pouring is completed; Based on the thermal expansion strain formula ε T(t)=α·ΔT(t), calculate the theoretical strain component ε caused by pure temperature T (t); According to the elastic modulus E, the theoretical strain component ε T (t) is converted into the temperature stress component σ T (t) = E·ε T (t), and σ T (t) inputting the strain and temperature vector correlation model; The temperature stress component σ in the measured strain data is converted into T (t) and load strain component σ L (t) separation, and obtain the real-time load strain change Δε L (t); The thermal expansion coefficient α ranges from 8×10 -6 / ℃ to 12×10 -6 / ℃, the elastic modulus E ranges from 25GPa to 35GPa; The present invention solves the problem that existing monitoring systems fail to quantify temperature-strain components based on concrete parameters, resulting in the inability to physically separate temperature and load strain. By comparing theoretical calculations with measured data, the invention eliminates the interference of temperature fluctuations on the strain threshold and improves the accuracy of the alarm signal.

[0020] Preferably, the method of decomposing the measured strain data into a component parallel to the sensor connection direction and a component perpendicular to the connection direction comprises the following steps: Determine the plane rectangular coordinate system based on the angle θ formed by the connection direction of each group of fiber Bragg grating strain sensors and temperature sensors and the longitudinal joint direction; take the connection direction as the first coordinate axis, and take the direction perpendicular to the connection direction and located in the plane of the concrete surface layer as the second coordinate axis; The original strain data ε0(t) collected by each fiber Bragg grating strain sensor is decomposed according to the following formula: Strain component parallel to the connecting line: ε x (t)=ε0(t)·cosθ, the strain component perpendicular to the connecting line direction: ε y (t)=ε0(t·sinθ; x (t) as a component of load strain, ε y (t) is used as the reference component of the seam displacement noise.

[0021] The decomposition method of the present invention solves the problem that the existing technology does not establish a strain vector coordinate system based on the angular relationship between the sensor connection direction and the joint, resulting in inaccurate calculation directions of the parallel and vertical components. Through coordinated processing of coordinate projection and sensor array, the strain interference of joint displacement in non-sensitive directions is eliminated, thereby improving the separation accuracy of load-strain components.

[0022] Parallel component (ε x ) physical meaning: the strain component parallel to the sensor line direction (ε x ) directly reflects the axial tensile / compressive strain of the concrete surface layer under load or temperature, and is the core indicator of the structural health status.

[0023] Vertical component (ε y ) physical meaning: the strain component perpendicular to the connecting line direction (ε y ) is mainly caused by the expansion and contraction displacement of the longitudinal joint, has nothing to do with the load, and is a measurement noise.

[0024] Preferably, the dynamic calibration method of the load-strain transfer coefficient comprises the following steps: The load-strain transfer coefficient k is calculated based on the angle θ between the connection direction of the fiber Bragg grating strain sensor and the temperature sensor in each sensor array and the longitudinal joint direction, as well as the thickness H of the concrete surface layer, where θ is an angle between 45° and 90° and H is the actual thickness of the concrete surface layer. Based on the concrete elasticity theory, the functional relationship between k, angle θ and thickness H is established. Specifically, k is proportional to sinθ and inversely proportional to the square of H. The strain distribution of the concrete surface layer under different load and temperature conditions is simulated by a finite element simulation model to verify the validity of the functional relationship and generate a correction coefficient β; Multiply the correction factor β by the theoretical calculated value to obtain the final load-strain transfer coefficient ; Input the dynamically calibrated k into the correlation equation ε x (t)=k·σ L (t)+ε T (t) linear combination relationship, where ε x (t) is the component of load strain parallel to the sensor line, σ L (t) is the load stress component, ε T (t) is the theoretical strain component caused purely by temperature, through ε T (t)=α·ΔT(t) is calculated.

[0025] The dynamic calibration method of the present invention solves the problem that the existing technology does not dynamically correct the load-strain transfer coefficient in combination with the sensor angle direction and the thickness of the concrete surface layer, resulting in insufficient accuracy in the separation of temperature strain and load strain. Through theoretical modeling and simulation verification, the parameters are collaboratively optimized to adapt to the mechanical response characteristics of different engineering scenarios and improve the accuracy of strain component separation.

[0026] Derivation based on elastic mechanics: Assuming that the concrete surface layer is an isotropic elastic body, the component of the load stress in the direction of the sensor connection line is: σ L =(F·sinθ) / (A·H 2 ), where F is the load force and A is the sensitive area of the sensor.

[0027] Combined strain and stress relationship ε L =σ L / E, derive: The results of finite element simulations for the working conditions of θ=45°, 60°, 90° and H=25cm, 30cm, and 35cm show that the theoretical formula error is less than 5%.

[0028] The calibration method of the load-strain transfer coefficient k is based on the theory of elastic mechanics, and the specific relationship is: Among them, β is the correction coefficient determined through finite element simulation and field tests, with a value range of 0.95-1.05; θ is the angle between the connection line of the two sensors in the group and the joint direction; H is the actual thickness of the concrete surface layer; this formula is derived based on the theory of elastic mechanics, assuming that the concrete surface layer is an isotropic elastic body. The component of the load stress in the direction of the sensor connection is related to the angle θ and the thickness H.

[0029] Preferably, the time window sliding average processing and the method for calculating the real-time load strain variation include the following steps: The sliding time window length is set to 24 hours, and the window contains the strain data of 48 vertical components collected periodically; For each vertical component strain data ε within the sliding time window y (t) Perform arithmetic averaging to obtain the reference value ε of load strain y base; The reference value ε y base is the static background value of load strain in the current time window; Load-strain component ε according to the parallel component x (t) and the theoretical strain component ε caused by pure temperature T (t), calculate the real-time load strain change Δε L (t)=ε x (t)-ε T (t)-ε y base; When the sliding window moves forward one data period, the oldest data point in the window is removed and the latest data point is added, and the benchmark value ε is recalculated. y base.

[0030] The method of the present invention solves the problem that the existing technology fails to remove the random noise of the vertical component through sliding average processing, resulting in drift of the load-strain reference value. By dynamically updating the reference value and synchronously subtracting the temperature strain component, the true load-strain change is separated, thereby improving the reliability of structural damage identification.

[0031] The present invention adopts a 24-hour window corresponding to the diurnal temperature cycle, and 48 data points to ensure statistical stability; the window sliding step is consistent with the data acquisition cycle to ensure data continuity; the static reference value is extracted from the vertical component, and the dynamic change is calculated by combining the parallel component and the temperature component to realize multi-dimensional data fusion.

[0032] The present invention has at least the following beneficial effects: This invention utilizes a specially arranged sensor array, periodic data transmission, and dynamic threshold determination to achieve real-time monitoring of the health of the concrete surface layer, providing data support for structural maintenance and reducing the risk of structural damage caused by strain accumulation or sudden temperature changes. This addresses the problem of existing concrete surfaces in storage yards lacking long-term, real-time monitoring of internal strain and temperature changes, making it difficult to promptly identify concrete cracking risks and abnormal temperature stresses.

