Real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations
By coordinating distributed sensing nodes and intelligent data processing units, the problems of monitoring blind spots, inaccurate assessments, and delayed responses in the hoisting of ultra-large components in UHV substations have been solved. This has enabled high-precision, multi-dimensional safety assessments and interventions throughout the entire process, improving construction safety and efficiency.
Patent Information
- Application Number
- CN202611523797.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-29
- Publication Date
- 2026-11-17
AI Technical Summary
Existing hoisting monitoring technologies cannot adapt to the strong electromagnetic environment and wide temperature conditions in ultra-high voltage substations. They suffer from problems such as monitoring blind spots, inaccurate assessments, delayed responses, and insufficient intervention, making it difficult to ensure the safety and quality of hoisting ultra-large components.
It employs distributed sensing node units, data acquisition and transmission units, intelligent data processing units, and hierarchical early warning and linkage control units, combined with fiber optic stress sensing modules, MEMS inertial measurement modules, temperature-sensitive adaptive magnetic-photocuring composite fixing structures, movable binocular scanning units, and machine-sensor collaborative prediction sampling and scheduling modules, to achieve real-time monitoring and multi-dimensional intervention with no blind spots and high precision.
It achieves full-process, all-area coverage monitoring, accurate and reliable assessment, adaptability to wide temperature environments, high efficiency in response, and proactive prevention and control, thereby improving hoisting safety and construction efficiency.
Smart Images

Figure CN123105472A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power equipment installation and construction monitoring technology in ultra-high voltage substations, specifically to a real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in ultra-high voltage substations. Background Technology
[0002] Ultra-high voltage (UHV) power transmission is the core carrier for the cross-regional optimization of energy allocation in my country. As a key node in the power grid, the construction quality of UHV substations directly determines the safety and stability of the power grid operation. With the continuous improvement of construction standards for UHV projects, core components within substations, such as steel pipe A-type columns, GIS roof structures, and main transformer radiators, are generally becoming larger, heavier, and more complex. A single component can weigh tens of tons, and its span or height can reach tens of meters. The hoisting of these ultra-large components is a core and critical process in substation construction. The entire process encompasses multiple conditions, including lifting off the ground, adjusting the aerial posture, long-distance relocation, and precise docking. The stress state of the components is dynamically affected by multiple factors, such as the arrangement of lifting points, hoisting posture, on-site wind load, lifting acceleration, and cable force deviation. This makes them highly susceptible to localized stress concentration, excessive structural deformation, and even safety accidents such as component buckling, weld cracking, and lifting point failure. These accidents not only cause huge economic losses but also severely delay the project construction schedule.
[0003] Currently, monitoring technologies for the hoisting of large components in the engineering field are mainly divided into three categories: manual inspection, static sampling inspection, and simple online monitoring. Manual inspection relies on the experience and judgment of construction personnel, and verifies the appearance deformation through visual observation and simple tools. It cannot obtain internal stress data of the components, resulting in a large number of monitoring blind spots and a serious lag in risk response. Traditional static testing mostly uses resistance strain gauges, which are welded or glued to the surface of the components. Data can only be read during the downtime, and it cannot achieve continuous dynamic monitoring throughout the entire process. At the same time, resistance strain gauges are electrical sensors, and data drift is serious in the environment of ultra-high voltage and strong electromagnetic fields, making it difficult to guarantee measurement accuracy. Welding fixation will damage the base material of the component, which does not meet the requirements of power equipment installation process. Ordinary adhesive methods are prone to falling off under hoisting vibration and high-altitude wind loads, resulting in insufficient reliability.
[0004] In recent years, some projects have begun to adopt fiber optic sensing online monitoring systems, which have solved the problem of electromagnetic interference to a certain extent, but there are still many significant defects in practical applications.
[0005] Firstly, the layout of measuring points relies on experience and can only cover known key sections. For irregularly shaped components or components with assembly deviations, it is easy to miss actual stress concentration areas, forming monitoring blind spots and making it impossible to fully grasp the stress state of the components.
[0006] Secondly, the accuracy of safety assessment is insufficient. The existing system directly uses the residual stress data from the factory and the theoretical finite element model as the assessment benchmark. However, after the components are transported over long distances and assembled on site, the residual stress will be redistributed, and there will inevitably be manufacturing geometric deviations, which will lead to the distortion of the additional stress calculation and the full-field inversion results, and the safety threshold setting will be inaccurate, which will easily result in false alarms or missed alarms.
[0007] Third, the sensor fixing structure has poor environmental adaptability. The commonly used anaerobic adhesive adhesion method has a significantly longer curing time in low temperature environment, which seriously affects the construction efficiency. In high temperature environment, the bonding strength decreases and there is a risk of falling off. Moreover, the adhesive residue after removal is difficult to clean, which will affect the subsequent anti-corrosion coating process of the component.
[0008] Fourth, the sampling strategy has a lag. The existing adaptive sampling passively identifies the change of working conditions based only on the rate of change of sensor data, which has a lag of hundreds of milliseconds to several seconds. It cannot accurately capture the peak impact load at the moment of working condition change, and key data is easily lost.
[0009] Fifth, the safety intervention methods are limited. Most systems can only send alarm information to on-site personnel, rely on manual operation of cranes to adjust working conditions, have long emergency response times, and cannot actively eliminate stress concentration causes from the source of load. Frequent shutdowns will also significantly reduce construction efficiency.
[0010] In summary, existing hoisting monitoring technologies are ill-suited to the strong electromagnetic environment and wide temperature range of ultra-high voltage (UHV) construction conditions. They also fail to meet the comprehensive needs of full monitoring coverage, high-precision assessment, timely response, and multi-dimensional intervention. There is an urgent need to develop a targeted real-time monitoring system to systematically address these technical deficiencies and comprehensively ensure the safety and quality of UHV ultra-large component hoisting operations. Summary of the Invention
[0011] To address the problems in existing technologies, this invention provides a real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations. Through the coordinated use of multiple technical means, it achieves blind-spot-free monitoring, high-precision safety assessment, and proactive multi-dimensional intervention for hoisting ultra-large components in strong electromagnetic and wide temperature environments, thereby comprehensively improving the safety management level and construction efficiency of hoisting operations.
