An adaptive positioning device and method for steel structure health monitoring sensor

Abstract

The application provides a kind of self-adapting positioning device and method for steel structure health monitoring sensor, through multidimensional acquisition environment, structure and position attitude data, combined with improved LSTM (long short-term memory network) and adaptive Kalman filtering fusion algorithm and double-thread data fusion strategy, precise positioning calibration and dynamic model updating are realized, with micro-displacement drive, mechanical locking and double power supply, supplemented by encrypted hierarchical storage, multi-mode communication, then through piezoelectric ceramic drive three-axis fine tuning and inclination self-adapting calibration linkage, the function of efficient positioning and calibration sensor is realized.The application effectively improves the positioning accuracy and environmental interference resistance of the sensor, ensures the positioning stability under dynamic deformation, and ensures the reliability of the monitoring data and the continuous controllability of the positioning process.The application solves the problems of low calibration efficiency and insufficient real-time performance in the prior art, supports fast repositioning and calibration, and improves the overall efficiency of steel structure health monitoring.

Description

Technical Field

[0001] This invention belongs to the field of steel structure health monitoring technology, specifically relating to an adaptive positioning device and method for steel structure health monitoring sensors. Background Technology

[0002] With the widespread application of steel structures in bridges, high-rise buildings, industrial plants, and other fields, fatigue damage, corrosion, and deformation during long-term service directly threaten structural safety. Steel structure health monitoring relies on sensors to accurately collect data such as vibration, strain, and temperature, and the positioning accuracy of these sensors directly determines the reliability of the monitoring results.

[0003] Existing sensor positioning methods mostly rely on fixed installation or manual calibration, which have significant drawbacks: First, vibration and temperature changes in steel structures can easily cause sensor position shifts, and traditional fixed devices cannot adaptively correct these shifts; second, in complex structures (such as long-span bridges and irregularly shaped steel structures), it is difficult to dynamically adjust the sensor's positioning after deployment, resulting in blind spots in key areas; third, positioning error compensation relies on only a single parameter and does not incorporate multi-source data fusion, leading to poor environmental adaptability; and fourth, there is a lack of adaptive positioning logic tailored to the deformation characteristics of steel structures, resulting in low calibration efficiency and insufficient real-time performance.

[0004] Therefore, there is an urgent need for an efficient sensor positioning and calibration method. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an adaptive positioning device and method for a steel structure health monitoring sensor, which is used for efficient positioning and calibration of the sensor.

[0006] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: an adaptive positioning device for a steel structure health monitoring sensor, comprising a data sensing unit, a computing unit, a mechanical and power supply unit, a communication and storage unit, and an adaptive execution unit; the data sensing unit is connected to the computing unit; the computing unit is connected to the communication and storage unit and the adaptive execution unit respectively; the mechanical and power supply unit is used to supply power to all units; The data sensing unit is used to collect environmental data, structural status data, and sensor position and attitude data of the steel structure; The computing unit is used to perform positioning calibration of the sensor based on the data collected by the data sensing unit, while dynamically adjusting the logic control and linking with the adaptive execution unit to complete the positioning optimization. The mechanical and power supply unit is used to provide physical fixation support, servo adjustment capability, and stable power supply for the device; The communication and storage unit is used for real-time transmission of location data, as well as for storing and encrypting historical monitoring data and location parameters. The adaptive execution unit is used to fine-tune the position and calibrate the attitude of the sensor according to the control instructions of the computing unit.

[0007] According to the above scheme, the data sensing unit includes an environmental and structural data acquisition module and a position and attitude acquisition module; The environmental and structural data acquisition module is used to collect surface temperature, humidity, vibration amplitude and strain data of the steel structure, which serve as the basis for environmental correction for positioning accuracy compensation. The position and attitude acquisition module is used to acquire the GPS coordinates, tilt angle and displacement data of the sensor in real time. Combined with the installation parameters, the position compensation amount is calculated according to the position compensation amount calculation formula to correct the positioning deviation caused by environmental interference.

[0008] Furthermore, This is the location compensation value; This refers to the sensor sensitivity coefficient; This refers to the vibration amplitude of the steel structure. Install the tilt angle for the sensor; This refers to the change in ambient temperature. The coefficient of thermal expansion of the steel structure; This represents the initial distance between the sensor and the monitoring point. The reference diameter for sensor installation is given; the formula for calculating the position compensation is: .

[0009] According to the above scheme, the computing unit includes an edge computing chip and a lightweight fusion algorithm module; the edge computing chip is the hardware carrier for deploying the algorithm; the lightweight fusion algorithm module is used to process multi-source sensor data, extract steel structure deformation features and sensor position offset patterns using a fusion algorithm that combines an improved long short-term memory network with an adaptive Kalman filter, and establish and dynamically update the positioning calibration model.

[0010] Furthermore, the computing unit also includes an algorithm hardware coordination module and a positioning logic control module; The algorithm hardware collaboration module is used to coordinate with the fusion algorithm and the benchmark template matching and particle filter algorithm to calculate the error compensation value based on the positioning benchmark data to correct the sensor coordinates. It adopts vibration-temperature dual-thread data fusion for structural dynamic deformation, updates the positioning calibration model in real time, and repositions the sensor when the sensor offset exceeds the threshold through template matching and motion trend prediction strategies. The positioning logic control module is used to dynamically update the calibration model based on the positioning deviation threshold, calculate the overlap between the actual positioning coordinates and the theoretical monitoring coordinates according to the overlap calculation method, calculate the positioning reliability using the positioning reliability calculation formula and judge the positioning stability. When the positioning reliability is less than the preset reliability threshold and the overlap is greater than the preset overlap threshold, the recalibration process is triggered.

[0011] According to the above scheme, the mechanical and power supply unit includes a micro-displacement drive system, a positioning actuator, a power management module, and an emergency power supply module; The micro-displacement drive system is used to drive the sensor to perform three-axis micro-displacement adjustment; The positioning actuator is used to mount sensors and perform mechanical locking after positioning; The power management module is used to regulate and distribute electrical energy, and supports switching between solar and lithium battery power supply. The emergency power supply module is used to ensure the complete execution of the positioning calibration process in the event of a sudden power outage.

[0012] According to the above scheme, the communication and storage unit includes an industrial communication module, a hierarchical storage module, and a reference template management module; The industrial communication module integrates LoRa wireless communication and Ethernet communication for real-time transmission of monitoring data and positioning parameters; The hierarchical storage module is used to store historical monitoring data and positioning calibration records, encrypts and archives them, and stores them hierarchically based on data importance, prioritizing the retention of positioning data and calibration logs from key monitoring areas; The benchmark template management module is used to dynamically update the benchmark template based on the positioning calibration results, and to realize rapid repositioning and calibration of sensors based on the historical template library.

[0013] According to the above scheme, the adaptive execution unit includes a position fine-tuning module and an attitude calibration module; The micro-displacement adjustment module is used to receive control commands from the computing unit and precisely adjust the three-axis micro-displacement of the sensor through a piezoelectric ceramic drive. The attitude calibration module is used to adjust the sensor installation tilt angle based on the data fed back by the tilt sensor, so that the sensor is perpendicular to the monitoring point.

