Integrated monitoring system for settlement and verticality of tower crane foundation

Abstract

This invention discloses an integrated monitoring system for tower crane foundation settlement and verticality, belonging to the field of tower crane safety monitoring technology. Currently, tower crane settlement and verticality are mostly detected independently, resulting in weak data correlation and an inability to effectively distinguish the causes of tower tilt, leading to low monitoring accuracy and poor early warning reliability. This system deploys four displacement sensors at the four corners of the tower crane foundation and a dual-axis tilt sensor at the bottom of the tower. Settlement and tilt angle data are simultaneously collected by a data acquisition unit. The processor, based on initial parameters, sensor spacing, and preset thresholds, determines whether foundation settlement and tower tilt angle exceed limits, and performs coupling verification by comparing the vector difference between the theoretical and measured tilt angles corresponding to uneven settlement. This invention achieves integrated and accurate monitoring of tower crane settlement and verticality, improving the ability to identify tower crane safety risks, and is suitable for routine safety monitoring of tower cranes in various construction projects.

Description

Technical Field

[0001] This invention relates to the field of tower crane safety monitoring technology. More specifically, this invention relates to an integrated monitoring system for tower crane foundation settlement and verticality. Background Technology

[0002] In the safety management of tower crane operation at construction sites, foundation settlement and tower verticality are two core safety indicators that require mandatory monitoring. Currently, the industry generally adopts a separate monitoring architecture, where foundation settlement monitoring and tower verticality monitoring use independent equipment systems and collect data independently according to their respective sampling periods.

[0003] In terms of settlement monitoring, conventional methods include manual leveling or the deployment of single-point settlement sensors. Manual measurement is affected by weather and operator skill, with sampling intervals typically ranging from 1 to 7 days, making continuous real-time monitoring impossible. Single-point sensors can only acquire local settlement data and cannot fully reflect the differentiated settlement characteristics at the four corners of the foundation. Tower verticality monitoring often uses independently installed tilt sensors, with sampling cycles typically ranging from 1 to 10 minutes, which cannot be precisely matched with the sampling time of settlement data.

[0004] Because the two types of data are sampled asynchronously, existing technologies cannot establish a correspondence between the foundation settlement state and the tower tilt state at the same moment. The industry has attempted to achieve data synchronization through software timestamp alignment, but due to clock drift within each device, the synchronization error is typically on the order of hundreds of milliseconds to several seconds. For equipment like tower cranes with frequently changing dynamic loads, this error is sufficient to distort the results of coupled calculations.

[0005] From a mechanical perspective, tower tilting can be caused by two independent factors: uneven foundation settlement and elastic deformation of the tower under load. Existing separate monitoring systems lack a linkage verification mechanism for these two types of parameters, making it impossible to separate the tilt components corresponding to the two factors. They can only issue warnings based on a single parameter exceeding the standard, which easily leads to false alarms or missed alarms. Some solutions attempt to establish a correlation between settlement and tilt through manual calculation, but the loads at construction sites change in real time, and manual calculations cannot meet the real-time requirements. Summary of the Invention

[0006] Another objective of this invention is to provide an integrated monitoring system for tower crane foundation settlement and verticality. Existing tower crane foundation settlement and tower verticality monitoring systems use a separate monitoring architecture, resulting in asynchronous data acquisition. This makes it impossible to establish a correspondence between the two types of parameters at the same time, and warnings can only be issued based on a single parameter exceeding the standard. It is difficult to distinguish the real cause of tower tilting, and the reliability of the warning is insufficient.

[0007] The lack of a standardized method for calculating the theoretical inclination angle from uneven foundation settlement has resulted in a lack of unified basis for comparing the theoretical and measured inclination angles. This leads to significant deviations in the vector difference calculation results, rendering the coupling verification meaningless.

[0008] Existing technologies do not consider the influence of the elastic deformation of the tower body caused by the load on the measured tilt angle. The measured tilt angle contains an elastic component that is unrelated to the foundation settlement, and cannot accurately reflect the true tilt state caused by uneven foundation settlement.

[0009] Existing technologies neglect the fact that the tilting state of the tower body changes the force boundary conditions of elastic deformation, resulting in inherent systematic errors in the query results of elastic tilt angle and load curve, and the calculation accuracy of the net measured tilt angle is difficult to meet engineering requirements.