[0033] This invention uses fiber Bragg grating (FBG) strain sensors embedded in the middle of the concrete surface layer to reduce baseline interference with strain data caused by bending deformation. A 2-3 cm deep temperature sensor is also used to capture critical surface thermal stress mutation points caused by diurnal temperature differences, improving sensitivity. Furthermore, the sensor's angled placement suppresses strain measurement noise caused by joint displacement. This addresses the coupling error between strain and temperature data in existing monitoring systems, which occurs when the sensor's embedded position does not align with the temperature mutation zone between the middle and surface of the concrete surface layer and fails to eliminate interference from joint displacement on the measurement direction.

[0034] This invention leverages thermodynamic theory to isolate the influence of temperature on strain, avoiding false alarms caused by ambient temperature fluctuations. It also utilizes spatial directional decomposition to highlight the mechanical characteristics of load strain, suppressing the impact of joint displacement noise on the data. It also dynamically calibrates the transfer coefficient to adapt to the actual needs of different engineering parameters, improving the model's applicability. This addresses the problem that existing monitoring systems fail to physically separate temperature strain from load strain, resulting in alarm signals unable to distinguish between structural damage and normal thermal expansion and contraction, reducing the relevance of maintenance decisions.

[0035] This invention triggers alarms based on load-strain data separated by a vectorized model, avoiding false alarms caused by temperature fluctuations. It also incorporates a temperature-load correlation coefficient to identify the synergistic effects of sudden temperature changes and load-strain (such as the superimposed stresses caused by heavy load operations during cold waves). It is compatible with existing threshold parameters (400με, 20°C), eliminating the need for system recalibration and maintaining manageable implementation costs. This addresses the problem that existing threshold alarms rely directly on raw strain data without stripping out the temperature-strain component, leading to thermal expansion during hot seasons being easily misinterpreted as structural damage and failing to identify the coupled effects of temperature and load-strain.

[0036] The present invention solves the problem that existing monitoring systems fail to quantify temperature-strain components based on concrete parameters, resulting in the inability to physically separate temperature and load strain. By comparing theoretical calculations with measured data, the invention eliminates the interference of temperature fluctuations on the strain threshold and improves the accuracy of the alarm signal.

[0037] Other advantages, objectives and features of the present invention will be reflected in part from the following description and will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a top view of the structure of the on-site installation implementation of the storage yard concrete surface layer monitoring system of the present invention; Figure 2 It is a cross-sectional structural diagram of an on-site installation implementation of the storage yard concrete surface layer monitoring system of the present invention; Among them, the concrete surface layer surface 1; the longitudinal joint direction 2; the first coordinate axis 3; the second coordinate axis 4; the temperature sensor 5; the fiber Bragg grating strain sensor 6; the connection direction of the fiber Bragg grating strain sensor and the temperature sensor 7; the middle position 8 in the thickness direction of the concrete surface layer. DETAILED DESCRIPTION

[0039] The present invention is further described in detail below with reference to the embodiments so that those skilled in the art can implement the invention with reference to the description.

[0040] It should be noted that the experimental methods described in the following embodiments are conventional methods unless otherwise specified, and the reagents and materials are commercially available unless otherwise specified. In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "set" should be understood in a broad sense, for example, they can be fixedly connected or set, or detachably connected or set, or integrally connected or set. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The orientations or positional relationships indicated by the terms "transverse", "longitudinal", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc. are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0041] Figure 1-2 The figure shows an implementation form of on-site installation of a storage yard concrete surface monitoring system.

[0042] Example 1 A storage yard concrete surface layer monitoring system of the present invention comprises: Multiple sensor arrays are embedded within the concrete surface layer, each sensor array consisting of a fiber Bragg grating strain sensor and a temperature sensor. Each sensor array is spaced apart along the longitudinal joint direction 2 of the concrete surface layer, and the fiber Bragg grating strain sensor 6 is embedded in the middle position 8 in the thickness direction of the concrete surface layer, and the temperature sensor 5 is embedded within a depth of 2-4 cm below the surface 1 of the concrete surface layer. A data acquisition module connected to the sensor array, used to acquire real-time strain data of the fiber Bragg grating strain sensor and real-time temperature data of the temperature sensor; A wireless transmission module that communicates with the data acquisition module, the wireless transmission module adopts the LoRa communication protocol and transmits the real-time strain data and real-time temperature data to a remote server at a preset time (e.g., 30 minutes); A data analysis platform deployed on the remote server is configured with a threshold comparison algorithm. The threshold comparison algorithm sets the strain threshold to 400 microstrain and the temperature change threshold to 20 degrees Celsius. When the real-time strain data exceeds the strain threshold, or the change amplitude of the real-time temperature data within 24 hours exceeds the temperature change threshold, the data analysis platform generates an alarm signal and marks the corresponding sensor array position.

[0043] This invention utilizes a specially arranged sensor array, periodic data transmission, and dynamic threshold determination to achieve real-time monitoring of the health of the concrete surface layer, providing data support for structural maintenance and reducing the risk of structural damage caused by strain accumulation or sudden temperature changes. This addresses the problem of existing concrete surfaces in storage yards lacking long-term, real-time monitoring of internal strain and temperature changes, making it difficult to promptly identify concrete cracking risks and abnormal temperature stresses.

[0044] The fiber Bragg grating strain sensor is buried within the range of ±10% of the 1 / 2 thickness H of the concrete surface layer (i.e., the middle position in the direction of the concrete surface thickness); each sensor array is arranged at intervals of 4-5 meters along the longitudinal joint direction, and the distance between two sensors in the group is 0.1-0.6 meters. The connection direction of the fiber Bragg grating strain sensor and the temperature sensor forms an angle of 45° to 90° with the longitudinal joint direction, and the buried axis of the two (i.e., the sensor length direction) is consistent with the connection direction 7, see Figure 1-2 shown.

[0045] This invention uses fiber Bragg grating (FBG) strain sensors embedded in the middle of the concrete surface layer to reduce baseline interference with strain data caused by bending deformation. A 2-3 cm deep temperature sensor is also used to capture critical surface thermal stress mutation points caused by diurnal temperature differences, improving sensitivity. Furthermore, the sensor's angled placement suppresses strain measurement noise caused by joint displacement. This addresses the coupling error between strain and temperature data in existing monitoring systems, which occurs when the sensor's embedded position does not align with the temperature mutation zone between the middle and surface of the concrete surface layer and fails to eliminate interference from joint displacement on the measurement direction.