[0014] An adaptive positioning method for sensors used in steel structure health monitoring includes the following steps: S101: Acquire steel structure environmental data, structural status data, and sensor position and attitude data; S102: Based on the improved convolutional neural network model, process the data obtained in step S101, identify the steel structure monitoring points and complete the initial sensor positioning; use a multi-source data fusion algorithm to extract structural features and environmental features, establish a positioning error prediction model and generate a positioning template; S103: The servo adjustment system is controlled by the position compensation formula to dynamically adjust the position and attitude of the sensor and correct the positioning deviation caused by environmental interference. When the positioning error is less than the preset positioning error threshold, the sensing data and the internal status data of the device are periodically synchronized, the data is stored and the positioning template is updated. When the positioning error exceeds the preset positioning error threshold, the Kalman filter algorithm is used to select the optimal positioning coordinates and re-initialize the positioning. When environmental factors cause positioning drift, the positioning accuracy is maintained and the positioning status is preserved through a temperature-vibration dual-thread compensation mechanism. S104: Based on the sensor position and attitude adjustment results in step S103, the positioning data is output in real time through an improved convolutional neural network model, and the positioning reliability between the actual monitoring area and the theoretical monitoring area is calculated. When the location confidence level is greater than the preset confidence level threshold, the sensor can stably locate the monitoring point. If the location reliability remains below the threshold or the monitoring point location shifts, a secondary positioning adjustment is performed based on the structural deformation data and positioning deviation data.

[0015] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the adaptive positioning method for a steel structure health monitoring sensor as described in claim 9.

[0016] The beneficial effects of this invention are as follows: 1. The present invention provides an adaptive positioning device and method for a steel structure health monitoring sensor. This device collects environmental, structural, and positional attitude data from multiple dimensions through a data sensing unit. It combines an improved long short-term memory network with an adaptive Kalman filter fusion algorithm and a dual-thread data fusion strategy in the computing unit to achieve precise positioning calibration and dynamic model updates. It is further equipped with a mechanical and power supply unit featuring micro-displacement drive, mechanical locking, and dual power supply, supplemented by a communication and storage unit with encrypted hierarchical storage and multi-mode communication. Finally, it is linked with an adaptive execution unit that drives three-axis micro-adjustment and tilt angle adaptive calibration via piezoelectric ceramic, thus achieving efficient positioning and sensor calibration.

[0017] 2. This invention effectively improves the sensor's positioning accuracy and resistance to environmental interference, ensures positioning stability under dynamic deformation, and guarantees reliable monitoring data and continuous and controllable positioning process.

[0018] 3. This invention solves the problems of low calibration efficiency and insufficient real-time performance in the prior art, supports rapid repositioning and calibration, and improves the overall efficiency of steel structure health monitoring.

[0019] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a structural schematic diagram of an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the adaptive positioning device according to an embodiment of the present invention.

[0023] Figure 3 This is a flowchart of an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the adaptive positioning method according to an embodiment of the present invention.

[0025] Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0027] Example 1 See Figure 1 This embodiment is used to implement the principle of the above method embodiment to construct an adaptive positioning device 10 for a steel structure health monitoring sensor, including a data sensing unit 100, a computing unit 200, a mechanical and power supply unit 300, a communication and storage unit 400, and an adaptive execution unit 500; the data sensing unit is connected to the computing unit, the computing unit is connected to the communication and storage unit and the adaptive execution unit respectively, and the mechanical and power supply unit supplies power to all units; The data sensing unit 100 is used to collect environmental data, structural status data, and sensor position and attitude data of the steel structure. The data sensing unit 100 includes an environmental and structural data acquisition module and a position and attitude acquisition module. The environmental and structural data acquisition module is used to collect surface temperature, humidity, vibration amplitude, and strain data of the steel structure to provide environmental correction basis for positioning accuracy compensation. The position and attitude acquisition module is used to collect GPS coordinates, tilt angle, and displacement data of the sensors in real time, and calculate the position compensation amount in combination with installation parameters to correct the positioning deviation caused by environmental interference.

[0028] This embodiment provides comprehensive data support for positioning calibration and directly corrects positioning deviations for environmental interference, effectively improving the accuracy and anti-interference capability of sensor positioning, and laying a reliable data foundation for the calibration optimization of the computing unit and the positioning adjustment of the adaptive execution unit.

[0029] The formula for calculating the position compensation amount is:

[0030] in, This is the location compensation value; This refers to the sensor sensitivity coefficient; This refers to the vibration amplitude of the steel structure. Install the tilt angle for the sensor; This refers to the change in ambient temperature. The coefficient of thermal expansion of the steel structure; This represents the initial distance between the sensor and the monitoring point. Set the reference diameter for the sensor.

[0031] This embodiment performs quantitative fusion calculations on the acquired sensor position and attitude data with environmental and structural data to accurately obtain position compensation values. It quantifies the positioning deviations caused by factors such as environmental interference, structural deformation, and installation parameters, providing accurate quantitative basis for sensor positioning calibration. It effectively corrects positioning errors caused by dynamic changes in the environment and structure, significantly improves sensor positioning accuracy, and provides reliable data support for the computing unit to build a positioning calibration model and for the adaptive execution unit to carry out positioning optimization.

[0032] The calculation unit 200 is used to calibrate the sensor positioning based on the steel structure environment data, structural status data and sensor position and attitude data. At the same time, it dynamically adjusts the logic control and links the adaptive execution unit to complete the positioning optimization. The computing unit 200 includes an edge computing chip and a lightweight fusion algorithm module. The edge computing chip serves as the hardware carrier for algorithm deployment and reduces resource consumption through algorithm pruning and parallel computing optimization. The lightweight fusion algorithm module uses a fusion algorithm that combines an improved long short-term memory network with an adaptive Kalman filter to process multi-source sensor data, extract steel structure deformation characteristics and sensor position offset patterns, establish a positioning calibration model, and update it dynamically.

[0033] In this embodiment, the edge computing chip serves as the hardware carrier for algorithm deployment. Through algorithm pruning and parallel computing optimization, it reduces resource consumption and provides hardware support for the efficient operation of the algorithm. The lightweight fusion algorithm module adopts a fusion algorithm that combines an improved long short-term memory network with an adaptive Kalman filter. It efficiently processes multi-source sensor data and extracts the deformation characteristics of the steel structure and the sensor position offset patterns. It establishes and dynamically updates the positioning calibration model, which not only ensures the efficiency and economy of the algorithm operation, but also achieves accurate processing of multi-source data and real-time optimization of the calibration model. This provides accurate algorithmic support for sensor positioning calibration and significantly improves the accuracy and real-time performance of positioning calibration.

[0034] For example, taking the health monitoring of a large-span steel structure factory as an example, the lightweight fusion algorithm module, supported by the edge computing chip, accesses environmental and structural data such as surface temperature, vibration amplitude, and strain of the steel structure collected by the data sensing unit, as well as position and attitude data such as GPS coordinates and tilt angles of the sensors. First, it mines the temporal variation characteristics of vibration and strain data by improving the long short-term memory network, accurately capturing the long-term laws of steel structure deformation and sensor position offset caused by equipment operation and temperature fluctuations. Then, it uses an adaptive Kalman filter to filter out the interference of environmental noise on GPS coordinate and tilt angle data in real time. Finally, it fuses the results of the two to establish a positioning calibration model. When the factory experiences instantaneous strong vibrations due to the start-up and shutdown of heavy equipment or thermal expansion and contraction caused by diurnal temperature differences, the algorithm module rapidly iterates and updates the model parameters at the edge based on the lightweight optimization characteristics, instantly outputting accurate position compensation (positioning accuracy is improved by more than 40% compared to a single algorithm, reaching the centimeter level), providing a reliable basis for the fine-tuning calibration of the adaptive execution unit. At the same time, because the model is optimized by pruning, resource consumption is reduced by 60%, ensuring real-time response under complex working conditions.