[0012] The technical solution adopted by this invention to solve its technical problem is: a real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in ultra-high voltage substations, including distributed sensing node units, data acquisition and transmission units, intelligent data processing units, and hierarchical early warning and linkage control units; The distributed sensing node unit includes several fixed composite sensing nodes and portable supplementary sensing nodes. Both types of nodes integrate fiber optic stress sensing modules and MEMS inertial measurement modules. The bottom of each node is equipped with a temperature-sensitive adaptive magnetic-photocuring composite fixing structure. The system is equipped with a movable binocular scanning unit to acquire the full-field geometry and deformation data of the component, identify the deformation concentration areas not covered by the fixed measuring points, and guide the portable supplementary sensing nodes to dynamically supplement the points and connect to the monitoring network. The data acquisition and transmission unit includes a ground acquisition host, an air-to-ground dual-path redundant communication module, and a machine-sensor collaborative prediction sampling and scheduling module. The machine-sensor collaborative prediction sampling and scheduling module accesses the crane's operating parameters, predicts the switching nodes of the hoisting conditions based on the working condition state machine, and adjusts the sampling frequency of the sensor nodes in advance. The intelligent data processing unit includes a residual stress in-situ recalibration module, a finite element model in-situ self-calibration module, and a stress-deformation coupling inversion module, which are used to correct the actual stress baseline of the component and the calculation model, remove the additional stress from the hoisting and invert the stress and deformation distribution of the component in the whole field. The graded early warning and linkage control unit is connected to both the crane control system and the multi-point sling force adjustment mechanism, and is used to trigger graded early warning and multi-dimensional operation intervention based on the safety assessment results.
[0013] Specifically, the temperature-sensitive adaptive magnetic-photocurable composite fixing structure includes a neodymium iron boron strong magnetic adsorption seat at the bottom of the node housing, a butyl rubber modified pressure-sensitive adhesive layer at the center of the adsorption seat, an acrylic ester ultraviolet light curable adhesive layer in the outer annular closed adhesive groove, and an annular micro ultraviolet LED light source with a wavelength of 365nm embedded in the inner side of the annular adhesive groove; the node has a built-in ceramic micro heating element, which triggers the ultraviolet light source to accelerate the curing of the photocurable adhesive in low-temperature environments, compensates for the bonding strength through the pressure-sensitive adhesive layer in high-temperature environments, and activates the heating element to heat and de-adhere the adhesive to achieve non-destructive peeling when removing it.
[0014] Specifically, the movable binocular scanning unit consists of two structured light binocular cameras. Before hoisting, the camera scans to obtain the initial three-dimensional shape of the component. During hoisting, the camera scans at preset intervals to obtain displacement and deformation data of the entire field. Areas with deformation gradients higher than preset thresholds are marked as deformation concentration areas, and the location coordinates are pushed to the on-site terminal to guide the point supplementation. After the portable supplementation sensor node is deployed, it automatically accesses the acquisition network via wireless protocol.
[0015] Specifically, the machine-sensor collaborative prediction sampling and scheduling module synchronously collects parameters such as crane hoist speed, lifting height, hook load, and trolley travel speed to construct a state machine that includes four working conditions: lifting off the ground, smooth aerial transfer, attitude adjustment, and positioning and docking. When the operating parameters reach the switching threshold of the corresponding working condition, the sampling frequency of the sensor node is adjusted 200ms in advance. The sampling frequency is 500~1000Hz during the lifting off the ground and positioning and docking stages, 100~300Hz during the attitude adjustment stage, and 10~30Hz during the smooth aerial transfer stage.
[0016] Specifically, the execution steps of the residual stress in-situ recalibration module are as follows: S1: Collect initial three-dimensional topographic data of the component and initial strain data of each measuring point before hoisting; S2: Based on the theory of elastic-plastic residual stress release, establish the mapping relationship between component morphology deviation and residual stress redistribution. S3: Using the factory residual stress field as the initial iteration field, combined with the measured initial strain and morphological deviation, the in-situ residual stress baseline of the component is obtained through iterative correction. S4: In real-time monitoring, the total strain data is subtracted from the in-situ residual strain baseline of the corresponding measuring point to obtain the additional strain of hoisting and convert it into additional stress of hoisting.
[0017] Specifically, the execution steps of the finite element model in-situ self-calibration module are as follows: S1: Import component theory design finite element initial model; S2: Map the actual geometric dimensions, wall thickness deviation, and assembly form and position deviation obtained from binocular scanning to the initial model mesh, and use mesh deformation technology to correct the geometric model; S3: Correct the model boundary constraints based on the actual connection form of the assembled nodes on site; S4: Verify the accuracy of the model by measuring the first-order natural frequency. If the error is less than 5%, generate a calculation benchmark model.
[0018] Specifically, the stress-deformation coupled inversion module is based on the in-situ calibrated finite element model, and constructs a weighted objective function containing four types of constraints: fixed measuring point stress, supplementary measuring point stress, nodal attitude deformation, and binocular full-field deformation. It is solved using an improved Gauss-Newton iteration method with an adaptive step size, and the iteration convergence threshold is 10. -3 The final output is a cloud map of the full-field stress field and deformation field of the component.
[0019] Specifically, the multi-point cable force adjustment mechanism includes cable force sensors and hydraulic adjustment cylinders installed on each cable; the system pre-stores a cable force-stress sensitivity matrix calculated based on a benchmark model. When local stress concentration caused by uneven cable force is detected, the sensitivity matrix is called to calculate the incremental adjustment amount of each cable and control the hydraulic cylinder to extend and retract to balance the load.
[0020] Specifically, the graded early warning and linkage control unit is set with four safety levels: normal, attention, early warning, and danger. The attention level pushes observation prompts to on-site technicians, the early warning level triggers on-site audible and visual alarms and initiates cable tension fine-tuning, and the danger level simultaneously triggers crane deceleration / pause commands and large-scale cable tension adjustment.
[0021] The beneficial effects of this invention are: Combining point and surface monitoring for full coverage: By complementing high-precision monitoring of fixed measuring points with binocular vision full-field scanning, and combining with a dynamic compensation mechanism, monitoring blind spots can be automatically identified and filled, completely solving the problem of incomplete coverage of traditional preset measuring points, realizing full-area force perception throughout the entire process of component hoisting, and avoiding the risk of missed detection.