[0010] Existing technologies calculate the rigid body tilt angle based solely on settlement changes, without considering the foundation tilt that existed during the initial installation of the tower crane. This results in inaccurate rigid body tilt angles in the input coupled correction model, significantly reducing the correction effect.

[0011] The existing technology does not deduct the elastic deformation angle of the tower body in the initial state, which results in the calculation of the change in elastic deformation including the initial value deviation and cannot accurately reflect the actual change in elastic deformation during the monitoring period.

[0012] Existing technologies all assume that foundation settlement is planar deformation. When the foundation undergoes non-coplanar torsional deformation, the conventional planar conversion model completely fails, resulting in incorrect theoretical tilt angles and triggering a large number of false alarms.

[0013] Existing technologies use software timestamp alignment to achieve data synchronization. However, due to the drift of the internal clocks of each device, the synchronization error can reach hundreds of milliseconds to several seconds, which cannot meet the time synchronization requirements of high-precision coupled computing.

[0014] Existing monitoring systems can only trigger audible and visual alerts, cannot fully preserve the original data before and after abnormal events, and cannot predict future risk trends, thus failing to provide a scientific basis for on-site emergency response.

[0015] To achieve these objectives and other advantages according to the present invention, an integrated monitoring system for tower crane foundation settlement and verticality is provided, comprising: Four displacement sensors are fixed at the four corners of the tower crane foundation. Each displacement sensor is used to measure the settlement of the corresponding corner relative to a reference point. A dual-axis tilt sensor is fixed to the outer side of the standard section at the bottom of the tower crane tower body to measure the actual tilt angle of the tower body along two orthogonal directions; A data acquisition unit is electrically connected to four displacement sensors and a dual-axis tilt sensor to synchronously acquire various settlement amounts and measured tilt angles. A processor is connected in communication with the data acquisition unit. The processor stores the initial settlement of each displacement sensor, the initial measured tilt angle of the dual-axis tilt sensor, the horizontal distance data between the four displacement sensors, a preset settlement threshold, a preset tilt angle threshold, and a preset difference threshold. The horizontal distance data between the four displacement sensors includes the distance between adjacent sensors along two orthogonal directions. The processor calculates the settlement change at each corner based on the settlement values ​​collected from the four displacement sensors and the initial settlement value corresponding to each displacement sensor, and selects the maximum value among the settlement changes to compare with the preset settlement threshold. The processor subtracts the measured tilt angle of the currently acquired dual-axis tilt sensor from the initial measured tilt angle to obtain the measured tilt angle change, which includes components in two orthogonal directions. The processor calculates the magnitude of the measured tilt angle change and compares it with a preset tilt angle threshold. The processor calculates the theoretical change in tilt angle caused by uneven settlement based on the settlement changes at each corner and the horizontal distance data between the four displacement sensors. The theoretical change in tilt angle includes components in two orthogonal directions. The processor uses the theoretical change in tilt angle and the measured change in tilt angle as vectors, calculates the magnitude of the vector difference, uses the magnitude as the difference value, and compares the difference value with a preset difference threshold. When the maximum value of the settlement change exceeds the preset settlement threshold, or the magnitude of the measured tilt angle change exceeds the preset tilt angle threshold, or the difference exceeds the preset difference threshold, the processor outputs an alarm signal.

[0016] Preferably, when the processor calculates the theoretical change in inclination caused by uneven settlement, it calculates the ratio of the settlement difference to the horizontal distance in the corresponding direction along two orthogonal directions to obtain the theoretical inclination components in the two directions.

[0017] Preferably, the processor also stores the current working height of the tower crane, as well as the tower body elastic tilt angle and load curve obtained in advance through finite element analysis or on-site calibration; The processor acquires the tower crane's working status in real time, including lifting capacity, luffing range, and slewing angle. Based on the current working height and load status, the processor retrieves the additional tilt component caused by the elastic deformation of the tower body from the tower body elastic tilt angle and load curve; Before comparing the theoretical tilt angle change with the measured tilt angle change using vector difference, the processor first subtracts the additional tilt angle component from the measured tilt angle change to obtain the net measured tilt angle change after deducting elastic deformation. The processor calculates the magnitude of the vector difference between the net measured change in tilt angle and the theoretical change in tilt angle, uses this magnitude as the difference value, and compares it with a preset difference threshold.