[0035] The computing unit 200 also includes an algorithm hardware coordination module and a positioning logic control module. The algorithm hardware coordination module is used to coordinate with the fusion algorithm and the reference template matching and particle filtering algorithm to calculate the error compensation value based on the positioning reference data to correct the sensor coordinates. It also uses vibration-temperature dual-thread data fusion to deal with the dynamic deformation of the structure, updates the positioning calibration model in real time, and repositions the sensor when the sensor offset exceeds the threshold through template matching and motion trend prediction strategies. The positioning logic control module is used to dynamically update the calibration model according to the positioning deviation threshold, calculate the overlap between the actual positioning coordinates and the theoretical monitoring coordinates according to the overlap calculation method, calculate the positioning reliability using the positioning reliability calculation formula, and judge the positioning stability. When the positioning reliability is less than the preset reliability threshold and the overlap is greater than the preset overlap threshold, the recalibration process is triggered.

[0036] In this embodiment, the algorithm hardware collaboration module uses a fusion algorithm and a benchmark template matching, along with a particle filtering algorithm, to collaboratively correct sensor coordinates. It combines vibration-temperature dual-thread data fusion to handle dynamic deformation and update the model, achieving repositioning when the deviation exceeds a threshold. The positioning logic control module updates the model based on the deviation threshold, judges stability by the overlap and positioning reliability, and triggers recalibration. It performs precise error correction and model optimization under dynamic deformation to ensure positioning reliability. Through threshold handling and recalibration, it ensures that the deviation is controllable, significantly improving positioning accuracy and dynamic adaptability, and providing stable and accurate positioning support for steel structure health monitoring.

[0037] For example, taking the health monitoring scenario of a long-span steel bridge as an example, the positioning logic control module first presets a positioning overlap threshold of 85% and a positioning reliability threshold of 90%. It receives the actual sensor positioning coordinates and theoretical monitoring coordinates output by the computing unit in real time, calculates the overlap using the overlap calculation method, and obtains the reliability value using the positioning reliability calculation formula. When the bridge experiences thermal deformation due to temperature differences between day and night, coupled with vibrations from vehicle traffic, causing fluctuations in sensor positioning, and the module calculates an overlap of 92% (above the preset threshold) and a positioning reliability of 88% (below the preset threshold), it immediately triggers the recalibration process. Simultaneously, it dynamically updates the positioning calibration model based on the real-time positioning deviation threshold. This process accurately judges positioning stability through dual indicators, avoiding false or missed calibration triggers, and ensures that the sensor maintains high-precision positioning amidst dynamic structural changes through dynamic model updates and timely recalibration, providing reliable positioning assurance for accurate monitoring of the bridge's health status.

[0038] The mechanical and power supply unit 300 provides physical fixation support, servo adjustment capability, and stable energy supply for the device. The mechanical and power supply unit 300 includes a micro-displacement drive system, a positioning actuator, a power management module, and an emergency power supply module. The micro-displacement drive system drives the sensor to perform three-axis micro-displacement adjustments, adapting to dynamic deformation compensation of the steel structure. The positioning actuator mounts the sensor and mechanically locks it after positioning to prevent secondary displacement caused by vibration. The power management module regulates and distributes electrical energy, supporting dual power supply switching between solar and lithium batteries to ensure safe and efficient circuit operation. The emergency power supply module is used to cope with sudden power outages, ensuring the complete execution of the positioning calibration process.

[0039] This embodiment uses a micro-displacement drive system to drive the sensor to adjust its three-axis micro-displacement to adapt to the dynamic deformation compensation of the steel structure. The positioning actuator is equipped with a sensor and is mechanically locked after positioning to prevent vibration and secondary offset. The power management module regulates and distributes power and supports dual power supply switching between solar and lithium batteries to ensure circuit safety and efficiency. The emergency power supply module responds to sudden power outages to ensure the integrity of the positioning calibration process, improves positioning accuracy and stability, and ensures continuous energy supply through dual power supply and emergency design. This ensures that the positioning calibration of the device is continuous and controllable under complex working conditions, laying a solid mechanical and energy foundation for the health monitoring of steel structures.

[0040] For example, taking the health monitoring scenario of a blast furnace body in a large steel structure smelting plant as an example, the high temperature during blast furnace smelting causes thermal expansion and contraction deformation of the furnace body steel structure. At the same time, changes in internal material load will cause minute vibrations and displacements of the furnace body. The data sensing unit collects the surface temperature, strain, and sensor position offset data of the furnace body in real time and transmits them to the computing unit. After analysis, the computing unit issues precise adjustment commands to the micro-displacement drive system. This system quickly drives the sensors installed in the key weld areas of the furnace body to perform micro-displacement compensation adjustments along the X, Y, and Z axes. This accurately corrects the sensor and monitoring point offset deviations caused by the thermal deformation and vibration of the furnace body, ensuring that the sensors are always in close contact with the monitoring area and maintain the optimal acquisition posture. This effectively avoids the distortion of key monitoring data such as weld strain and temperature caused by the dynamic deformation of the furnace body, providing high-precision data support for fatigue damage assessment and safety early warning of the blast furnace body structure.

[0041] The communication and storage unit 400 is used for real-time transmission of positioning data and storage and encrypted archiving of historical monitoring data and positioning parameters. The communication and storage unit 400 includes an industrial communication module, a hierarchical storage module, and a reference template management module. The industrial communication module is used to integrate LoRa wireless communication and Ethernet communication to transmit monitoring data and positioning parameters in real time. The hierarchical storage module is used to store historical monitoring data and positioning calibration records, encrypt and archive them, and store them hierarchically based on data importance, prioritizing the retention of positioning data and calibration logs in key monitoring areas. The reference template management module is used to dynamically update the reference template according to the positioning calibration results and realize rapid repositioning and calibration of sensors based on the historical template library.

[0042] This embodiment integrates LoRa wireless communication and Ethernet communication through an industrial communication module for real-time transmission of monitoring data and positioning parameters. The hierarchical storage module encrypts and archives historical data and calibration records, storing them hierarchically according to importance, prioritizing the retention of data and logs from key areas. The benchmark template management module dynamically updates templates based on calibration results and utilizes a historical template library to achieve rapid sensor repositioning and calibration. Encrypted hierarchical storage ensures data security and efficient storage, while dynamic updates and reuse of benchmark templates improve the efficiency of sensor repositioning and calibration. This provides stable support for data flow, retention, and subsequent calibration optimization in steel structure health monitoring, ensuring the continuous and efficient conduct of monitoring work.

[0043] For example, taking the health monitoring scenario of the roof truss of a large steel structure stadium as an example, during initial deployment, the benchmark template management module records the coordinates, tilt angle, and environmental parameters after the initial positioning and calibration of the sensor, generating an initial benchmark template and storing it in the historical template library. When seasonal temperature differences cause thermal deformation of the truss or the load of crowds during events causes structural displacement, after positioning and calibration, the module will dynamically update the benchmark template based on the new results to ensure matching with the structural state. If a sensor at a monitoring point fails and needs to be replaced, the module can directly call the benchmark template in the historical template library for that point and quickly complete the repositioning and calibration by combining it with the current environmental data. This shortens the original calibration process, which required 4 hours, to 30 minutes, improving calibration efficiency and ensuring the consistency between the positioning accuracy of the new sensor and historical data, providing efficient and accurate template support for long-term health monitoring of the truss.