[0022] In-situ calibration ensures accurate and reliable assessment: By recalibrating residual stress in situ and self-calibrating the finite element model in situ, the influence of residual stress redistribution during transportation and assembly and manufacturing geometric deviations on the assessment results is eliminated. The calculation error of additional stress can be controlled within 2%, which greatly improves the accuracy of safety assessment and effectively reduces the probability of false alarms and missed alarms.
[0023] Wide temperature adaptability and efficient and non-destructive installation: The temperature-sensitive magnetic-photocurable composite fixing structure requires no welding and does not damage the base material of the component. It can guarantee bonding strength and curing efficiency in environments ranging from -10℃ to 50℃. The installation efficiency is more than twice that of traditional methods, and it can be removed without damage or residue. It is fully adaptable to the construction needs of different seasons and regions.
[0024] Predictive scheduling ensures stable and efficient transmission: Machine-sensor collaborative predictive sampling scheduling eliminates the lag of traditional sampling, fully captures transient loads during operating condition switching, and significantly reduces data redundancy during stable phases; coupled with dual-path redundant communication between air and ground, it can eliminate signal blind spots in complex sites, achieving 100% data transmission continuity.
[0025] Multi-dimensional intervention and proactive prevention improve efficiency: Constructing a multi-dimensional intervention system of active cable force adjustment and crane control eliminates stress concentration causes from the source of load. Compared with single shutdown intervention, it can quickly control risks and reduce unnecessary downtime, balancing safety and construction efficiency; the graded early warning mechanism realizes gradient management of risks, and improves emergency response speed by an order of magnitude. Attached Figure Description
[0026] The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0027] Figure 1 This is a schematic diagram of the overall architecture of the monitoring system described in this invention; Figure 2 This is a cross-sectional schematic diagram of the temperature-sensitive adaptive composite fixing structure described in this invention; Figure 3 This is a schematic diagram of the machine-sensor collaborative prediction sampling scheduling logic described in this invention; Figure 4 This is a block diagram of the hierarchical early warning and multi-dimensional linkage control logic described in this invention. Detailed Implementation
[0028] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.
[0029] like Figures 1-4 As shown, the real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in ultra-high voltage substations according to the present invention includes distributed sensing node units, data acquisition and transmission units, intelligent data processing units, and hierarchical early warning and linkage control units. Each unit is connected in sequence and works in coordination to form a complete monitoring and control closed loop.
[0030] Distributed sensing node unit: This unit consists of a fixed composite sensing node, a portable supplementary sensing node, and a movable binocular scanning unit, enabling full-coverage perception that combines point and surface sensing.
[0031] Both the fixed composite sensing node and the portable replacement sensing node integrate a fiber Bragg grating stress sensing module, a MEMS inertial measurement module, a node control module, and an energy storage module. The fiber Bragg grating stress sensing module uses a fiber Bragg grating as the sensing element; its all-optical measurement principle naturally eliminates electromagnetic interference. It incorporates an independent temperature-compensated grating to eliminate the interference of temperature and strain on stress measurement in real time, achieving a strain measurement resolution of up to 1 με. The MEMS inertial measurement module integrates a three-axis accelerometer and a three-axis gyroscope to acquire the three-dimensional attitude angles and linear accelerations of the measurement points. The attitude angle measurement accuracy is better than 0.1°, used for overall component deformation and attitude analysis. The core sensing performance of both types of nodes is consistent. The portable replacement node is 40% smaller and uses integrated energy storage, making it suitable for temporary, rapid replacement scenarios.
[0032] Temperature-sensitive adaptive magnetic-photocurable composite fixing structure: This fixing structure is installed at the bottom of all sensing nodes, including a strong magnetic adsorption base, a composite adhesive layer, and auxiliary functional components. The strong magnetic adsorption base is made of N52 grade neodymium iron boron permanent magnets and can be adapted to the surface shape of the component as a flat or curved contact surface. The adsorption force of a single node is not less than 200N, enabling rapid temporary positioning. The central area of the adsorption base is covered with a 0.3mm thick butyl rubber modified pressure-sensitive adhesive layer, and the outer periphery is set with a 0.5mm deep and 3mm wide annular closed adhesive groove, which is filled with an acrylic UV-curable adhesive layer. Eight micro UV LEDs with a wavelength of 365nm and a power of 5mW are uniformly embedded inside the adhesive groove. A ceramic micro heating element is installed inside the node housing at the position of the adhesive groove, with a heating temperature range of 60~80℃, which is uniformly driven by the node control module.
[0033] During installation, remove the release film from the adhesive layer, align the magnetic base with the measuring point, and apply it to the component surface. Lightly press the joint; the pressure-sensitive adhesive layer will immediately form a preliminary bond, resisting minor impacts during construction. At room temperature, the UV-cured adhesive naturally cures for 2 hours to reach its rated bonding strength. When the built-in temperature sensor detects an ambient temperature below 5℃, it automatically activates the UV LED light source to irradiate the adhesive layer, achieving over 90% bonding strength within 8 minutes, solving the problem of slow low-temperature curing. When the ambient temperature is above 40℃, the high elasticity and viscosity of the pressure-sensitive adhesive layer compensates for the high-temperature creep of the UV-cured adhesive, maintaining an overall bonding strength of no less than 85% of the rated value. When removing the joint, the control module activates a micro-heating plate for 3 minutes to reduce the cohesion of the adhesive layer, allowing for easy peeling. Only a very thin adhesive residue remains on the component surface, which can be completely removed with ordinary solvents without affecting subsequent anti-corrosion coating processes.
[0034] Portable binocular scanning unit: Employing two industrial-grade structured light binocular cameras with a measurement accuracy of 0.1mm / m, it can be flexibly deployed on supports around the hoisting operation area. Before hoisting, the binocular scanning unit performs a full-field 3D scan of the component, acquiring its initial geometric shape and dimensional parameters, and outputting a point cloud model. During hoisting, the component is rapidly scanned at preset time intervals of 5 minutes to acquire full-field 3D displacement and deformation data, and to calculate the deformation gradient of each region. The system marks areas with deformation gradients higher than a preset threshold as deformation concentration areas, determines whether fixed sensor nodes are deployed in these areas, and if they are uncovered blind spots, pushes the area coordinates and deformation magnitude to the on-site construction terminal, guiding construction personnel to deploy portable supplementary sensor nodes to the corresponding positions. After the supplementary node is powered on, it automatically connects to the acquisition network via an industrial wireless LAN, synchronizes the timestamp, and begins data acquisition to supplement the high-precision stress monitoring data of the area, eliminating monitoring blind spots.