[0018] Preferably, the processor also stores a coupling correction model, which is used to calculate the amount of correction required to be applied to the elastic tilt angle and load curve of the tower body under different rigid body tilt angles; After the processor finds the additional tilt angle component from the elastic tilt angle and load curve of the tower, it further inputs the theoretical tilt angle change obtained by the current calculation as the rigid tilt angle of the tower into the coupled correction model to obtain the corrected additional tilt angle component. The processor then subtracts the corrected additional tilt component from the measured tilt angle change to obtain the net measured tilt angle change after deducting the effects of elastic deformation and rigid-elastic coupling.

[0019] Preferably, the processor also stores an initial rigid body tilt angle, which is calculated from the initial settlement of the four displacement sensors in the initial state. Before inputting the currently calculated theoretical tilt angle change into the coupled correction model, the processor first adds the initial rigid body tilt angle to the theoretical tilt angle change to obtain the current absolute rigid body tilt angle. The processor then inputs the current absolute rigid body tilt angle into the coupled correction model to obtain the corrected additional tilt angle component.

[0020] Preferably, the processor also stores the initial elastic deformation tilt angle, which is obtained by looking up the tower elastic tilt angle and load curve based on the tower working height, initial load and initial rigid body tilt angle in the initial state and then correcting it through the coupling correction model. The processor subtracts the initial elastic deformation tilt angle from the corrected additional tilt angle component to obtain the elastic deformation change. Then, it subtracts the elastic deformation change from the measured tilt angle change to obtain the net measured tilt angle change after deducting the elastic deformation change.

[0021] Preferably, the processor also stores a preset non-coplanar error threshold; The processor uses the least squares method to fit an optimal fitting plane based on the settlement changes at the four corners and the horizontal distance data between the four displacement sensors. It also calculates the absolute value of the residual settlement difference at the four corners relative to the optimal fitting plane and takes the maximum value of the absolute value of the four residual settlement differences as the non-coplanar error. When the non-coplanar error exceeds the preset non-coplanar error threshold, the processor outputs a basic torsion alarm signal and skips the operation of comparing the vector difference between the theoretical tilt angle change and the measured tilt angle change at the current sampling time. At the same time, it skips the alarm judgment operation of the difference exceeding the threshold corresponding to the vector difference comparison and continues to execute the operation of comparing the maximum value of the settlement change with the preset settlement threshold, and continues to execute the operation of comparing the magnitude of the measured tilt angle change with the preset tilt angle threshold. When the non-coplanar error does not exceed the preset non-coplanar error threshold, the processor extracts the theoretical tilt angle change from the slope of the two directions of the best-fit plane, and uses the theoretical tilt angle change and the measured tilt angle change as the modulus of the vector difference.

[0022] Preferably, a hardware trigger synchronization circuit is also provided between the data acquisition unit and the processor; the hardware trigger synchronization circuit sends an edge trigger pulse to the four displacement sensors and the dual-axis tilt sensor at the beginning of each data acquisition cycle. Four displacement sensors and a dual-axis tilt sensor are synchronously sampled and latched on the rising edge of the same pulse. The latched signal initiates analog-to-digital conversion on the rising edge of the same pulse or the falling edge that follows it. The digital values ​​output by the four displacement sensors and the dual-axis tilt sensor after analog-to-digital conversion are all associated with the period identifier of the current trigger pulse; The processor combines four settlement measurements and one tilt measurement collected under the same trigger pulse into a set of synchronous measurement data based on the period identifier, so that the maximum acquisition time difference between the settlement measurements of the four displacement sensors and the tilt measurements of the dual-axis tilt sensor is less than 1 microsecond.

[0023] Preferably, the processor packages the raw settlement data measured by the four displacement sensors, the raw tilt data measured by the dual-axis tilt sensor, the theoretical tilt change calculated by the processor, the measured tilt change, and the vector difference magnitude into an encrypted data packet according to a preset periodic upload cycle, and sends it to the cloud server via a wireless network. The cloud server runs a time-series-based deep learning model to predict the foundation settlement trend and tilt development trend in the next 24 hours based on the current data characteristics, and sends the prediction results and corresponding level-three decision suggestions back to the processor display screen on site for display. Furthermore, each time the processor outputs an alarm signal, in addition to triggering the audible and visual alarm, it also packages the following data from 30 seconds before the alarm to 10 seconds after the alarm into an encrypted event data packet: raw settlement data measured by four displacement sensors, raw tilt data measured by dual-axis tilt sensors, theoretical tilt change calculated by the processor, measured tilt change, and vector difference magnitude. The processor sends encrypted event data packets to the cloud server via a wireless network; the cloud server runs the time-series-based deep learning model, which uses historical normal data and known accident cases as training sets, predicts the foundation settlement trend and tilt angle development trend in the next 24 hours based on the data characteristics of this alarm event, and provides three-level decision suggestions, including continuing to observe, recommending reduced load and immediate suspension of work and reinforcement. The cloud server transmits the three-level decision recommendations back to the on-site processor display screen.