[0044] The adaptive execution unit 500 is used to fine-tune the sensor position and calibrate its attitude according to the control commands from the computing unit. The adaptive execution unit 500 includes a position fine-tuning module and an attitude calibration module; the micro-displacement adjustment module is used to receive control commands from the computing unit and precisely adjust the three-axis micro-displacement of the sensor through a piezoelectric ceramic drive; the attitude calibration module automatically adjusts the sensor mounting tilt angle based on feedback data from the tilt sensor, so that the sensor is perpendicularly attached to the monitoring point.

[0045] In this embodiment, the position fine-tuning module receives control commands from the computing unit and precisely adjusts the three-axis micro-displacement of the sensor via piezoelectric ceramic drive. The attitude calibration module relies on feedback data from the tilt sensor to automatically adjust the sensor's installation tilt angle to ensure it is perpendicular to the monitoring point. This not only corrects positional deviations caused by dynamic deformation through high-precision micro-displacement adjustment but also ensures that the sensor is always in the optimal acquisition posture through attitude calibration. This effectively avoids data distortion caused by positional offset or attitude tilt, significantly improving sensor positioning accuracy and data acquisition reliability, and providing solid execution support for accurate monitoring of the health status of steel structures.

[0046] For example, taking the health monitoring scenario of key nodes in the tower body of a large steel structure transmission tower of a high-voltage transmission line as an example, the transmission tower may experience slight torsion and tilting due to continuous strong winds at high altitudes and diurnal temperature differences. This causes the tilt angle of the monitoring sensors installed at the nodes to shift. At this time, the attitude calibration module receives the offset data fed back by the tilt sensor in real time, quickly activates the adjustment mechanism to automatically adjust the installation tilt angle of the sensor, and accurately corrects the offset to ensure that the sensor detection surface is always perpendicular to the surface of the monitoring node. Without attitude calibration, sensor tilt can lead to errors of more than 20% in monitoring data such as strain and vibration. After module calibration, the data error can be controlled within 5%, effectively avoiding the distortion of monitoring data caused by attitude tilt, and providing high-precision data support for the fatigue state assessment and wind vibration damage early warning of the transmission tower structure.

[0047] This embodiment acquires environmental, structural, and positional attitude data from multiple dimensions through a data sensing unit. Combined with an improved long short-term memory network and adaptive Kalman filter fusion algorithm and a dual-thread data fusion strategy in the computing unit, it achieves precise positioning calibration and dynamic model updates. This is further enhanced by a mechanical and power supply unit with micro-displacement drive, mechanical locking, and dual power supply, supplemented by a communication and storage unit with encrypted hierarchical storage and multi-mode communication. Finally, it is linked with an adaptive execution unit that drives three-axis micro-adjustment and tilt adaptive calibration via piezoelectric ceramics. This effectively improves sensor positioning accuracy and resistance to environmental interference, ensures positioning stability under dynamic deformation, guarantees reliable monitoring data, and ensures a continuous and controllable positioning process. It also supports rapid repositioning and calibration, improving the overall efficiency of steel structure health monitoring. This solves the problems of low calibration efficiency and insufficient real-time performance in existing technologies.

[0048] Example 2 The structure and principle of this embodiment are the same as those of Embodiment 1, the difference being in its application to a specific instance. For example... Figure 2 As shown, it specifically includes: Taking the health monitoring of a long-span steel truss suspension bridge as an application scenario, this bridge has a main span of 500m. The steel truss is made of Q690 high-strength steel. The bridge site is located in a subtropical monsoon climate zone, which is constantly affected by high temperature and humidity, typhoons and rainstorms, and heavy vehicle impacts. The mid-span, support, and node areas of the steel truss are prone to dynamic deformation and stress concentration. Traditional sensor positioning methods suffer from a data distortion rate of over 25% due to environmental interference and structural deformation. To solve this problem, an adaptive positioning device based on sensors used for steel structure health monitoring was deployed. The device adopts a modular design, and each unit can be quickly assembled through industrial-grade connectors. The core deployment areas are the bottom of the mid-span of the steel truss, the top of the side span support, and the main truss nodes. One complete device is configured at each monitoring point, with a total of 24 devices deployed throughout the bridge, achieving full bridge coverage through a distributed architecture. The data sensing unit is installed close to the surface of the steel truss, the computing unit is integrated into a waterproof and dustproof stainless steel control box, the mechanical and power supply units are fixed to the embedded parts of the steel truss by expansion bolts, the communication and storage units are connected to the existing communication network of the bridge, and the adaptive execution unit and the sensor form an integrated installation structure. The overall protection level reaches IP67, which is suitable for the complex service environment of the bridge.

[0049] The data sensing unit adopts a "multi-module collaborative acquisition + precise synchronization" design to ensure the comprehensiveness and timeliness of data acquisition. In the environmental and structural data acquisition modules, temperature acquisition uses a PT100 platinum resistance sensor, installed on the lower flange surface of the steel truss, and fixed with high-temperature thermally conductive adhesive. The measurement range is -50℃ to 200℃, with an accuracy of ±0.1℃. Humidity acquisition uses an SHT30 digital humidity sensor, integrated inside the temperature sensor housing, achieving simultaneous temperature and humidity acquisition. The humidity measurement range is 0% to 100%RH, with an accuracy of ±2%RH. Vibration data acquisition uses an IEPE accelerometer, installed at the mid-span node of the steel truss, fixed by magnetic adsorption, with a sensitivity of 100mV / g, a measurement range of ±5g, and a sampling frequency of 100Hz. Strain data is acquired using a BF120-3AA strain gauge, adhered to the tension zone of the steel truss, using a semi-bridging method, with a sensitivity of 2.0±1%, in conjunction with a DH3816 static strain gauge. The position and attitude acquisition module uses the UBLOXNEO-7MGPS module, with a positioning accuracy of ±1m. It is fixed to the top of the sensor housing with a bracket to ensure stable satellite signal reception. Tilt acquisition uses an MMA8452 triaxial tilt sensor, integrated inside the sensor mounting base, with a measurement range of ±180° and an accuracy of ±0.1°. Displacement acquisition uses a KeyenceIL-300 laser displacement sensor, mounted opposite each other on the fixed support of the steel truss, with a measurement distance of 50~300mm and an accuracy of ±0.01mm. For data synchronization, each acquisition module is connected to the main controller via an SPI bus, and synchronous sampling at a frequency of 10Hz is achieved by an STM32H743 microcontroller. When calculating the position compensation, the sensor sensitivity coefficient is determined to be 0.85, and the thermal expansion coefficient of the steel structure is 1.2×10⁻⁶, taking into account the bridge design parameters. -5 With an ambient temperature of 5℃, an initial distance of 0.5m, and an installation reference diameter of 0.3m, when the ambient temperature changes by 5℃, the vibration amplitude is 0.2mm, and the installation tilt angle shifts by 0.5°, the calculated position compensation value is 0.12mm, providing accurate data support for subsequent calibration.