[0046] The physical mechanism by which sensor angle placement suppresses joint displacement noise is as follows: The expansion and contraction (e.g., thermal expansion and contraction) of a longitudinal joint primarily manifests as horizontal movement along the length of the joint. When the sensor line forms an angle with the joint (e.g., 60°), the joint displacement decomposes into a component along the line and a component perpendicular to the line. Fiber Bragg grating strain sensors are only sensitive to strain along their axis (i.e., the line), while the perpendicular component has no effect. The magnitude of the joint displacement component along the line is: ΔL·cosθ (ΔL is the total joint displacement, θ is the angle). When θ ≥ 45°, cosθ ≤ 0.707, indicating that the effect of joint displacement on the sensor is weakened. When θ = 90° (the line is perpendicular to the joint), cosθ = 0, indicating that the joint displacement has no direct effect on the sensor data.

[0047] Finite element analysis of the temperature distribution from the concrete surface to a depth of 5 cm revealed that the temperature fluctuation at a depth of 0-2 cm from the surface is 90%-95% of environmental fluctuations, while the fluctuation at a depth of 2-4 cm is 80%-85%, filtering out transient wind chill and sunlight interference. At a depth of 4 cm, the fluctuation is less than 50%, with significant response lag. This indicates that the 2-4 cm depth balances surface thermal stress sensitivity and interference resistance, making it the optimal burial range and enhancing sensitivity.

[0048] The setting of the strain threshold of 400 με is based on the following principles: Critical value of concrete cracking: The ultimate tensile strain of C30-C40 concrete is about 100-150 με. Considering the safety factor of 2.5-3.0, 400 με is taken; Actual measurement verification: In the preliminary test, after a 50-ton container was continuously subjected to stress for 30 days, the maximum strain value of the sensor was 380 με (no cracking), so the threshold was set at 400 με.

[0049] The setting basis of the temperature change threshold of 20℃ is: the critical point of concrete thermal stress is when ΔT>20℃, the temperature stress σ T =E·α·ΔT>6MPa (E=30GPa, α=10×10 -6 / ℃), which is more than twice the tensile strength of C30 concrete (2.0-2.5MPa), and is prone to cause cracking risks.

[0050] During implementation: The fiber Bragg grating strain sensor may be HBM FS62 (fiber Bragg grating strain sensor, measuring range ±1500με), and the temperature sensor may be a temperature sensor with a stainless steel package and an accuracy of ±0.5°C.

[0051] The concrete surface material is C30-C40 ordinary silicate concrete, and the thermal expansion coefficient α is 10×10 -6 / ℃, the elastic modulus E is 30GPa.

[0052] The fiber Bragg grating strain sensor is buried in the middle of the thickness of the concrete surface layer (for example, the buried depth of a 30cm thick surface layer is 15cm±3cm), and the temperature sensor is buried within a depth range of 2-4cm below the surface.

[0053] Each group of sensor arrays is arranged at intervals of 4-5 meters along the longitudinal seam direction, with a distance of 0.1 meters between two sensors in the group, and an angle (eg, 60°) is formed between the sensor connection direction and the seam direction.

[0054] Before concrete pouring, pre-install sensor fixing brackets (such as stainless steel U-shaped card slots) on the formwork, with the error of bracket spacing ≤ 5%. The fiber Bragg grating sensor is fixed at the center of the bracket through waterproof tape, and the temperature sensor is installed in the shallow preset hole position after being encapsulated with heat shrinkable sleeves. When vibrating the concrete, use an insertion vibrator with a low vibration frequency (≤ 50Hz), avoiding the range within 1 meter of the sensor to prevent displacement or damage. Within 7 days after pouring, detect the baseline drift of the sensor daily through a handheld fiber optic demodulator. When the drift amount > 10με, trigger recalibration.

[0055] For the data acquisition module, a National Instruments NI-9237 type multifunctional acquisition card can be selected, which supports synchronous acquisition of multi-channel data with a sampling frequency of 10Hz. The wireless transmission module uses a Semtech SX1276 type LoRa chipset and is equipped with a microcontroller. Each acquisition includes the average value of stable data within 10 seconds. The transmission power of the LoRa module is dynamically adjusted according to RSSI: when RSSI ≤ -110dBm, the power is 20dBm; when -110dBm < RSSI ≤ -90dBm, it is 17dBm; when RSSI > -90dBm, it is 14dBm. The module is installed in a waterproof control box at the edge of the concrete surface layer, with the box size of 300mm × 200mm × 150mm and the protection level of IP67.

[0056] During operation, the acquisition card is connected to the sensor through a shielded cable, and the strain and temperature data are read once per cycle (such as every 30 minutes). After the data is differentially encoded and compressed (only record data with a change amount ≥ 0.5με or a temperature change ≥ 0.1°C), it is transmitted to the server through the LoRa module. The transmission time slots are allocated according to the sensor numbers (for example, the 1st array sends at 00:00, and the 2nd sends at 00:05) to avoid channel conflicts.

[0057] The data analysis platform is developed based on Python and deployed on an Alibaba Cloud ECS server, using a MySQL database to store historical data. The threshold parameters are set as: the strain threshold is 400 microstrains, and the 24-hour temperature change threshold is 20°C. For the real-time load strain change amount Δε L (t) When it exceeds 400 microstrains, or when the temperature change amplitude ΔT ≥ 20°C within 24 hours, the platform generates an alarm signal and locates the corresponding sensor. The alarm information is pushed to the maintenance personnel through text messages and emails, and the event log is recorded simultaneously.

[0058] Through the above implementation manner of the present invention, the system can achieve long-term monitoring of the strain and temperature changes of the concrete surface layer, distinguish abnormal loads from temperature stress fluctuations, reduce the false alarm rate, and provide real-time data support for the maintenance of the yard structure.