[0035] Data acquisition and transmission unit: This unit includes a ground acquisition host, an air-to-ground dual-path redundant communication module, and a machine-sensor collaborative prediction sampling scheduling module, ensuring the synchronous acquisition and stable and efficient transmission of multi-source data.
[0036] The ground acquisition host connects to the fiber Bragg grating modules of all sensing nodes via an optical fiber bus. It uses wavelength demodulation technology to achieve high-speed acquisition of grating signals, with a demodulation frequency of up to 2000Hz. At the same time, it receives attitude data from the MEMS modules of each node and full-field deformation data from the binocular scanning unit through an industrial wireless local area network. It uses GPS time synchronization to unify high-precision timestamps, with a time synchronization error of less than 1ms, to achieve precise time alignment of stress data, attitude data and visual deformation data, providing a data foundation for subsequent coupled analysis.
[0037] The air-to-ground dual-redundant communication module includes two independent transmission channels: a terrestrial 5G private network communication link and a UAV relay communication link. Under normal operating conditions, the terrestrial 5G private network is used preferentially for data transmission, with bandwidth sufficient for data transmission at all sampling frequencies. When the module detects that the terrestrial link signal strength is below -90dBm, it automatically and seamlessly switches to the UAV relay link with a switching latency of less than 50ms. The relay UAV hovers 30m directly above the hoisting operation area, acting as a mobile communication relay node. This effectively bypasses the obstruction of metal structures and mountain terrain, eliminating signal blind spots and ensuring the continuity of data transmission throughout the hoisting process.
[0038] The machine-sensor collaborative prediction sampling and scheduling module: This module simultaneously connects to the CAN bus of the crane control system to acquire real-time operating parameters such as hoist speed, lifting height, hook load, and trolley travel speed. It constructs a finite state machine encompassing four working conditions: lifting off the ground, smooth aerial transfer, attitude adjustment, and positioning and docking. The module sets corresponding switching thresholds for each working condition. For example, when the lifting height reaches 90% of the component's height off the ground and the trolley travels 5 meters from the positioning area, it is determined to be a critical point for working condition switching. When the operating parameters reach the switching threshold, the module sends a sampling frequency adjustment command to all sensor nodes 200ms in advance, achieving synchronous matching between sampling frequency switching and working condition switching. This eliminates the lag of traditional passive identification and fully captures the transient impact load at the moment of working condition switching.
[0039] The sampling frequency corresponds to the operating conditions as follows: During the lifting and docking phases, when the load impact is large and the stress changes rapidly, the highest sampling frequency of 500~1000Hz is used to accurately capture transient stress characteristics; during the attitude adjustment phase, when the stress and attitude continuously change, a medium sampling frequency of 100~300Hz is used to track the changing trend; during the stable aerial transfer phase, when the stress state is stable, the lowest sampling frequency of 10~30Hz is used to significantly reduce data redundancy and system power consumption. Through predictive dynamic scheduling, while ensuring the monitoring accuracy of key stages, the amount of redundant data can be reduced by more than 70%, significantly reducing transmission and storage pressure and extending node endurance.
[0040] Intelligent data processing unit: This unit is deployed on the field edge gateway and cloud server, and includes a residual stress in-situ recalibration module, a finite element model in-situ self-calibration module, and a stress-deformation coupling inversion module to achieve in-depth processing and high-precision evaluation of monitoring data.
[0041] The residual stress in-situ recalibration module is executed as follows: S1: Before hoisting operations, collect the initial three-dimensional topography point cloud data of the component obtained by binocular scanning, as well as the initial strain data of all fixed sensor nodes, and record the ambient temperature. S2: Based on the residual stress release theory of elastoplastic mechanics, establish the mapping relationship between component morphology deviation (relative to the ideal design morphology) and residual stress redistribution, and distinguish between stress release and redistribution components generated during transportation and assembly. S3: The residual stress detection field when the component leaves the factory is used as the initial iteration field, the measured initial strain of each measuring point is used as the constraint condition, and the stress correction amount corresponding to the morphology deviation is combined for iterative calculation. The iteration convergence threshold is that the strain error is less than 2με. S4: After iterative convergence, the in-situ residual stress baseline of the component is obtained, corresponding to the actual residual strain value of each measuring point; during real-time monitoring, the total strain data collected by the fiber optic grating is subtracted from the in-situ residual strain baseline of the corresponding measuring point to obtain the additional strain generated purely by the hoisting load, which is then converted into the additional hoisting stress through the material elastic modulus.
[0042] This module eliminates the interference of residual stress redistribution caused by transportation and assembly on safety assessment, making the calculation of additional stress more consistent with the actual stress state of the component.
[0043] The in-situ self-calibration module for finite element models executes the following steps: S1: Pre-import the theoretical design finite element initial model of the component, including element properties, material parameters, design geometry and theoretical boundary conditions; S2: The actual geometric parameters of the components obtained by binocular scanning, including actual wall thickness, cross-sectional dimensions, member length deviation, assembly form and position deviation, etc., are mapped to the mesh nodes of the initial model. The geometric model is corrected by mesh deformation technology based on radial basis function to ensure that the mesh quality does not degrade. S3: Based on the actual connection form of the assembled nodes on site (such as the actual preload of bolted ball joints and the actual constraint stiffness of welded joints), modify the boundary constraint conditions and node stiffness parameters of the model. S4: The first natural frequency of the component is measured by on-site hammering method and compared with the calculated frequency of the corrected model. If the error is less than 5%, the model verification is passed and the final calculation benchmark model is generated; if the error exceeds the limit, the boundary parameters are repeatedly corrected until the accuracy requirements are met.
[0044] This module eliminates the impact of manufacturing and assembly errors on the inversion calculation, providing a reliable model benchmark for subsequent full-field analysis.