[0050] The computing unit employs a collaborative design of "hardware optimization + algorithm iteration" to ensure efficient and accurate data processing. The edge computing chip utilizes the NVIDIA Jetson Nano developer kit, which integrates a quad-core ARM Cortex-A57 CPU and a 128-core NVIDIA Maxwell GPU, supporting CUDA parallel computing. To reduce resource consumption, the algorithm was pruned and optimized: the number of hidden layer neurons in the improved Long Short-Term Memory network was reduced from 256 to 128, redundant convolutional kernels were removed, reducing the model size by 60%; and the multi-source data processing task was decomposed into three threads—environmental data preprocessing, location data fusion, and calibration model update—through GPU parallel computing, reducing processing latency from 500ms to 150ms. In the lightweight fusion algorithm module, an improved long short-term memory network is introduced with an attention mechanism to assign higher weights to key features of time-series data such as vibration and strain. For example, during periods of heavy vehicle traffic, the weight of vibration data is increased to 0.6, accurately capturing the dynamic deformation pattern of the steel truss girder at mid-span of 1.2mm. Adaptive Kalman filtering filters out random noise in GPS signals by updating the noise covariance matrix of the state equation and observation equation in real time (the initial value of the Q matrix is ​​set to diag[0.01,0.01,0.01], and the R matrix is ​​adjusted in real time according to the data noise), improving the coordinate positioning accuracy from ±1m to ±0.5m. The algorithm hardware collaboration module performs secondary optimization on the fused data using a particle filtering algorithm with 500 particles. Based on the benchmark coordinate data of the bridge design, error compensation values ​​are calculated. When the steel truss girder undergoes a thermal expansion deformation of 0.8mm due to temperature changes, the calibration model is updated in real time through vibration-temperature dual-thread data fusion. The positioning logic control module has a preset overlap threshold of 85% and a positioning reliability threshold of 90%. When typhoon weather causes the actual positioning coordinates of the sensor to overlap with the theoretical coordinates by 92% and the reliability to be 88%, the recalibration process is immediately triggered to ensure positioning stability.

[0051] The mechanical and power supply units adopt a "precise drive + stable power supply" design concept to provide reliable protection for the operation of the device. The micro-displacement drive system uses a PZT-5H piezoelectric ceramic actuator, which has a displacement resolution of 0.01μm and a maximum output force of 500N. The sensor is mounted on a three-axis displacement platform with cross roller guides, achieving a repeatability of ±0.1μm. When the calculation unit issues a compensation command of 0.12mm, the actuator achieves synchronous adjustment of the three axes through voltage closed-loop control (control accuracy ±0.01V), adapting to the dynamic deformation of the steel truss. The positioning actuator uses an electromagnetic locking device. After the sensor completes positioning, the electromagnetic lock is driven by DC24V voltage to generate a locking force of 1000N to fix the sensor in the current position. After locking, the displacement deviation is less than 0.05mm, effectively preventing secondary displacement caused by vehicle vibration. The power management module employs a collaborative design between an MPPT solar controller and a lithium battery management system (BMS). It is equipped with two 100W monocrystalline silicon solar panels (23% conversion efficiency) and two 12V / 100Ah lithium iron phosphate batteries. When the light intensity is ≥20000 lux, the solar panels power the device and charge the lithium batteries. When light intensity is insufficient, the BMS automatically switches to lithium battery power, maintaining a stable supply voltage of DC24V±0.5V with a conversion efficiency of 95%. The emergency power supply module uses a 12V / 50Ah backup lithium battery. A voltage monitoring module monitors the main power status in real time. When the main power fails, the backup battery automatically switches power within 50ms, ensuring the complete execution of the positioning calibration process and providing up to 8 hours of runtime.

[0052] The adaptive execution unit, together with the communication and storage unit, forms a closed-loop system of "high-efficiency execution + data flow," improving monitoring efficiency. Within the adaptive execution unit, the position fine-tuning module is linked with the micro-displacement drive system. It receives PWM control signals from the computing unit and adjusts the extension and retraction of the piezoelectric ceramic actuator using a PID algorithm. For example, when the steel truss beam experiences a 0.15mm displacement at mid-span, the module completes precise fine-tuning within 200ms. The attitude calibration module provides real-time data feedback through the MMA8452 tilt sensor. When the sensor tilt angle deviates by 2°, the module drives the adjustment bracket via a stepper motor. The stepper motor has a step angle of 1.8°, a reduction ratio of 1:100, and an adjustment accuracy of ±0.05°, ensuring that the sensor and monitoring point are perpendicularly aligned (verticality error ≤0.1°), avoiding strain data errors caused by attitude misalignment (error 20% before calibration, error ≤5% after calibration). In the communication and storage unit, the industrial communication module integrates an SX1278 LoRa module and an industrial Ethernet switch. The LoRa module has a communication distance of up to 3km, uses spread spectrum communication technology (spread spectrum gain 18dB), and has strong anti-interference capabilities, used for transmitting real-time monitoring data. The Ethernet switch connects to the bridge monitoring center via fiber optic cable, with a transmission rate of 100Mbps, used for transmitting large-capacity historical data. The hierarchical storage module uses a 128GB industrial-grade SD card and a 1TB solid-state drive. The SD card stores data for ordinary monitoring areas (storage period of 30 days), and the solid-state drive stores data for critical areas (storage period of 1 year). Data is encrypted using the AES-256 encryption algorithm before archiving. The benchmark template management module initially establishes benchmark templates for 24 monitoring points (including coordinates, tilt angle, environmental parameters, etc.). The templates are dynamically updated every 7 days based on calibration results. When a sensor at a monitoring point fails and needs to be replaced, the module calls up data from the historical template library and, combined with the current environmental parameters, completes rapid repositioning calibration within 30 minutes, saving 87.5% of the time compared to traditional calibration processes. Since its operation six months ago, the device has maintained a sensor positioning accuracy of ±0.5mm under complex conditions such as high temperature, heavy rain, and typhoons, and the accuracy of monitoring data has reached over 95%, providing reliable technical support for health assessment and safety early warning of long-span steel structure bridges.

[0053] In summary, this application's embodiments, through modular design and multi-unit collaborative operation, combined with precise multi-sensor acquisition, deep integration of edge computing chips and optimization algorithms, and micro-displacement drive and stable power supply, can accurately achieve sensor positioning accuracy of ±0.5mm in key areas such as the mid-span and supports of steel truss beams. The monitoring data accuracy is improved to over 95%, reducing data distortion by more than 25% compared to traditional solutions. Multi-source data processing and calibration results can be completed within 150ms. During dynamic operation, adaptive Kalman filtering and particle filtering algorithms compensate for environmental and structural deformation effects. Combined with electromagnetic locking to prevent secondary offset and seamless emergency power supply switching mechanisms, it can still operate stably under complex conditions such as high temperature and humidity, typhoons and heavy vehicle loads. Sensor positioning deviation is controlled within ±0.5mm, and repositioning calibration can be completed within 30 minutes after fault replacement, saving 87.5% of time compared to traditional processes. This device effectively solves the problems of inaccurate positioning, data distortion, and weak anti-interference ability caused by environmental interference and structural deformation of traditional sensors through multi-unit collaboration and deep adaptation of algorithms and hardware, providing efficient and reliable technical support for health assessment and safety early warning of long-span steel truss suspension bridges.