[0024] The present invention has at least the following beneficial effects: This invention achieves integrated monitoring of foundation settlement and tower verticality by deploying displacement sensors at the four corners of the tower crane foundation and dual-axis tilt sensors at the bottom of the tower. It uses synchronous acquisition technology to establish the correspondence between the two types of parameters at the same time, and distinguishes the causes of tower tilt through triple alarm judgment logic, effectively reducing false alarms and missed alarms, and significantly improving the accuracy and reliability of tower crane safety alarms.

[0025] This invention obtains the theoretical dip angle component by calculating the ratio of the settlement difference to the corresponding horizontal distance along two orthogonal directions. This method has clear physical meaning, simple calculation process, and low engineering implementation difficulty. It unifies the calculation standard of theoretical dip angle, provides an accurate basis for vector comparison between theoretical dip angle and measured dip angle, and ensures the reliability of the coupling verification results.

[0026] This invention pre-establishes the tower body elastic tilt angle and load curve, and combines it with the real-time acquired tower crane working height, lifting capacity, luffing amplitude and slewing angle to accurately calculate the additional tilt angle component caused by the elastic deformation of the tower body. After subtracting this component from the measured tilt angle, the net measured tilt angle is obtained, which eliminates the interference of load changes on the tilt monitoring results and greatly improves the accuracy of tilt cause identification.

[0027] This invention establishes a coupled correction model that considers the mutual influence between the rigid body tilt and elastic deformation of the tower. The theoretical tilt angle change obtained from the current calculation is used as the input of the rigid body tilt angle into the model to correct the elastic tilt angle query result. This eliminates the systematic error caused by the change of the force boundary conditions and further improves the calculation accuracy of the net measured tilt angle.

[0028] This invention obtains the current absolute rigid body tilt angle by adding the initial rigid body tilt angle calculated from the initial settlement of four displacement sensors in the initial state to the current theoretical tilt angle change. This fully reflects the actual rigid body tilt state of the tower and provides accurate input parameters for the coupled correction model, ensuring the effectiveness of the correction effect throughout the entire life cycle of the tower crane.

[0029] This invention pre-calculates the elastic deformation tilt angle of the tower body in the initial state, subtracts the initial value from the corrected additional tilt angle component to obtain the change in elastic deformation, and then subtracts the change from the measured tilt angle change, thereby eliminating the influence of the elastic deformation in the initial state on the monitoring results, so that the net measured tilt angle can accurately reflect the tilt change caused by foundation settlement during the current monitoring period.

[0030] This invention uses the least squares method to fit the best-fit plane of foundation settlement, calculates the residual settlement difference of the four corners relative to this plane, and takes the maximum value as the non-coplanar error. It can accurately identify the torsional deformation condition of the foundation. At this time, invalid vector comparison operations are automatically skipped to avoid false alarms caused by model failure, while retaining independent alarm functions for settlement and tilt angle.

[0031] This invention achieves microsecond-level synchronous sampling, latching, and analog-to-digital conversion by setting up a hardware-triggered synchronization circuit to send edge-triggered pulses to all sensors simultaneously in each acquisition cycle. This eliminates clock drift errors caused by software synchronization, ensures a strict correspondence between settlement data and tilt angle data at the same moment, and provides a reliable data foundation for high-precision coupled calculations.

[0032] This invention, by setting up a periodic upload mechanism and an alarm-triggered upload mechanism, can not only conduct daily inspections and predictions of progressively accumulating risks, but also automatically package and encrypt complete raw data and calculated data from 30 seconds before the alarm to 10 seconds after the alarm and upload them to the cloud. Using a deep learning model trained on historical normal data and accident cases, it can predict risk trends for the next 24 hours and provide three levels of decision-making suggestions, thus realizing the traceability of abnormal events and scientific emergency response.

[0033] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0034] Figure 1 This is a flowchart of an integrated monitoring system for tower crane foundation settlement and verticality, as described in one of the technical solutions of the present invention. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to embodiments, so that those skilled in the art can implement it based on the description.