[0045] The stress-deformation coupled inversion module is based on the in-situ calibrated finite element baseline model and integrates multi-source monitoring data for coupled inversion. The module constructs a weighted least squares objective function, with constraints including four types of data: measured additional stress at fixed measuring points, measured additional stress at supplementary measuring points, measured attitude deformation at nodes, and full-field displacement deformation from binocular scanning. Corresponding weight coefficients are set according to the measurement accuracy of each type of data, with the fixed measuring point stress having the highest weight, followed by the full-field deformation. The solution algorithm employs an improved Gauss-Newton iteration method with an adaptive step size. During iteration, the step size is dynamically adjusted based on the descent of the objective function to avoid iteration oscillations. The iteration convergence threshold is when the change in the objective function is less than 10. -3 After iterative convergence, the module outputs stress and deformation distribution cloud maps of the entire component, which can intuitively display the location of maximum stress, maximum deformation, and overall stress state. This module integrates point-based high-precision data with surface-based full-field data, overcoming the limitations of discrete measurement points and comprehensively and accurately reflecting the overall stress state of the component.
[0046] Tiered early warning and linkage control unit: This unit includes an on-site audible and visual early warning module, a remote command terminal, a crane linkage interface, and a multi-point sling force adjustment mechanism, constructing a proactive prevention and control system with graded response and multi-dimensional intervention.
[0047] Multi-point sling force adjustment mechanism: For multi-point lifting conditions, tension-type sling force sensors and hydraulic adjustment cylinders are deployed on each sling. The sling force measurement accuracy is better than 1%FS, and the hydraulic cylinder stroke adjustment accuracy is better than 1mm, both connected to the system control terminal. The system pre-calculates the sling force-stress sensitivity matrix based on a finite element reference model, i.e., the stress change in each region of the component corresponding to a unit change in sling force. During real-time monitoring, when local stress concentration caused by uneven sling force is identified, the sensitivity matrix is called to calculate the incremental adjustment amount for each sling. A step-by-step adjustment method is used to control the extension and retraction of the corresponding hydraulic cylinders to balance the load on each sling, eliminate the cause of stress concentration at the source of the load, and prevent further stress deterioration.
[0048] A tiered early warning mechanism is established based on the yield strength of the component material, the design allowable stress, and the deformation limits. Four safety levels are set: Normal, Attention, Warning, and Danger, corresponding to different over-limit ratios and response actions. Normal level: Stress and deformation are both within 70% of the allowable value; data is displayed in real time, and no warning action is taken. Attention level: Stress reaches 70%~85% of the allowable value; a warning message is sent to on-site technicians, suggesting enhanced monitoring. Warning level: Stress reaches 85%~95% of the allowable value; an on-site audible and visual alarm is triggered, and a warning message and adjustment suggestions are sent to the hoisting commander, simultaneously initiating automatic cable tension fine-tuning. Danger level: Stress exceeds 95% of the allowable value; two interventions are triggered simultaneously: first, a deceleration or suspension command is sent to the crane control system via a hard-wired interface, with a response delay of less than 200ms; second, a significant cable tension adjustment is initiated to quickly reduce the component stress level.
[0049] Digital twin visualization: The remote command terminal constructs a lightweight digital twin model of the component, and maps stress cloud map, deformation cloud map and hoisting posture in real time to realize the three-dimensional visualization of monitoring data, which helps the command personnel to intuitively grasp the stress state of the component and improve decision-making efficiency.
[0050] Example 1: Application of monitoring for hoisting A-type steel pipe columns in low-temperature winter environments This embodiment was applied to the hoisting operation of the steel pipe A-type column in the main transformer area of a 1000kV UHV substation in northern China. The construction environment temperature was about -3℃. The A-type column was made of Q355B seamless steel pipe, with a total height of 42 meters, a maximum diameter of 1.2 meters, and a total weight of 28 tons. The hoisting process adopted was a single-point lifting by the main crane plus bottom guide rope assistance. The entire hoisting process consisted of four stages: lifting off the ground, vertical adjustment in the air, horizontal transfer, and cup-shaped positioning.
[0051] This monitoring involved installing eight fixed composite sensor nodes at the welded joints of the lifting lugs on the A-type columns, at the upper 1 / 3 of the column height section, at the middle 2 / 3 of the column height section, and at the column base bevel. During sensor installation, workers removed the release film from the bottom of each node, aligned the curved neodymium iron boron magnetic base with the pre-set measuring point, and applied it to the steel pipe surface. After lightly pressing the node, the pressure-sensitive adhesive layer immediately formed a preliminary fixation, ensuring stable adhesion without detachment. When the built-in temperature sensor detected an ambient temperature below 5℃, it automatically activated a 365nm ring-shaped ultraviolet LED light source to cure the adhesive layer. After 8 minutes, the adhesive layer reached 90% of its rated bonding strength, meeting the requirements for lifting stress. The entire installation process required no welding equipment or additional heating tools. Installing all sensors on a single A-type column took approximately 35 minutes and caused no damage to the base material.
[0052] Before the hoisting operation began, two mobile binocular scanning units were activated to perform a full-field 3D scan of the A-type column, acquiring the actual geometric dimensions and initial topographic point cloud data of the component. Simultaneously, the intelligent data processing unit completed two core calibration tasks: first, in-situ recalibration of residual stress. Using the component's factory residual stress data as the initial field, combined with the topographic deviation obtained from the binocular scan and the initial strain data from eight measuring points, the baseline of in-situ residual stress was obtained after five iterations. Comparison showed that the residual stress at some measuring points had changed by 8% to 12% compared to the factory value, verifying the necessity of in-situ calibration. Second, in-situ self-calibration of the finite element model. Parameters such as actual wall thickness and straightness deviation were substituted into the theoretical model, and the geometric model was corrected using mesh deformation technology. The model was then checked against the measured first-order natural frequency, with a final frequency error of 3.2%, meeting the accuracy requirements, and generating a calculation benchmark model. Simultaneously, the machine-sensor collaborative prediction sampling and scheduling module was connected to the crane control system to acquire real-time operating parameters such as hoist speed and lifting height. A preset lifting height of 38 meters was designated as the critical ground-lift prediction point.
[0053] During the lifting phase, the crane slowly raises the hook. When the lifting height reaches the predicted threshold of 38 meters, the sampling and scheduling module increases the sampling frequency of the sensor nodes to 800Hz 200ms in advance, ensuring complete and accurate acquisition of the transient impact stress of the component at the moment of liftoff. The data is transmitted in real time to the on-site edge gateway via the terrestrial 5G private network. The intelligent processing unit simultaneously completes residual stress stripping and inverts the overall stress distribution of the column based on the corrected finite element model. Once the component is completely off the ground and the crane begins its trolley movement, the module determines that it has entered the stable aerial transfer phase based on the travel speed and stress change data, automatically reducing the sampling frequency to 20Hz, significantly reducing data transmission volume and node power consumption.