[0054] Example 3 See Figure 3 The specific steps of an adaptive positioning method for a steel structure health monitoring sensor are as follows: S101: Acquire steel structure environmental data, structural status data, and sensor position and attitude data; This embodiment acquires environmental data, structural state data, and sensor position and attitude data of the steel structure. It can accurately capture the dynamic deformation and stress concentration characteristics of the steel structure caused by environmental factors and loads such as high temperature and humidity and heavy vehicle loads. It can also monitor the positioning attitude deviation of the sensors in real time. This provides comprehensive and accurate raw data support for the multi-source data fusion and adaptive calibration algorithm operation of the subsequent edge computing unit. It avoids positioning calibration deviations caused by missing or inaccurate data. It ensures the positioning accuracy of subsequent sensors and the reliability of monitoring data from the data source. This lays a key foundation for solving the problem of monitoring data distortion caused by environmental interference and structural deformation in traditional solutions.

[0055] S102: Based on the improved convolutional neural network model, the steel structure environmental data, structural state data and sensor position and attitude data are processed. After identifying the steel structure monitoring points, the initial sensor positioning is completed. The multi-source data fusion algorithm is used to extract structural features and environmental features, establish a positioning error prediction model and generate a positioning template. Multi-source data fusion algorithms are used in scenarios such as steel structure health monitoring to integrate data from different sources, such as environmental parameters, structural status, and sensor attitude. By removing redundancy and filling in missing information, they achieve information complementarity, thereby improving data reliability and positioning or monitoring accuracy.

[0056] This embodiment utilizes a multi-source data fusion algorithm to integrate multi-source information such as environmental parameters, structural state, and sensor attitude after processing by an improved CNN. By removing redundancy and filling in missing information, it achieves information complementarity, accurately extracts key structural and environmental features, and provides high-quality support for the establishment of a positioning error prediction model and the generation of positioning templates. At the same time, it solves the problems of large interference and one-sided information in single-source data, improves data reliability and feature representation capabilities, directly ensures the initial positioning accuracy of the sensor, and lays a solid foundation for subsequent positioning calibration.

[0057] S103: The servo adjustment system is controlled by a position compensation formula to dynamically adjust the sensor position and attitude, correcting positioning deviations caused by environmental interference. Simultaneously, when the positioning error is less than a preset positioning error threshold, the sensing data and internal device status data are periodically synchronized, stored, and the positioning template is updated. When the positioning error exceeds the preset positioning error threshold, a Kalman filter algorithm is used to select the optimal positioning coordinates and re-initialize the positioning. When environmental factors cause positioning drift, a temperature-vibration dual-thread compensation mechanism is used to maintain positioning accuracy and preserve the positioning status. The position compensation formula is a mathematical expression used in steel structure health monitoring to calculate the positioning deviation compensation value in order to correct the sensor position and improve positioning accuracy, by combining environmental parameters, structural state data and sensor attitude information.

[0058] This embodiment utilizes a position compensation formula to accurately calculate the positioning deviation compensation value by combining environmental parameters, structural state data, and sensor attitude information. It controls the servo adjustment system to dynamically adjust the sensor position and attitude to correct positioning deviations caused by environmental interference. At the same time, it provides core data support for positioning error threshold judgment and temperature-vibration dual-thread compensation to deal with positioning drift, ensuring the effectiveness of data synchronization, template updates, and re-initialization of positioning. It effectively solves the accuracy problem of traditional positioning affected by environmental interference and maintains the stability and reliability of sensor positioning.

[0059] S104: Based on the positioning status, the positioning data is output in real time through an improved convolutional neural network model, and the positioning reliability between the actual monitoring area and the theoretical monitoring area is calculated; when the positioning reliability is greater than the preset reliability threshold, the sensor stably positions the monitoring point; if the positioning reliability is continuously lower than the threshold or the monitoring point position is offset, a secondary positioning adjustment is performed based on the structural deformation data and positioning deviation data.

[0060] Improved convolutional neural network models are convolutional neural network models that are optimized based on traditional convolutional neural networks by adjusting the network structure (such as adding or removing convolutional layers, optimizing convolutional kernel parameters), incorporating attention mechanisms, etc., in order to improve feature extraction accuracy, reduce computational costs, or adapt to specific tasks (such as steel structure monitoring data processing).

[0061] This embodiment utilizes an improved convolutional neural network model, leveraging optimizations such as structural adjustment and the integration of attention mechanisms to achieve high-precision feature extraction and efficient data processing. It can output real-time steel structure monitoring and positioning data, and accurately calculate the positioning reliability of the actual and theoretical monitoring areas, providing a reliable basis for determining whether the sensor is stably positioned. Furthermore, when the positioning reliability remains below a threshold or the monitoring point deviates, the output precise positioning data provides high-quality data support for secondary positioning adjustments based on structural deformation and positioning deviation data. This effectively solves problems such as inaccurate feature extraction and low processing efficiency in traditional models, leading to distorted positioning judgments and delayed adjustments, thus ensuring the real-time performance and stability of positioning monitoring.

[0062] For example, in the scenario of detecting microcracks on the surface of steel structures, an improved convolutional neural network model integrating channel-spatial attention modules and depthwise separable convolutions is adopted. This addresses the problems of traditional CNNs' insufficient feature extraction for microcracks with widths less than 0.1 mm and the difficulty of embedded deployment due to redundant model parameters. Using ResNet-50 as the base encoder, the model strengthens the weights of crack grayscale difference features through channel attention modules and accurately locates defect areas through spatial attention modules. Simultaneously, the introduction of depthwise separable convolutions to replace standard convolutions reduces computational load by 85%. Tested on a dataset of 1500 images of steel structure cracks containing high-temperature oxidation and corrosion interference, the model achieved a detection accuracy of 95.2%, an improvement of 8.7% compared to the traditional VGG16, a recall rate of 91.5% for microcracks, and a single-image inference time of only 86 ms. This not only solves the problems of missed and false detections in traditional models but also adapts to the embedded real-time detection requirements of bridge steel structure health monitoring.

[0063] This embodiment proposes an adaptive positioning method for a steel structure health monitoring sensor. This method utilizes a data sensing unit to collect multi-dimensional environmental, structural, and positional attitude data. Combined with an improved long short-term memory network and adaptive Kalman filter fusion algorithm and a dual-thread data fusion strategy in the computing unit, it achieves precise positioning calibration and dynamic model updates. This is further enhanced by a mechanical and power supply unit with micro-displacement drive, mechanical locking, and dual power supply, along with a communication and storage unit featuring encrypted hierarchical storage and multi-mode communication. Finally, an adaptive execution unit with piezoelectric ceramic-driven three-axis micro-adjustment and tilt adaptive calibration works in conjunction with this method. This effectively improves the sensor's positioning accuracy and resistance to environmental interference, ensures positioning stability under dynamic deformation, guarantees reliable monitoring data, and maintains a continuous and controllable positioning process. It also supports rapid repositioning and calibration, improving the overall efficiency of steel structure health monitoring. This method solves the problems of low calibration efficiency and insufficient real-time performance in existing technologies.