[0059] Embodiment 2 In another embodiment of the present invention, based on embodiment 1, the data analysis platform is further configured with a strain and temperature vectorized correlation model, and the construction method thereof includes the following steps: S1, collecting real-time strain data of fiber Bragg grating strain sensors and real-time temperature data of temperature sensors in the same sensor array, and synchronously storing them to a remote server at a fixed period; S2. Based on the thermal expansion characteristics and elastic parameters of concrete, the theoretical strain component caused by pure temperature is calculated using temperature data, including the following steps: S2.1. Obtain the thermal expansion coefficient α and elastic modulus E of concrete; S2.2. Calculate the difference ΔT(t) between the current temperature T(t) and the reference temperature T0 based on the real-time temperature data collected by the temperature sensor, where the reference temperature T0 is the initial curing temperature of the concrete surface layer upon completion of pouring; S2.3, based on the thermal expansion strain formula ε T (t)=α·ΔT(t), calculate the theoretical strain component ε caused by pure temperature T (t); S2.4, according to the elastic modulus E, the theoretical strain component ε T (t) is converted into the temperature stress component σ T (t) = E·ε T (t), and σ T (t) inputting the strain and temperature vector correlation model; S2.5, the temperature stress component σ in the measured strain data is converted into T (t) and load strain component σ L (t) separation, and obtain the real-time load strain change Δε L (t); S2.6, the value range of the thermal expansion coefficient α is 8×10 -6 / ℃ to 12×10 -6 / ℃, the elastic modulus E ranges from 25GPa to 35GPa; S3, based on the angular position relationship between the fiber Bragg grating strain sensor and the temperature sensor, decomposing the measured strain data into a component parallel to the sensor connection direction and a component perpendicular to the connection direction, comprising the following steps: S3.1. Determine the plane rectangular coordinate system based on the angle θ formed between the connecting direction of each set of fiber Bragg grating strain sensors and temperature sensors and the longitudinal joint direction; use the connecting direction 7 as the first coordinate axis 3, and the direction perpendicular to the connecting direction and located in the plane of the concrete surface layer as the second coordinate axis 4, such as Figure 1-2 As shown; S3.2. Decompose the original strain data ε0(t) collected by each fiber Bragg grating strain sensor according to the following formula: S3.3. Strain component parallel to the connecting line: ε x (t)=ε0(t)·cosθ, the strain component perpendicular to the connecting line direction: ε y (t)=ε0(t·sinθ; x (t) as a component of load strain, ε y (t) is used as the reference component of the seam displacement noise.

[0060] S4. Establish a correlation equation to relate the parallel components to a linear combination of the temperature strain component and the load strain component, where the load strain transfer coefficient is dynamically calibrated according to the sensor angle direction and the concrete surface thickness, including the following steps: S4.1. Calculate the load-strain transfer coefficient k based on the angle θ between the connection direction of the fiber Bragg grating strain sensor and the temperature sensor in each sensor array and the longitudinal joint direction, and the thickness H of the concrete surface layer, where θ is an angle between 45° and 90°, and H is the actual thickness of the concrete surface layer; S4.2. Based on concrete elasticity theory, establish a functional relationship between k, angle θ, and thickness H. Specifically, k is proportional to sinθ and inversely proportional to the square of H. S4.3. Simulate the strain distribution of the concrete surface layer under different load and temperature conditions using a finite element simulation model to verify the validity of the functional relationship and generate a correction factor β; S4.4. Multiply the correction factor β by the theoretical calculated value to obtain the final load-strain transfer coefficient ; S4.5. Input the dynamically calibrated k into the correlation equation ε x (t)=k·σ L (t)+ε T (t) linear combination relationship, where ε x (t) is the component of load strain parallel to the sensor line, σ L (t) is the load stress component, ε T (t) is the theoretical strain component caused purely by temperature, through ε T (t)=α·ΔT(t) is calculated.

[0061] S5. Performing time window sliding average processing on the strain data of the vertical component to extract the baseline value of the load strain, and calculating the real-time load strain variation by the difference between the parallel component and the temperature strain component, including the following steps: S5.1. Set the sliding time window length to 24 hours, and the window contains the strain data of 48 vertical components collected periodically; S5.2. For each vertical component strain data ε within the sliding time window y (t) Perform arithmetic averaging to obtain the reference value ε of load strain y base; S5.3, the reference value ε y base is the static background value of load strain in the current time window; S5.4 Load-strain component ε according to the parallel component x (t) and the theoretical strain component ε caused by pure temperature T (t), calculate the real-time load strain change Δε L (t)=ε x (t)-ε T (t)-ε y base; S5.5. When the sliding window moves forward one data period, the oldest data point in the window is removed and the latest data point is added, and the baseline value ε is recalculated. y base.

[0062] S6. When the data analysis platform generates an alarm signal, the real-time load strain change Δε output by the strain and temperature vector correlation model is converted into L (t) as input data for the threshold comparison algorithm; When Δε L (t) when the strain threshold of 400 microstrain is exceeded, it is determined to be a load abnormality and an alarm signal is generated; When the real-time temperature data changes by more than 20 degrees Celsius within 24 hours, and Δε L (t) When the Pearson correlation coefficient with the temperature variation amplitude meets one of the following conditions, it is determined to be a temperature stress anomaly and an alarm signal is generated: (1) |r|>0.7 and the signs are consistent in two consecutive sliding time windows; (2) The temperature variation range ΔT>25℃ and |r|>0.7.

[0063] This invention leverages thermodynamic theory to isolate the influence of temperature on strain, avoiding false alarms caused by ambient temperature fluctuations. It also utilizes spatial directional decomposition to highlight the mechanical characteristics of load strain, suppressing the impact of joint displacement noise on the data. It also dynamically calibrates the transfer coefficient to adapt to the actual needs of different engineering parameters, improving the model's applicability. This addresses the problem that existing monitoring systems fail to physically separate temperature strain from load strain, resulting in alarm signals unable to distinguish between structural damage and normal thermal expansion and contraction, reducing the relevance of maintenance decisions.

[0064] This invention triggers alarms based on load-strain data separated by a vectorized model, avoiding false alarms caused by temperature fluctuations. It also incorporates a temperature-load correlation coefficient to identify the synergistic effects of sudden temperature changes and load-strain (such as the superimposed stresses caused by heavy load operations during cold waves). It is compatible with existing threshold parameters (400με, 20°C), eliminating the need for system recalibration and maintaining manageable implementation costs. This addresses the problem that existing threshold alarms rely directly on raw strain data without stripping out the temperature-strain component, leading to thermal expansion during hot seasons being easily misinterpreted as structural damage and failing to identify the coupled effects of temperature and load-strain.

[0065] Add a data preprocessing module to the data analysis platform and convert Δε L (t) is synchronized with the temperature data to input the threshold comparison algorithm, without adding new hardware, through Δε L (t)>400με directly reflects the risk of structural damage; temperature changes exceeding 20℃ and with Δε L (t) is strongly correlated (|r|>0.7), suggesting that sudden temperature change induces additional stress.

[0066] The calculation of the correlation coefficient is achieved by the following steps: 1. Set a 24-hour sliding time window in the data analysis platform and extract the temperature change amplitude ΔT(t) and real-time load strain change Δε within the window L (t) time series data, where: ΔT(t) represents the difference between the real-time temperature data of the temperature sensor at time point t and the temperature data 24 hours ago; Δε L (t) represents the real-time load strain change output by the strain and temperature vectorization correlation model; The sliding time window is defined as 48 data points collected at a fixed period (e.g., 30 minutes) over a continuous 24-hour period; 2. Calculate ΔT(t) and Δε based on the Pearson correlation coefficient formula L The linear correlation of (t) is calculated as follows: in: n is the total number of data points in the sliding time window, n=48; ΔT i is the temperature change amplitude corresponding to the i-th data point; Δε Li is the real-time load strain change corresponding to the i-th data point; For all ΔT in the window i The arithmetic mean of For all Δε in the window Li The arithmetic mean of 3. When the following conditions are met, it is determined to be abnormal temperature stress: |r|>0.7 (indicating strong linear correlation); If the correlation coefficients in two consecutive sliding time windows have the same sign (both positive or both negative), or if the temperature change amplitude ΔT exceeds 25°C, only one window needs to meet |r|>0.7 to trigger an alarm.