[0054] During the relocation, the binocular scanning unit scanned the column every 5 minutes. One scan detected a deformation gradient exceeding the threshold in a non-measuring area in the middle of the column, marking it as a deformation concentration zone. The system immediately pushed the coordinates of this area to the on-site terminal. Construction personnel then deployed a portable supplementary sensor node, which automatically connected to the acquisition network, transmitting stress data from this area in real time. It was confirmed that the localized stress fluctuation was caused by lateral gusts, with the maximum additional stress at 72% of the allowable value, within the concern level and not exceeding the safety threshold. When the trolley traveled 5 meters from the cup-shaped rim, the sampling and scheduling module anticipated entering the positioning and docking phase, increasing the sampling frequency to 800Hz in advance to track stress changes throughout the positioning process until the component was precisely seated and temporarily fixed.
[0055] After hoisting, the miniature heating elements of all nodes were activated via the control terminal. After heating for 3 minutes, the nodes were peeled off one by one. Only a very thin layer of adhesive remained on the column surface, which could be completely removed by wiping with alcohol, without affecting the subsequent anti-corrosion coating process. Throughout the hoisting process, monitoring data was continuous and without any data loss. The error in the measurement of additional stress was less than 1.5%. One unexpected deformation area was successfully filled. The sensors remained reliably fixed in the low-temperature environment, ensuring the safety of the hoisting operation.
[0056] Comparative Example 1: A fiber optic monitoring system using traditional anaerobic adhesive for fixing was used to monitor the simultaneous hoisting of Type A columns of the same specifications, with the same number of measuring points deployed. Due to the slow curing of anaerobic adhesive in low-temperature environments, the deployment of a single column took approximately 2 hours, significantly reducing construction efficiency. The system directly used the factory residual stress baseline and theoretical finite element model without in-situ calibration. A comparison showed that the calculated additional stress deviated by 12% from the actual value, indicating a significant assessment error. A fixed sampling frequency of 500Hz was used throughout, resulting in 25 times the data volume during the stable phase compared to this example. Furthermore, due to passive identification of the working condition switching, approximately 300ms of impact peak data was lost at the moment of liftoff. The lack of a full-field scanning and dynamic compensation mechanism meant that local deformation areas of the column were not monitored, creating blind spots. After node removal, a large amount of adhesive residue remained on the column, requiring additional grinding and cleaning, which took approximately 1 hour. In contrast, the system in this example improved deployment efficiency by 243% in low-temperature environments, improved stress assessment accuracy by an order of magnitude, achieved full-area monitoring without blind spots, and removed the column without damage or residue, demonstrating significant overall performance advantages.
[0057] Example 2: Application of monitoring for the four-point lifting and hoisting of a space frame roof This embodiment is applied to the overall lifting and hoisting project of the GIS room roof of a certain ultra-high voltage substation. The roof is a square pyramidal bolted ball roof structure with a span of 48 meters, a width of 24 meters, a lifting height of 18 meters, and a total weight of 65 tons. The four-point hoisting synchronous lifting process is adopted. The hoisting process is divided into three stages: synchronous lifting, aerial attitude correction, and roof placement.
[0058] This monitoring involved deploying 24 fixed composite sensing nodes at the four lifting point ball joints of the space frame, the mid-span of the lower chord, and key load-bearing members of the upper chord, covering the main load-bearing members and critical nodes. Considering the curved surface characteristics of the circular tube members in the space frame, the magnetic mounting bases of the sensing nodes were fitted with arc-shaped contact surfaces to ensure a tight fit against the outer wall of the tubes. The installation of all 24 nodes was completed by three construction workers in approximately 1.8 hours, requiring no hot work and meeting on-site safety management requirements. Each sling of the four-point lifting device was equipped with a tension-type cable force sensor and a hydraulic adjustment cylinder, connected to the system's linkage control unit. The system pre-calculated a 4D cable force-stress sensitivity matrix based on the space frame's baseline model, clearly defining the influence coefficient of each cable force adjustment on the stress of key members.
[0059] Before the hoisting operation, the binocular scanning unit performs a full-field scan of the assembled space frame to acquire actual geometric parameters and assembly deviation data; the intelligent data processing unit completes in-situ recalibration of residual stress and in-situ self-correction of the finite element model. Due to deviations in member length and misalignment of nodes during on-site assembly, the maximum geometric deviation of the corrected finite element model compared to the theoretical model reaches 12mm, and the first-order frequency deviation is 7.8% before correction and 4.1% after correction, meeting the accuracy requirements and generating a calculation benchmark model that conforms to the actual state. In the initial stage of lifting, the four lifting points are lifted synchronously. The machine-sensor collaborative prediction sampling and scheduling module predicts the entry into the synchronous lifting stage based on the hoisting speed and load data, and sets the sampling frequency to 500Hz. The system calculates the overall deflection and torsion angle of the space frame in real time by fusing multi-node MEMS attitude data and binocular scanning data. At the same time, it obtains the stress values of each member through fiber optic grating data. After residual stress stripping, the stress values are input into the coupled inversion module and iteratively calculated using the improved Gauss-Newton method. After six iterations, the system converges to obtain the stress cloud map and deformation cloud map of the entire space frame, which are updated and displayed in real time on the digital twin model of the command terminal.
[0060] When the space frame was lifted to a height of 10 meters, a slight synchronization error occurred in one of the lifting cylinders, resulting in excessive tension in the corresponding slings. This caused a slight twisting of the space frame, and the local stress in a certain member on the side continuously increased to 88% of the allowable value, reaching the warning threshold. The system recognized that the stress concentration was caused by uneven cable tension and immediately called upon the pre-stored sensitivity matrix to calculate the incremental adjustment amount for each sling. It controlled the hydraulic cylinder of the sling on the higher side to retract by 2mm, while simultaneously extending the cylinder on the lower side by 1.5mm. After step-by-step adjustment, the cable tension of each sling returned to equilibrium in about 2 seconds, and the member stress dropped to 76% of the allowable value. No shutdown was triggered throughout the process, and the lifting progress was not affected. Once the space frame entered the stable lifting phase, the system automatically reduced the sampling frequency to 30Hz, significantly reducing the amount of data processing. During this process, binocular scanning detected concentrated deformation in a certain secondary web member. No measuring point was preset at this location. The system immediately guided the construction personnel to set up portable supplementary nodes to track the stress changes of this member in real time, confirming that it was within the safe range and eliminating monitoring blind spots.