[0064] Example 4 The steps in this embodiment are the same as in embodiment 3, the difference being that each step is applied to a specific instance. For example... Figure 4 As shown, the specific steps include: Taking the construction of a health monitoring system for a large-span steel truss bridge as an application scenario, the goal is to achieve adaptive positioning of fiber optic strain sensors at key monitoring points of the bridge's main truss, ensuring the accuracy and continuity of monitoring data. First, a multi-source data acquisition system was built. For environmental data acquisition, a PT100 platinum resistance temperature sensor (measurement range -50℃~200℃, accuracy ±0.1℃), an SHT30 humidity sensor (accuracy ±2%RH), and an FC-2A ultrasonic anemometer (wind speed accuracy ±0.1m / s) were used, deployed at three environmental monitoring nodes on the top of the main truss of the bridge and near the supports. For structural status data, an FBG-3000 series fiber optic strain sensor (range -1500με~1500με, accuracy ±1με) and an ADXL345 accelerometer (range ±16g, accuracy ±0.01g) were used, initially selecting 20 potential monitoring points. Sensor position and attitude data were acquired through a WT901C high-precision IMU inertial measurement unit (heading accuracy ±0.5°, acceleration accuracy ±0.001g), which was integrated with the sensor in the same package. The data acquisition frequency is set to 1Hz for environmental data, 100Hz for structural status data, and 50Hz for position and attitude data. The data is transmitted in real time to the edge computing node via a 4G industrial module, with the transmission latency controlled within 50ms.

[0065] In the initial localization stage, an improved convolutional neural network model was used to identify and locate monitoring points. The model was based on the VGG16 architecture, with the last three fully connected layers removed and one global average pooling layer and two 1×1 convolutional layers added, reducing the parameter scale while improving feature extraction efficiency. During data preprocessing, environmental data such as temperature, humidity, and wind speed were normalized to the [0,1] interval and fused with structural state data such as strain and acceleration, as well as three-dimensional position (x,y,z) and three-dimensional attitude (pitch angle, roll angle, yaw angle) data output by the IMU, into a 64×64×12 feature map as input. The model training dataset included simulated loading test data (5000 sets) of the 1:10 scale model of the bridge and measured data of similar bridges (3000 sets), with the labels being the coordinates of the monitoring points calibrated by a total station (accuracy ±0.1mm). After training, the model achieved a recognition accuracy of 98.2%. After outputting the initial positioning coordinates of the sensor, the DS evidence theory was used to fuse multi-source data: the mean and variance of the environmental data, the time domain peak value and frequency domain dominant frequency of the structural data were used as evidence. The environmental feature weight was set to 0.3 and the structural feature weight to 0.7. After fusion, 12-dimensional key features were extracted and input into the BP neural network to construct a positioning error prediction model. The model input was the fused features, and the output was the positioning error in the x, y, and z directions. After testing with 1000 sets of validation data, the error prediction accuracy reached 96.5%. Finally, a positioning template containing the optimal positioning coordinates and environmental adaptation parameters was generated.

[0066] The dynamic adjustment system uses an STM32H743 main control chip to control the servo adjustment mechanism. This mechanism consists of three stepper motors (positioning accuracy ±0.02mm) and two servo motors (angle accuracy ±0.1°), respectively realizing the position adjustment of the sensor in the x, y, and z directions, as well as the pitch and roll attitude adjustment. The position compensation formula is designed as: X'=X0+ΔX1+ΔX2, where X0 is the initial positioning coordinate, ΔX1 is the static error compensation amount output by the error prediction model, and ΔX2 is the real-time environmental deviation compensation amount. When the system detects that the positioning error (measured jointly by the IMU and laser rangefinder, accuracy ±0.05mm) is less than the preset threshold of 0.3mm, it synchronizes the sensor strain data and internal status data such as servo motor current and temperature every 5 minutes via the SPI interface, stores them on a local 128GB industrial-grade SD card, and uploads them to the cloud database via the 5G network. At the same time, it uses the sliding window method (window size 50 sets of data) to update the feature parameters of the positioning template. When the positioning error is greater than 0.3mm, the Kalman filter algorithm is activated: the state equation is set as X'. k =AX k-1 +BU k-1 +W k-1 The observation equation is Z k =HX k +V k The state matrix A contains position and attitude parameters, the observation matrix H is constructed from laser rangefinder observations, and the process noise Q and observation noise R are calibrated into diagonal matrices using offline data. After filtering, the optimal positioning coordinates are output and reinitialized. To address environmental drift, temperature compensation uses the linear fitting formula ΔT=k×(T-T0) (k is calibrated to 0.002mm / ℃ by high and low temperature tests). Vibration compensation uses adaptive filtering to extract vibration frequency components and generate a reverse compensation signal. This dual-thread collaboration keeps the environmental drift within 0.1mm.

[0067] In the location verification stage, the improved convolutional neural network model outputs location data in real time. The location reliability calculation adopts a cosine similarity and error weighted model: C=α×(X·X0) / (|X|×|X0|)+β×(1-|X-X0| / X) max ), where α=0.6 and β=0.4 are weighting coefficients, X is the real-time positioning coordinate, X0 is the theoretical monitoring point coordinate, and X maxThe maximum allowable positioning range is set to 5mm, with a preset confidence threshold of 0.8. When C > 0.8, the sensor is considered to be in stable positioning, and continuous monitoring mode is initiated. When C remains below 0.8 for 30 seconds or the displacement sensor detects a position shift of more than 0.5mm at the monitoring point, secondary positioning adjustment is triggered. During secondary positioning, structural deformation data is first collected by the strain sensor, and a corrected feature vector is constructed by combining it with the positioning deviation data output by the IMU. This vector is then input into a convolutional neural network model optimized through transfer learning (fine-tuned based on the initial model, with 2000 sets of structural deformation condition data added to the training data). The corrected positioning coordinates are output, and the servo system is controlled to perform fine-tuning. The adjustment range is set to 0.05~0.2mm depending on the deviation. After adjustment, the confidence is recalculated until the threshold requirement is met.

[0068] In summary, this application's embodiments achieve comprehensive real-time acquisition of environmental, structural, and sensor attitude data by building a multi-source high-precision data acquisition system. Combined with an improved convolutional neural network model and a multi-source data fusion algorithm, high-precision initial positioning of monitoring points is achieved. Then, a servo adjustment mechanism controlled by an STM32H743 microcontroller, along with Kalman filtering and environmental compensation strategies, dynamically corrects positioning errors and controls drift. Supplemented by a secondary positioning mechanism based on position confidence determination and transfer learning optimization, this effectively ensures the adaptive and accurate positioning of fiber optic strain sensors at key monitoring points. Positioning errors are stably controlled within a low threshold, and environmental drift is less than 0.1 mm, significantly improving the accuracy and continuity of monitoring data and providing reliable data support for the health assessment of long-span steel truss bridges.

[0069] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0070] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0071] Figure 5 This is a schematic diagram of the electronic device provided in this embodiment. This embodiment also includes a memory 501, a processor 502, and a computer program stored in the memory 501 and executable on the processor 502.

[0072] When the processor 502 executes the program, it implements an adaptive positioning method for a steel structure health monitoring sensor provided in the above embodiments.

[0073] Furthermore, electronic devices also include: Communication interface 503 is used for communication between memory 501 and processor 502.

[0074] The memory 501 is used to store computer programs that can run on the processor 502.

[0075] The memory 501 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0076] If the memory 501, processor 502, and communication interface 503 are implemented independently, then the communication interface 503, memory 501, and processor 502 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0077] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.

[0078] Processor 502 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of this application.

[0079] This embodiment also provides a computer-readable storage medium storing executable instructions that, when executed by a processor, enable the processor to implement an adaptive positioning method for a steel structure health monitoring sensor.

[0080] In the description of this specification, the references to "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0081] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0082] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0083] Furthermore, this application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] This application is described with reference to the flowchart of the method and computer program product according to Embodiment 1 and the block diagram of the device (system) according to Embodiment 3. It should be understood that each step or block in the flowchart or block diagram, as well as combinations of steps or blocks in the flowchart or block diagram, can be implemented by computer program instructions.