[0067] The present invention solves the problem that existing monitoring systems fail to quantify temperature-strain components based on concrete parameters, resulting in the inability to physically separate temperature and load strain. By comparing theoretical calculations with measured data, the invention eliminates the interference of temperature fluctuations on the strain threshold and improves the accuracy of the alarm signal.

[0068] The decomposition method of the present invention solves the problem that the existing technology does not establish a strain vector coordinate system based on the angular relationship between the sensor connection direction and the joint, resulting in inaccurate calculation directions of the parallel and vertical components. Through coordinated processing of coordinate projection and sensor array, the strain interference of joint displacement in non-sensitive directions is eliminated, thereby improving the separation accuracy of load-strain components.

[0069] Parallel component (ε x ) physical meaning: the strain component parallel to the sensor line direction (ε x ) directly reflects the axial tensile / compressive strain of the concrete surface layer under load or temperature, and is the core indicator of the structural health status.

[0070] Vertical component (ε y ) physical meaning: the strain component perpendicular to the connecting line direction (ε y ) is mainly caused by the expansion and contraction displacement of the longitudinal joint, has nothing to do with the load, and is a measurement noise.

[0071] The dynamic calibration method of the present invention solves the problem that the existing technology does not dynamically correct the load-strain transfer coefficient in combination with the sensor angle direction and the thickness of the concrete surface layer, resulting in insufficient accuracy in the separation of temperature strain and load strain. Through theoretical modeling and simulation verification, the parameters are collaboratively optimized to adapt to the mechanical response characteristics of different engineering scenarios and improve the accuracy of strain component separation.

[0072] Derivation based on elastic mechanics: Assuming that the concrete surface layer is an isotropic elastic body, the component of the load stress in the direction of the sensor connection line is: σ L =(F·sinθ) / (A·H 2 ), where F is the load force and A is the sensitive area of the sensor.

[0073] Combined strain and stress relationship ε L =σ L / E, derive: The results of finite element simulations for the working conditions of θ=45°, 60°, 90° and H=25cm, 30cm, and 35cm show that the theoretical formula error is less than 5%.

[0074] The calibration method of the load-strain transfer coefficient k is based on the theory of elastic mechanics, and the specific relationship is: Where β is a correction factor determined through finite element simulation and field testing, with a value range of 0.95-1.05; θ is the angle between the line connecting the two sensors in the group and the joint direction; and H is the actual thickness of the concrete surface layer. This formula is derived from elastic mechanics theory and assumes that the concrete surface layer is an isotropic elastic body. The component of the load stress in the direction of the sensor line is related to the angle θ and the thickness H. The simulation verification process for the correction factor β includes the following steps: a. Create a 3D finite element model of the concrete surface layer and set material parameters (elastic modulus E = 30 GPa, Poisson's ratio ν = 0.2); b. Apply a vertical load F (range: 10kN-100kN) and a temperature gradient ΔT (range: 10℃-40℃) to the model; c. Extract the simulated strain value ε in the direction of the sensor connection sim , and calculate k based on the theoretical formula theory = sinθ / H²; d. Calculate the correction factor β=ε sim / (k theory ·σ L ); e. Repeat steps a to b for θ = 45°, 60°, 90° and H = 25cm, 30cm, 35cm to generate the fitting formula for β: β=0.98+0.02·sinθ−0.001·H The correction coefficient β ranges from 0.95 to 1.05.

[0075] The method of the present invention solves the problem that the existing technology fails to remove the random noise of the vertical component through sliding average processing, resulting in drift of the load-strain reference value. By dynamically updating the reference value and synchronously subtracting the temperature strain component, the true load-strain change is separated, thereby improving the reliability of structural damage identification.

[0076] The present invention adopts a 24-hour window corresponding to the diurnal temperature cycle, and 48 data points to ensure statistical stability; the window sliding step is consistent with the data acquisition cycle to ensure data continuity; the static reference value is extracted from the vertical component, and the dynamic change is calculated by combining the parallel component and the temperature component to realize multi-dimensional data fusion.

[0077] The LoRa communication protocol configuration of the wireless transmission module includes the following optimization methods: According to the real-time RSSI value between the sensor array and the remote server, dynamically adjust the transmission power of the LoRa module. The mapping relationship between the transmission power P and RSSI is as follows: when RSSI ≤ -110dBm, P = 20dBm; when -110dBm < RSSI ≤ -90dBm, P = 17dBm; when RSSI > -90dBm, P = 14dBm; Before data transmission, perform differential coding compression on the real-time strain data and temperature data. The compression rule is: only record the change amount of adjacent cycle data, and mark it as invalid data and discard it when the change amount is less than the sensor accuracy; Adopt a time division multiple access mechanism to allocate data transmission time slots. The wireless transmission module of each group of sensor arrays sends data at a fixed time offset within a preset cycle, and the offset is set according to an arithmetic progression according to the array number; The optimization method solves the problem of high data loss rate caused by signal attenuation and channel competition in the complex environment of the yard in existing LoRa communication. Through the synergistic effect of power adaptation, data compression and time slot allocation, while ensuring the transmission distance, it reduces energy consumption and conflict probability, and improves the integrity and real-time of monitoring data.

[0078] The present invention matches the transmission power in real time based on RSSI, avoids packet loss of long-distance sensors due to insufficient power, and at the same time suppresses signal overload of short-distance sensors; eliminates redundant data, reduces the amount of data transmitted each time, and reduces the channel occupancy time; allocates time slots according to the number, avoids channel conflicts caused by multiple sensors sending at the same time, and ensures that the data transmission timing is controllable.

[0079] During the implementation process: The fiber Bragg grating strain sensor can adopt HBMFS62 (fiber Bragg grating strain sensor, measuring range ±1500με), and the temperature sensor can select a temperature sensor with a stainless steel package and an accuracy of ±0.5℃.

[0080] The concrete surface layer material is C30-C40 ordinary portland cement concrete, the thermal expansion coefficient α is taken as 10×10⁻ 6 / ℃, and the elastic modulus E is taken as 30GPa.

[0081] The fiber Bragg grating strain sensor is buried in the middle position of the concrete surface layer thickness (for example, the buried depth of a 30cm thick surface layer is 15cm ± 3cm), and the temperature sensor is buried within the depth range of 2-4cm below the surface.

[0082] Each group of sensor arrays is arranged at intervals of 4-5 meters along the longitudinal joint direction. The distance between the two sensors in the group is 0.1 meters, and the angle formed by the sensor connection direction and the joint direction is (for example, 60°).