[0061] When the space frame was raised to 1 meter below the design elevation, the sampling and scheduling module anticipated the entry into the roof positioning and docking stage, and increased the sampling frequency to 500Hz in advance. Throughout the docking process, it monitored the stress changes of the members, and coordinated with cable tension fine-tuning to ensure the stability of the space frame until the space frame was fixed and the hoisting operation was completed. No stress exceeding limits occurred during the entire hoisting process. The system successfully eliminated one synchronization deviation condition through active cable tension adjustment and compensated for one blind spot in a secondary member, effectively avoiding the risk of space frame deformation and damage.
[0062] Comparison Example 2: A conventional fixed-point fiber optic stress monitoring system was used for monitoring the hoisting of similar space frames, lacking active cable force adjustment and full-field scanning capabilities. This system could only display stress values at preset measuring points and could not reflect the overall torsional state of the space frame. In previous similar projects, uneven cable force caused plastic deformation of members, requiring work stoppages for manual cable force adjustment, resulting in delays of approximately 30 minutes per instance. The system used theoretical models for evaluation, leading to significant deviations in stress calculations, resulting in three false alarms and unnecessary work stoppages, accumulating a one-hour delay. Without a dynamic compensation mechanism, stress concentration in secondary members could not be detected in a timely manner, leading to instances where member buckling risks were not identified early. In contrast, the system in this embodiment achieves full-field perception of space frame stress, and active cable force adjustment can eliminate stress concentration at its source, reducing the number of work stoppages. The time for handling single stress anomalies is shortened from 30 minutes to 2 seconds, while significantly improving evaluation accuracy, combining safety and economy.
[0063] Example 3: Application of Monitoring and Multi-dimensional Linkage Control for Lifting Components in Complex Mountainous Terrain This embodiment is applied to the hoisting operation of the main transformer radiator of an ultra-high voltage substation in a mountainous area. The site has a large terrain undulation, with dense surrounding mountains and power distribution equipment structures, severe ground wireless signal obstruction, and a complex communication environment. The total weight of the radiator component to be hoisted is 32 tons and the height is 12 meters. It is hoisted using a four-point hoisting method. The hoisting path needs to pass through two sets of structures, and there are many areas with obvious signal obstruction and lateral wind vents.
[0064] This monitoring system deployed 12 fixed composite sensor nodes at the four lifting points, side wall reinforcing ribs, and bottom support locations of the component. The slings were equipped with tension sensors and hydraulic adjustment mechanisms. Data transmission employed a dual-path redundant configuration (air-ground). In addition to a dedicated 5G ground network link, an industrial-grade relay drone was deployed, pre-positioned to hover 30 meters directly above the lifting path, ready for deployment. The tiered early warning and linkage control unit simultaneously connected to the crane control system and tension adjustment mechanism via hard-wired interfaces, with a pre-set four-level safety response logic; the response delay for hazardous commands did not exceed 200ms. The machine-sensor collaborative prediction sampling and scheduling module pre-set operating condition switching thresholds: predicting ground clearance when the lifting height reaches 10.8 meters, predicting entry into the obstructed section when the trolley is 10 meters from the obstructed area, and predicting positioning when 5 meters from the positioning area.
[0065] After the hoisting operation began, the initial 5G signal in the work area was good. The system defaulted to using the ground-based 5G private network link for data transmission. Based on the crane's lifting parameters, the sampling and scheduling module increased the sampling frequency to 800Hz 200ms before liftoff to fully collect the impact stress data at the moment of liftoff. When the component was hoisted to a position 10 meters away from the first set of frames, the module anticipated entering a signal obstruction area and prepared for link switching in advance. When the component entered the obstruction area and the ground signal strength dropped to -92dBm, the communication module immediately and automatically switched to the UAV relay link with a switching delay of 42ms. The monitoring data was forwarded to the ground receiving station via the UAV, and there was no data interruption throughout the process. After the component left the obstruction area, the ground signal recovered to -75dBm, and the system automatically switched back to the 5G private network link, dynamically switching the optimal transmission channel based on signal quality throughout the entire process.
[0066] During hoisting, the intelligent data processing unit calculates the additional stress and deformation of the component in real time, continuously assessing the safety level. When the component passes through an open windy area, the instantaneous lateral wind load causes the stress on one side of the component to rise rapidly, reaching 87% of the allowable value, triggering the warning threshold. The system triggers an on-site audible and visual alarm and initiates active cable tension adjustment, fine-tuning the length of each sling to distribute the load, gradually reducing the component stress. A few seconds later, the wind load increases further, and the stress rises to 96% of the allowable value, reaching the danger threshold. The graded warning unit simultaneously triggers two interventions within 180ms: first, it sends a deceleration command to the crane control system, and the crane automatically reduces its travel speed to 30% of its original speed; second, it increases the cable tension adjustment range to quickly reduce the local stress peak. The entire emergency response process requires no manual intervention, the component stress quickly returns to a safe range, no structural damage occurs, and work is not completely halted. Normal hoisting speed resumes after the wind load weakens.
[0067] The hoisting operation traversed three areas with signal obstruction, with zero data interruption and zero packet loss. Through multi-dimensional linkage, it successfully withstood a strong wind load impact, verifying the reliability of dual-path communication and multi-dimensional linkage control.
[0068] Comparison with Example 3: A conventional monitoring system using single-channel 5G transmission was used in the same site for hoisting operations. Lacking cable tension adjustment capabilities, it only supported manual alarms. Under the same hoisting path, there were four signal interruptions, with the longest single interruption lasting 8 seconds, indicating a significant monitoring blind spot. When strong wind loads caused stress to exceed limits, the system only sent alarm messages to on-site personnel, requiring manual crane operation for adjustment. The average emergency response time was approximately 4 seconds, missing the optimal intervention window. The peak stress of the components exceeded the limit by approximately 8%, posing a risk of structural damage. Furthermore, the only way to address the risk was through a complete shutdown, resulting in a delay of approximately 15 minutes per stoppage. In contrast, the dual-channel redundant communication scheme in this embodiment achieves blind-spot-free monitoring throughout the entire hoisting process in complex sites, with 100% data continuity. The multi-dimensional linkage of cable tension adjustment and crane control reduces the emergency response time from seconds to milliseconds, while simultaneously reducing the number of shutdowns, significantly improving the safety redundancy and construction efficiency of the hoisting operation.