[0085] The various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0086] These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which are executable by the processor of the computer or other programmable data processing device, produce instructions for implementing the process. Figure 1 One or more processes or boxes Figure 1 An adaptive positioning device for a sensor used for health monitoring of steel structures, comprising one or more functions specified in the boxes.

[0087] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes or boxes Figure 1 The steps of an adaptive positioning method for a steel structure health monitoring sensor are specified in one or more boxes.

[0090] The above embodiments are only used to illustrate the design concept and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, all equivalent changes or modifications made based on the principles and design ideas disclosed in the present invention are within the protection scope of the present invention.

Claims

1. An adaptive positioning device for a steel structure health monitoring sensor, characterized in that: It includes a data sensing unit, a computing unit, a mechanical and power supply unit, a communication and storage unit, and an adaptive execution unit; the data sensing unit is connected to the computing unit; the computing unit is connected to the communication and storage unit and the adaptive execution unit respectively; the mechanical and power supply unit is used to supply power to all units; The data sensing unit is used to collect environmental data, structural status data, and sensor position and attitude data of the steel structure; The computing unit is used to perform positioning calibration of the sensor based on the data collected by the data sensing unit, while dynamically adjusting the logic control and linking with the adaptive execution unit to complete the positioning optimization. The mechanical and power supply unit is used to provide physical fixation support, servo adjustment capability, and stable power supply for the device; The communication and storage unit is used for real-time transmission of location data, as well as for storing and encrypting historical monitoring data and location parameters. The adaptive execution unit is used to fine-tune the position and calibrate the attitude of the sensor according to the control instructions of the computing unit.

2. The adaptive positioning device for a steel structure health monitoring sensor according to claim 1, characterized in that: The data sensing unit includes an environmental and structural data acquisition module and a position and attitude acquisition module; The environmental and structural data acquisition module is used to collect surface temperature, humidity, vibration amplitude and strain data of the steel structure, which serve as the basis for environmental correction for positioning accuracy compensation. The position and attitude acquisition module is used to acquire the GPS coordinates, tilt angle and displacement data of the sensor in real time. Combined with the installation parameters, the position compensation amount is calculated according to the position compensation amount calculation formula to correct the positioning deviation caused by environmental interference.

3. The adaptive positioning device for a steel structure health monitoring sensor according to claim 2, characterized in that: This is the location compensation value; This refers to the sensor sensitivity coefficient; This refers to the vibration amplitude of the steel structure. Install the tilt angle for the sensor; This refers to the change in ambient temperature. The coefficient of thermal expansion of the steel structure; This represents the initial distance between the sensor and the monitoring point. Install the reference diameter for the sensor; The formula for calculating the position compensation amount is: 。 4. The adaptive positioning device for a steel structure health monitoring sensor according to claim 1, characterized in that: The computing unit includes an edge computing chip and a lightweight fusion algorithm module; the edge computing chip is the hardware carrier for deploying the algorithm. The lightweight fusion algorithm module is used to process multi-source sensor data, extract steel structure deformation features and sensor position offset patterns using a fusion algorithm that combines an improved long short-term memory network with an adaptive Kalman filter, and establish and dynamically update the positioning calibration model.

5. The adaptive positioning device for a steel structure health monitoring sensor according to claim 4, characterized in that: The computing unit also includes an algorithm hardware coordination module and a positioning logic control module; The algorithm hardware collaboration module is used to coordinate with the fusion algorithm and the benchmark template matching and particle filter algorithm to calculate the error compensation value based on the positioning benchmark data to correct the sensor coordinates. It adopts vibration-temperature dual-thread data fusion for structural dynamic deformation, updates the positioning calibration model in real time, and repositions the sensor when the sensor offset exceeds the threshold through template matching and motion trend prediction strategies. The positioning logic control module is used to dynamically update the calibration model based on the positioning deviation threshold, calculate the overlap between the actual positioning coordinates and the theoretical monitoring coordinates according to the overlap calculation method, calculate the positioning reliability using the positioning reliability calculation formula and judge the positioning stability. When the positioning reliability is less than the preset reliability threshold and the overlap is greater than the preset overlap threshold, the recalibration process is triggered.

6. The adaptive positioning device for a steel structure health monitoring sensor according to claim 1, characterized in that: The mechanical and power supply unit includes a micro-displacement drive system, a positioning actuator, a power management module, and an emergency power supply module; The micro-displacement drive system is used to drive the sensor to perform three-axis micro-displacement adjustment; The positioning actuator is used to mount sensors and perform mechanical locking after positioning; The power management module is used to regulate and distribute electrical energy, and supports switching between solar and lithium battery power supply. The emergency power supply module is used to ensure the complete execution of the positioning calibration process in the event of a sudden power outage.

7. The adaptive positioning device for a steel structure health monitoring sensor according to claim 1, characterized in that: The communication and storage unit includes an industrial communication module, a hierarchical storage module, and a reference template management module; The industrial communication module integrates low-power local area network wireless standards and Ethernet communication for real-time transmission of monitoring data and positioning parameters; The hierarchical storage module is used to store historical monitoring data and positioning calibration records, encrypts and archives them, and stores them hierarchically based on data importance, prioritizing the retention of positioning data and calibration logs from key monitoring areas; The benchmark template management module is used to dynamically update the benchmark template based on the positioning calibration results, and to realize rapid repositioning and calibration of sensors based on the historical template library.

8. The adaptive positioning device for a steel structure health monitoring sensor according to claim 1, characterized in that: The adaptive execution unit includes a position fine-tuning module and an attitude calibration module; The micro-displacement adjustment module is used to receive control commands from the computing unit and precisely adjust the three-axis micro-displacement of the sensor through piezoelectric ceramic drive; The attitude calibration module is used to adjust the sensor installation tilt angle based on the data fed back by the tilt sensor, so that the sensor is perpendicular to the monitoring point.

9. An adaptive positioning method for a health monitoring sensor of steel structures, characterized in that: Includes the following steps: S101: Acquire steel structure environmental data, structural status data, and sensor position and attitude data; S102: Based on the improved convolutional neural network model, process the data obtained in step S101, identify the steel structure monitoring points and complete the initial sensor positioning; use a multi-source data fusion algorithm to extract structural features and environmental features, establish a positioning error prediction model and generate a positioning template; S103: The servo adjustment system is controlled by the position compensation formula to dynamically adjust the position and attitude of the sensor and correct the positioning deviation caused by environmental interference. When the positioning error is less than the preset positioning error threshold, the sensing data and the internal status data of the device are periodically synchronized, the data is stored and the positioning template is updated. When the positioning error exceeds the preset positioning error threshold, the Kalman filter algorithm is used to select the optimal positioning coordinates and re-initialize the positioning. When environmental factors cause positioning drift, the positioning accuracy is maintained and the positioning status is preserved through a temperature-vibration dual-thread compensation mechanism. S104: Based on the sensor position and attitude adjustment results in step S103, the positioning data is output in real time through an improved convolutional neural network model, and the positioning reliability between the actual monitoring area and the theoretical monitoring area is calculated. When the location confidence level is greater than the preset confidence level threshold, the sensor can stably locate the monitoring point. If the location reliability remains below the threshold or the monitoring point location shifts, a secondary positioning adjustment is performed based on the structural deformation data and positioning deviation data.

10. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the adaptive positioning method for a health monitoring sensor for steel structures as described in claim 9.

Images