[0036] like Figure 1 As shown, the present invention provides an integrated monitoring system for tower crane foundation settlement and verticality, comprising: Four displacement sensors are fixed at the four corners of the tower crane foundation. Each displacement sensor is used to measure the settlement of the corresponding corner relative to a reference point. A dual-axis tilt sensor is fixed to the outer side of the standard section at the bottom of the tower crane tower body to measure the actual tilt angle of the tower body along two orthogonal directions; A data acquisition unit is electrically connected to four displacement sensors and a dual-axis tilt sensor to synchronously acquire various settlement amounts and measured tilt angles. A processor is connected in communication with the data acquisition unit. The processor stores the initial settlement of each displacement sensor, the initial measured tilt angle of the dual-axis tilt sensor, the horizontal distance data between the four displacement sensors, a preset settlement threshold, a preset tilt angle threshold, and a preset difference threshold. The horizontal distance data between the four displacement sensors includes the distance between adjacent sensors along two orthogonal directions. The processor calculates the settlement change at each corner based on the settlement values ​​collected from the four displacement sensors and the initial settlement value corresponding to each displacement sensor, and selects the maximum value among the settlement changes to compare with the preset settlement threshold. The processor subtracts the measured tilt angle of the currently acquired dual-axis tilt sensor from the initial measured tilt angle to obtain the measured tilt angle change, which includes components in two orthogonal directions. The processor calculates the magnitude of the measured tilt angle change and compares it with a preset tilt angle threshold. The processor calculates the theoretical change in tilt angle caused by uneven settlement based on the settlement changes at each corner and the horizontal distance data between the four displacement sensors. The theoretical change in tilt angle includes components in two orthogonal directions. The processor uses the theoretical change in tilt angle and the measured change in tilt angle as vectors, calculates the magnitude of the vector difference, uses the magnitude as the difference value, and compares the difference value with a preset difference threshold. When the maximum value of the settlement change exceeds the preset settlement threshold, or the magnitude of the measured tilt angle change exceeds the preset tilt angle threshold, or the difference exceeds the preset difference threshold, the processor outputs an alarm signal.

[0037] In the above technical solution, four wire-type displacement sensors can be used as displacement sensors, respectively fixed to the pre-embedded steel plates at the four corners of the tower crane foundation. The measuring end of each sensor is connected to a pre-set reference point, which is either a deeply buried reference pile independent of the tower crane foundation or a reference structure stable on the ground. MEMS dual-axis tilt sensors can be used as tilt measuring elements, fixed to the outer side of the first standard section at the bottom of the tower crane, with the mounting plane perpendicular to the tower axis. An industrial-grade multi-channel data acquisition unit can be used for data acquisition, electrically connected to the four displacement sensors and the dual-axis tilt sensor via shielded cables. An embedded ARM processor can be used as the computing core, communicating with the data acquisition unit via an RS485 bus. The processor pre-stores the initial settlement of each displacement sensor, the initial measured tilt angle of the dual-axis tilt sensor, and the horizontal distance data between the four displacement sensors. The horizontal distance data includes the distance between adjacent sensors along two orthogonal directions. It also stores a preset settlement threshold of 10mm, a preset tilt angle threshold of 1 / 1000, and a preset difference threshold of 0.2 / 1000. The initial settlement of each displacement sensor and the initial measured tilt angle of the dual-axis tilt sensor stored in the processor are values ​​calibrated and recorded when the tower crane is installed, the foundation is stable, and the tower body is vertical.

[0038] The data acquisition unit synchronously acquires the settlement data from four displacement sensors and the measured tilt angle data from a dual-axis tilt sensor according to a set sampling period, and transmits the data to the processor. The processor subtracts the corresponding initial settlement data from the currently acquired settlement data of each displacement sensor to calculate the settlement change at each corner, and selects the maximum value to compare with a preset settlement threshold. The processor subtracts the initial measured tilt angle from the currently acquired measured tilt angle to obtain the measured tilt angle change containing two orthogonal directional components. It calculates the square root of the sum of the squares of the two orthogonal directional components to obtain the magnitude of the measured tilt angle change, and compares it with a preset tilt angle threshold. Based on the settlement change at each corner and the horizontal distance data, the processor calculates the theoretical tilt angle change caused by uneven settlement, containing two orthogonal directional components. It calculates the square root of the sum of the squares of the differences between the theoretical and measured tilt angles in the two directions to obtain the magnitude of the vector difference, and compares this magnitude with a preset difference threshold. If any of the above comparison results exceed the limit, the processor outputs an alarm signal.