[0083] Before pouring concrete, pre-install sensor mounting brackets (such as stainless steel U-shaped slots) on the formwork, with a bracket spacing tolerance of ≤5%. Secure the fiber Bragg grating sensor to the center of the bracket with waterproof tape. Enclose the temperature sensor in heat shrink tubing and install it in a pre-set shallow hole. When vibrating the concrete, use a low-frequency (≤50 Hz) insert vibrator, keeping it within 1 meter of the sensor to prevent displacement or damage. For 7 days after pouring, check the sensor baseline drift daily with a handheld fiber optic interrogator. Drift > 10 με triggers recalibration.

[0084] The data acquisition module can use the National Instruments NI-9237 multi-function acquisition card, which supports synchronous acquisition of multi-channel data with a sampling frequency of 10 Hz.

[0085] The strain and temperature vectorization correlation model configured in the data analysis platform is developed based on MATLAB Simulink, and uses a linear regression algorithm to separate the temperature and load strain components.

[0086] During operation, the strain data ε0(t) and temperature data T(t) of the same sensor array are collected and stored in the server in 30-minute cycles. According to the angle θ between the sensor line and the joint (for example, 60°), the original strain data is decomposed into parallel components ε x (t)=ε0(t·cosθ and the vertical component ε y (t)=ε0(t)·sinθ. According to the thermal expansion formula ε T (t)=α·ΔT(t) to calculate the pure temperature strain component, the elastic modulus E is taken as 30GPa, and ε T (t) is converted to temperature stress σ T (t) = E·ε T (t). Establish the correlation equation ε x (t)=k·σ L (t)+ε T (t), where the load-strain transfer coefficient k is dynamically calibrated according to the angle θ and the concrete thickness H (e.g., k = 0.85 when θ = 60° and H = 30 cm).

[0087] Finite element software (such as ANSYS) was used to simulate the strain distribution at different angles (45°, 60°, 90°) and thicknesses (25cm, 30cm, 35cm), verify the theoretical formula of the transfer coefficient k=sinθ / (H²), and add a correction factor β (β=0.95-1.05) based on the simulation results.

[0088] The data analysis platform adds a sliding average filter module to the vertical component ε y (t) Perform 24-hour window sliding average processing to extract the benchmark value ε ybase.

[0089] Real-time load strain change Δε L (t)=ε x (t)-ε T (t)-ε y The base exceeds 400 microstrains, or the temperature change ΔT ≥ 20°C in 24 hours and Δε in two consecutive sliding time windows L The Pearson correlation coefficient of (t) is greater than 0.7, or the temperature change ΔT is greater than 25°C and |r| is greater than 0.7 within one sliding time window.

[0090] Through the above-mentioned implementation mode of the present invention, the system can effectively separate temperature strain and load strain, suppress measurement noise caused by joint displacement, improve the accuracy of alarm signals, and provide reliable data support for long-term health monitoring of the concrete surface layer of the yard.

[0091] Test 1: Alarm accuracy test In Area A of a port container yard, the concrete surface layer of each test area has a size of 100m×50m and a thickness of 30cm. The material is C35 ordinary Portland concrete (cement: P.O42.5; sand: medium-coarse sand; gravel: 5-20mm crushed stone; water-cement ratio 0.45).

[0092] The comparison group's technical solution used a traditional resistance strain gauge (model: Vishay CEA-06-250UW-120), without an integrated temperature-strain separation algorithm. The alarm threshold was directly set to 400με based on the original strain data, and the temperature change threshold was 20°C.

[0093] Experimental group technical solution: adopt the specific implementation methods of Example 1 and Example 2.

[0094] Test period: March 2023 to September 2023, ambient temperature range: -10℃ to 45℃, humidity 30%-85%.

[0095] Comparison group: 20 resistance strain gauges were evenly arranged in the test area with a spacing of 5 meters, and the data was transmitted to the cloud via LoRa.

[0096] Experimental group: 20 sensor arrays (5 meters apart) were arranged in the same area. Each array contained one fiber Bragg grating strain sensor and one temperature sensor. The data was transmitted to the cloud via LoRa.

[0097] The load anomaly is as follows: a 50-ton container (size 12m×2.5m) is placed in the center of the area for 30 days, and the crack growth of the concrete surface layer is manually marked.

[0098] The temperature changes suddenly occurred by spraying with a high-pressure water gun at night (water temperature 5°C for 2 hours, regional temperature drop of 20°C) and covering with insulation film during the day (regional temperature increase of 15°C) to simulate extreme temperature changes.

[0099] The experimental group and the control group simultaneously collected strain and temperature data, and recorded alarm events (trigger threshold conditions) and manual inspection results.

[0100] Crack detection was performed manually using a Dino-Lite AM7915MZTS digital microscope (220× magnification) to record crack width (≥0.1 mm was considered effective damage). The results are shown in Table 1.

[0101] Table 1: index Example 1 Example 2 Control group (existing technology) Number of abnormal load alarms 9 11 times 8 times Temperature stress alarm times 5 9 times 3 times Number of false alarms (when there is no damage) 7 2 times 15 times Number of missed reports (real damage) 2 1 time 4 times Missed alarm rate = number of missed alarms / actual number of damages × 100%; False alarm rate = number of false alarms / total number of alarms × 100% Missing report rate of Example 2: 1 / 12×100%≈8.3%; Missing report rate of control group: 4 / 12×100%≈33.3%; False alarm rate in Example 2: 2 / (11+9+2)×100%≈9.1%; false alarm rate in the control group: 15 / (8+3+15)×100%≈41.7%; It can be seen that Example 2 significantly reduces the false alarm and missed alarm rates through the temperature-strain separation algorithm and optimized threshold setting.

[0102] Experiment 2: Long-term reliability test Test object: HBM FS62 fiber optic sensor (encapsulation material: polyimide coating + 304 stainless steel housing), embedded in a C35 concrete block (1m × 1m × 0.3m).

[0103] Salt spray environment: according to ASTM B117 standard, 5% NaCl solution, temperature 35 ° C, humidity 95%, for 12 months Freeze-thaw cycle: According to GB / T50082-2009 standard, -20℃ (4 hours) → 25℃ (4 hours), cycle 30 times.

[0104] Alkaline soaking: NaOH solution with pH=12.5, soak at room temperature for 6 months.

[0105] Strain measurement: Micron Optics sm125 fiber optic interrogator (resolution 1 με).

[0106] Package integrity: Olympus IPLEX GX / GT industrial endoscopes.

[0107] The test results are shown in Table 2.