[0069] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.
Claims
1. A real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in ultra-high voltage substations, characterized in that, It includes distributed sensing node units, data acquisition and transmission units, intelligent data processing units, and hierarchical early warning and linkage control units; The distributed sensing node unit includes several fixed composite sensing nodes and portable supplementary sensing nodes. Both types of nodes integrate fiber optic stress sensing modules and MEMS inertial measurement modules. The bottom of each node is equipped with a temperature-sensitive adaptive magnetic-photocuring composite fixing structure. The system is equipped with a movable binocular scanning unit to acquire the full-field geometry and deformation data of the component, identify the deformation concentration areas not covered by the fixed measuring points, and guide the portable supplementary sensing nodes to dynamically supplement the points and connect to the monitoring network. The data acquisition and transmission unit includes a ground acquisition host, an air-to-ground dual-path redundant communication module, and a machine-sensor collaborative prediction sampling and scheduling module. The machine-sensor collaborative prediction sampling and scheduling module accesses the crane's operating parameters, predicts the switching nodes of the hoisting conditions based on the working condition state machine, and adjusts the sampling frequency of the sensor nodes in advance. The intelligent data processing unit includes a residual stress in-situ recalibration module, a finite element model in-situ self-calibration module, and a stress-deformation coupling inversion module, which are used to correct the actual stress baseline of the component and the calculation model, remove the additional stress from the hoisting and invert the stress and deformation distribution of the component in the whole field. The graded early warning and linkage control unit is connected to both the crane control system and the multi-point sling force adjustment mechanism, and is used to trigger graded early warning and multi-dimensional operation intervention based on the safety assessment results.
2. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The temperature-sensitive adaptive magnetic-photocurable composite fixing structure includes a neodymium iron boron strong magnetic adsorption seat at the bottom of the node housing, a butyl rubber modified pressure-sensitive adhesive layer arranged in the center of the adsorption seat, an acrylic ester ultraviolet light curable adhesive layer arranged in the outer annular closed adhesive groove, and an annular micro ultraviolet LED light source with a wavelength of 365nm embedded in the inner side of the annular adhesive groove. The node has a built-in ceramic micro heating element. Under low temperature conditions, it triggers the ultraviolet light source to accelerate the curing of the photocurable adhesive. Under high temperature conditions, the bonding strength is compensated by the pressure-sensitive adhesive layer. When removing the device, the heating element is activated to heat and de-adhere the adhesive, achieving non-destructive peeling.
3. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The movable binocular scanning unit consists of two structured light binocular cameras. Before hoisting, the initial three-dimensional shape of the component is obtained by scanning. During the hoisting process, the displacement and deformation data of the entire field are obtained by scanning at preset intervals. Areas with deformation gradients higher than preset thresholds are marked as deformation concentration areas, and the position coordinates are pushed to the on-site terminal to guide the supplementary points. After being deployed, the portable supplementary sensor node automatically connects to the data acquisition network via wireless protocol.
4. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The machine-sensor collaborative prediction sampling and scheduling module synchronously collects parameters such as crane winch speed, lifting height, hook load, and trolley travel speed to construct a state machine that includes four working conditions: lifting off the ground, stable aerial transfer, attitude adjustment, and positioning docking. When the operating parameters reach the critical threshold for switching to the corresponding operating condition, the sampling frequency of the sensor node is adjusted 200ms in advance. The sampling frequency is 500~1000Hz during the lifting and docking phase, 100~300Hz during the attitude adjustment phase, and 10~30Hz during the smooth aerial transfer phase.
5. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The execution steps of the residual stress in-situ recalibration module are as follows: S1: Collect initial three-dimensional topographic data of the component and initial strain data of each measuring point before hoisting; S2: Based on the theory of elastic-plastic residual stress release, establish the mapping relationship between component morphology deviation and residual stress redistribution. S3: Using the factory residual stress field as the initial iteration field, combined with the measured initial strain and morphological deviation, the in-situ residual stress baseline of the component is obtained through iterative correction. S4: In real-time monitoring, the total strain data is subtracted from the in-situ residual strain baseline of the corresponding measuring point to obtain the additional strain of hoisting and convert it into additional stress of hoisting.
6. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The execution steps of the finite element model in-situ self-calibration module are as follows: S1: Import component theory design finite element initial model; S2: Map the actual geometric dimensions, wall thickness deviation, and assembly form and position deviation obtained from binocular scanning to the initial model mesh, and use mesh deformation technology to correct the geometric model; S3: Correct the model boundary constraints based on the actual connection form of the assembled nodes on site; S4: Verify the accuracy of the model by measuring the first-order natural frequency. If the error is less than 5%, generate a calculation benchmark model.
7. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The stress-deformation coupled inversion module is based on the in-situ calibrated finite element model. It constructs a weighted objective function containing four types of constraints: fixed measuring point stress, supplementary measuring point stress, nodal attitude deformation, and binocular full-field deformation. The function is solved using an improved Gauss-Newton iteration method with an adaptive step size, and the iteration convergence threshold is 10. -3 The final output is a cloud map of the full-field stress field and deformation field of the component.
8. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The multi-point sling force adjustment mechanism includes sling force sensors and hydraulic adjustment cylinders installed on each sling; The system pre-stores a cable force-stress sensitivity matrix calculated based on a benchmark model. When it detects local stress concentration caused by uneven cable force, it calls the sensitivity matrix to calculate the incremental adjustment of each cable and controls the hydraulic cylinder to extend and retract to balance the load.
9. The real-time stress-deformation monitoring system for the entire process of hoisting ultra-large components in UHV substations according to claim 1, characterized in that: The graded early warning and linkage control unit is set with four safety levels: normal, attention, warning, and danger. The "Attention" level sends observation prompts to on-site technicians; the "Early Warning" level triggers on-site audible and visual alarms and initiates cable tension fine-tuning; and the "Danger" level simultaneously triggers crane deceleration / pause commands and significant cable tension adjustments.