[0108] Table 2: Test conditions Signal drift (με / year) Package damage rate (%) Functional failure rate (%) Normal environment (24 months) 8±1.6 0 0 Salt spray environment (12 months) 12±2.1 3.2 2.5 Freeze-thaw cycles (30 times) 18±3.3 6.7 5.0 Alkaline soaking (6 months) 22±4.2 10.5 8.3 The measurement results in Table 2 show that under normal conditions, the average annual drift was 8.2 με, with a standard deviation of 1.6 με, indicating good sensor baseline stability. However, under salt spray, freeze-thaw, and alkaline conditions, the drift increased to 1.5 times, 2.3 times, and 2.8 times, respectively, compared to normal conditions. This is primarily due to degradation of the coating and packaging caused by environmental corrosion. In salt spray, the packaging failure rate was 3.2%, corresponding to 2.5% functional failures, primarily due to uneven stress on the internal optical fiber caused by corrosion of the stainless steel casing. Under alkaline immersion, the packaging failure rate was 10.5%, corresponding to 8.3% functional failures, primarily due to strain transmission distortion caused by swelling of the polyimide coating. This indicates that system durability can be effectively improved through appropriate packaging and under normal conditions.

[0109] Test 3: Joint displacement noise suppression rate test The dimensions of the concrete specimens were: 2 m (length) × 1 m (width) × 0.3 m (thickness); C35 ordinary Portland concrete, with a mix ratio (by weight) of cement: sand: gravel: water = 1:1.8:2.7:0.45; standard curing for 28 days, humidity ≥ 95%, temperature 20 ± 2°C.

[0110] Sensor array: Each group contains one fiber Bragg grating strain sensor and one temperature sensor; a longitudinal seam (5 mm wide) is preset along the center line of the specimen, and three groups of sensor arrays are installed on both sides of the seam, with angles θ = 45°, 60°, and 90°, respectively, and the distance between the two in the group is 0.5 m.

[0111] Electric translation stage (accuracy ±0.01mm), simulating the joint thermal expansion and contraction displacement ΔL=5mm; Test method: Specimen fixation: Fix the concrete specimen on a rigid base to ensure that the joints can expand and contract freely; Displacement application: A horizontal displacement of ΔL = 5 mm was applied at a rate of 0.1 mm / s using a displacement stage (simulating joint expansion and contraction). Data acquisition: synchronously record strain data of three sets of sensors for 10 minutes; Noise calculation: Theoretical noise component: ε 理论 =ΔL·cosθ / L (L=2m is the length of the specimen); Measured noise component: ε 实测 = sensor strain value − baseline value in no displacement state; Measured noise suppression rate: η1=1−(ε 实测 / ε 理论 ).

[0112] In the joint displacement noise suppression rate verification test, a C35 concrete specimen (2m × 1m × 0.3m) was used. A longitudinal joint (5mm wide) was pre-set along the specimen's centerline, and three sensor arrays (θ = 45°, 60°, and 90°) were installed. After applying a horizontal displacement of ΔL = 5mm using a motorized translation stage, the measured noise suppression rate η1 was compared with the theoretical formula η2 = 1 − cosθ. The following results are available: θ=45°: The theoretical noise suppression rate η2 is 29.3%, the measured value η1 is 27.5%, and the error is 1.8%; θ=60°: theoretical value 50.0%, measured value 48.1%, error 1.9%; θ=90°: theoretical value 100%, measured value 98.2%, error 1.8%.

[0113] The experiments show that the quantitative relationship between the sensor angle θ and the noise suppression rate is established, and the measured errors are all less than 2%. It can be seen that the angle arrangement can effectively suppress the joint displacement noise.

[0114] Experiment 4: Multi-scenario model experiment Specimen: Port yard simulation area (20m×20m×0.3m), longitudinal joints (5mm width) preset along the center line of the specimen, C40 concrete, reinforcement ratio 0.6%; Sensor array: One group (8 groups in total) is arranged every 5 meters along both sides of the longitudinal seam, θ=60°, the fiber Bragg grating strain sensor is buried at a depth of 15 cm, and the sensor arrays on both sides of the longitudinal seam are staggered.

[0115] Scenario 1 (asymmetric loading): 30-ton container (6m x 2.4m) placed eccentrically (1.5m from the joint) for 30 days.

[0116] Scenario 2 (sudden temperature change): Cover with black insulation film during the day (temperature rises by 15°C), and spray 10°C cold water at night (temperature drops by 20°C); the cycle period is 24 hours and lasts for 7 days.

[0117] Scenario 3 (combined action): 30-ton container + day and night temperature control cycle, lasting 15 days.

[0118] Determination method: Manual inspection reveals that crack width ≥ 0.1 mm (microscope measurement) is real damage; use the system monitoring of Example 2.

[0119] Accuracy = number of correct alarms / number of actual damages × 100%; False alarm rate = number of false alarms / total number of alarms × 100%.

[0120] In a simulated port yard area (50m×50m×0.3m, C40 concrete), the system's performance under different working conditions is as follows: Asymmetric Loading (Scenario 1): After a 30-ton container was placed eccentrically for 30 days, the system correctly reported seven alarms (eight instances of actual damage), with an accuracy rate of 87.5%. It also reported one false alarm (with a false alarm rate of 12.5%), likely due to the eccentric loading causing local strain concentration that exceeded the model's preset range.

[0121] Sudden temperature change (Scenario 2): The temperature suddenly changed by ΔT = 35°C (a 15°C rise and a 20°C drop) within 24 hours. The system correctly reported four alarms (five of which indicated actual damage), with an accuracy rate of 80.0%. There were also two false alarms (with a false alarm rate of 33.3%), possibly due to the instantaneous temperature gradient exceeding the linear thermal expansion range assumed by the model.

[0122] Combined action (Scenario 3): Load and temperature cycles were coupled for 15 days. The system correctly alarmed 8 times (10 times in the event of actual damage), with an accuracy rate of 80.0%. There were also 3 false alarms (false alarm rate of 27.3%), which may be due to the synergistic effect of load and temperature, which led to the accumulation of strain component separation errors.

[0123] The test results show that the system performs best under asymmetric loads (with an accuracy of 87.5%), verifying its adaptability to complex mechanical distributions; the accuracy rate remains at 80% in complex action scenarios, demonstrating the model's effective decoupling capabilities for the coupling of multiple physical fields.

[0124] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiment. They can be applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily realized.

Claims

1. The yard concrete surface monitoring system is characterized by: include: Multiple sensor arrays are embedded within the concrete surface layer, each sensor array consisting of a fiber Bragg grating strain sensor and a temperature sensor. Each sensor array is spaced apart along the longitudinal joint direction of the concrete surface layer, and the fiber Bragg grating strain sensor is embedded in the middle of the thickness direction of the concrete surface layer, and the temperature sensor is embedded within a depth of 2-4 centimeters below the surface of the concrete surface layer. A data acquisition module connected to the sensor array, used to acquire real-time strain data of the fiber Bragg grating strain sensor and real-time temperature data of the temperature sensor; A wireless transmission module that communicates with the data acquisition module, the wireless transmission module adopts the LoRa communication protocol and transmits the real-time strain data and real-time temperature data to a remote server at a preset time period; A data analysis platform deployed on the remote server is configured with a threshold comparison algorithm, which sets a strain threshold and a temperature change threshold. When the real-time strain data exceeds the strain threshold, or the change amplitude of the real-time temperature data within 24 hours exceeds the temperature change threshold, the data analysis platform generates an alarm signal and marks the corresponding sensor array position.

2. The yard concrete surface monitoring system according to claim 1, characterized in that: The fiber Bragg grating strain sensor is buried within the range of ±10% at 1 / 2 of the thickness H of the concrete surface layer; each group of sensor arrays is arranged at intervals of 4-5 meters along the longitudinal joint direction, and the distance between two sensors in the group is 0.1-0.6 meters. The connection direction of the fiber Bragg grating strain sensor and the temperature sensor forms an angle of 45° to 90° with the longitudinal joint direction, and the buried axis of the two is consistent with the connection direction.

3. The yard concrete surface monitoring system according to claim 2, characterized in that: The data analysis platform is also configured with a strain and temperature vectorized correlation model, the construction method of which includes the following steps: Collect the real-time strain data of the fiber Bragg grating strain sensor and the real-time temperature data of the temperature sensor in the same sensor array, and store them synchronously to the remote server at a fixed period; Based on the thermal expansion characteristics and elastic parameters of concrete, the theoretical strain component caused by pure temperature is calculated through temperature data; Decomposing the measured strain data into a component parallel to the sensor connection direction and a component perpendicular to the connection direction according to the angular position relationship between the fiber Bragg grating strain sensor and the temperature sensor; A correlation equation is established to relate the parallel components to a linear combination of the temperature strain component and the load strain component, where the load strain transfer coefficient is dynamically calibrated according to the sensor angle direction and the thickness of the concrete surface layer; The strain data of the vertical component are processed by time window sliding average to extract the baseline value of the load strain, and the real-time load strain change is calculated by the difference between the parallel component and the temperature strain component.

4. The yard concrete surface monitoring system according to claim 3, characterized in that: When the data analysis platform generates an alarm signal, the real-time load strain change Δε output by the strain and temperature vector correlation model is converted into L (t) as input data for the threshold comparison algorithm; When Δε L (t) when the strain threshold of 400 microstrain is exceeded, it is determined to be a load abnormality and an alarm signal is generated; When the real-time temperature data changes by more than 20 degrees Celsius within 24 hours, and Δε L (t) When the Pearson correlation coefficient with the temperature variation amplitude meets one of the following conditions, it is determined to be a temperature stress anomaly and an alarm signal is generated: (1) |r|>0.7 and the signs are consistent in two consecutive sliding time windows; (2) The temperature variation amplitude ΔT>25℃ and |r|>0.7 within one sliding time window.

5. The storage yard concrete surface layer monitoring system according to claim 3 or 4, characterized in that: The method for calculating the theoretical strain component caused purely by temperature comprises the following steps: Obtain the thermal expansion coefficient α and elastic modulus E of concrete; Calculate the difference ΔT(t) between the current temperature T(t) and the reference temperature T0 based on the real-time temperature data collected by the temperature sensor, where the reference temperature T0 is the initial curing temperature of the concrete surface layer when pouring is completed; Based on the thermal expansion strain formula ε T (t)=α·ΔT(t), calculate the theoretical strain component ε caused by pure temperature T (t); According to the elastic modulus E, the theoretical strain component ε T (t) is converted into the temperature stress component σ T (t) = E·ε T (t), and σ T (t) inputting the strain and temperature vector correlation model; The temperature stress component σ in the measured strain data is converted into T (t) and load strain component σ L (t) separation, and obtain the real-time load strain change Δε L (t); The thermal expansion coefficient α ranges from 8×10 -6 / ℃ to 12×10 -6 / ℃, the elastic modulus E ranges from 25GPa to 35GPa.

6. The yard concrete surface monitoring system according to claim 5, characterized in that: The method for decomposing the measured strain data into a component parallel to the sensor connection direction and a component perpendicular to the connection direction comprises the following steps: Determine the plane rectangular coordinate system based on the angle θ formed by the connection direction of each group of fiber Bragg grating strain sensors and temperature sensors and the longitudinal joint direction; take the connection direction as the first coordinate axis, and take the direction perpendicular to the connection direction and located in the plane of the concrete surface layer as the second coordinate axis; The original strain data ε0(t) collected by each fiber Bragg grating strain sensor is decomposed according to the following formula: Strain component parallel to the connecting line: ε x (t)=ε0(t)·cosθ, the strain component perpendicular to the connecting line direction: ε y (t)=ε0(t·sinθ; x (t) as a component of load strain, ε y (t) is used as the reference component of the seam displacement noise.

7. The yard concrete surface monitoring system according to claim 6, characterized in that: The dynamic calibration method of the load-strain transfer coefficient comprises the following steps: The load-strain transfer coefficient k is calculated based on the angle θ between the connection direction of the fiber Bragg grating strain sensor and the temperature sensor in each sensor array and the longitudinal joint direction, as well as the thickness H of the concrete surface layer, where θ is an angle between 45° and 90° and H is the actual thickness of the concrete surface layer. Establish a functional relationship between k, angle θ, and thickness H. Specifically, k is proportional to sinθ and inversely proportional to the square of H. The strain distribution of the concrete surface layer under different load and temperature conditions is simulated by a finite element simulation model to verify the validity of the functional relationship and generate a correction coefficient β; Multiply the correction factor β by the theoretical calculated value to obtain the final load-strain transfer coefficient ; Input the dynamically calibrated k into the correlation equation ε x (t)=k·σ L (t)+ε T (t), where ε x (t) is the component of load strain parallel to the sensor line, σ L (t) is the load stress component, ε T (t) is the theoretical strain component caused purely by temperature, through ε T (t)=α·ΔT(t) is calculated.

8. The yard concrete surface monitoring system according to claim 7, characterized in that: The time window sliding average processing and real-time load strain change calculation method includes the following steps: The sliding time window length is set to 24 hours, and the window contains the strain data of 48 vertical components collected periodically; For each vertical component strain data ε within the sliding time window y (t) Perform arithmetic averaging to obtain the reference value ε of load strain y base; The reference value ε y base is the static background value of load strain in the current time window; Load-strain component ε according to the parallel component x (t) and the theoretical strain component ε caused by pure temperature T (t), calculate the real-time load strain change Δε L (t)=ε x (t)-ε T (t)-ε y base; When the sliding window moves forward one data period, the oldest data point in the window is removed and the latest data point is added, and the benchmark value ε is recalculated. y base.

Images