[0039] By adopting this technical solution, the present invention can realize integrated synchronous monitoring of tower crane foundation settlement and tower verticality, establish the correspondence between the two types of parameters at the same time, realize multi-dimensional safety verification through triple alarm judgment logic, effectively distinguish different causes of tower tilt, reduce the occurrence of false alarms and missed alarms, and improve the reliability of tower crane operation safety early warning.

[0040] In another technical solution, when the processor calculates the theoretical change in tilt angle caused by uneven settlement, it calculates the ratio of the settlement difference to the horizontal distance in the corresponding direction along two orthogonal directions to obtain the theoretical tilt angle components in the two directions.

[0041] In the above technical solution, four displacement sensors are arranged along two orthogonal directions of the tower crane foundation, corresponding to the length and width directions of the foundation, respectively. The horizontal distance between adjacent sensors can be 4m, 5m, or 6m, with the specific value determined according to the actual dimensions of the tower crane foundation. The sampling frequency of the data acquisition unit can be set to 1Hz to ensure timely capture of changes in foundation settlement and tower tilt. The processor pre-stores identifiers for the two orthogonal directions, corresponding to the longitudinal and transverse directions of the tower crane, respectively.

[0042] When calculating the theoretical tilt angle change caused by uneven settlement, the processor first extracts the settlement changes of two adjacent displacement sensors in two orthogonal directions. The difference between the settlement changes of the two sensors is then calculated to obtain the settlement difference in the corresponding direction. This settlement difference is then divided by the horizontal distance between adjacent sensors in that direction to obtain the theoretical tilt angle component in that direction. The theoretical tilt angle components in the two orthogonal directions together form the theoretical tilt angle change vector, which is used for subsequent calculations of the vector difference between the theoretical and measured tilt angle changes.

[0043] By adopting this technical solution, the present invention clarifies a standardized calculation method for the theoretical tilt angle change, which has a clear physical meaning, a simple calculation process, and low engineering implementation difficulty. It provides a unified and accurate calculation basis for the vector comparison between the theoretical tilt angle and the measured tilt angle, and ensures the reliability of the coupling verification results.

[0044] In another technical solution, the processor also stores the current working height of the tower crane, as well as the tower body elastic tilt angle and load curve obtained in advance through finite element analysis or on-site calibration; The processor acquires the tower crane's working status in real time, including lifting capacity, luffing range, and slewing angle. Based on the current working height and load status, the processor retrieves the additional tilt component caused by the elastic deformation of the tower body from the tower body elastic tilt angle and load curve; Before comparing the theoretical tilt angle change with the measured tilt angle change using vector difference, the processor first subtracts the additional tilt angle component from the measured tilt angle change to obtain the net measured tilt angle change after deducting elastic deformation. The processor calculates the magnitude of the vector difference between the net measured change in tilt angle and the theoretical change in tilt angle, uses this magnitude as the difference value, and compares it with a preset difference threshold.

[0045] In the above technical solution, the processor pre-stores the current working height of the tower crane, as well as the elastic tilt angle and load curve of the tower body. This curve is obtained through finite element analysis or on-site calibration. During finite element analysis, a three-dimensional solid model of the tower body is established, and load combinations with different lifting capacities, luffing amplitudes, and slewing angles are applied to calculate the elastic tilt angle at the bottom of the tower body at different working heights. During on-site calibration, it should be carried out under the condition that the tower crane foundation is absolutely level and there is no settlement (e.g., on a rigid test bench). The tower body tilt angle is measured under the following conditions: no load, 25% rated load, 50% rated load, 75% rated load, and 100% rated load. The difference between the tilt angle measured under each load condition and the tilt angle under no-load condition is the pure elastic deformation tilt angle caused by the load. Based on this, the curve corresponding to the load and elastic tilt angle is plotted. If calibration on a rigid test bench is not possible due to on-site limitations, the no-load tilt angle can be measured first, and then the rigid body tilt component caused by foundation settlement can be compensated and corrected before extracting the elastic deformation component. The tower crane's current working height can be obtained in real time using its built-in height sensor; the lifting weight can be obtained in real time using a weight sensor installed on the hoisting mechanism; the luffing amplitude can be obtained in real time using a displacement sensor installed on the luffing mechanism; and the slewing angle can be obtained in real time using an encoder installed on the slewing mechanism. The output signals from all these sensors are connected to a data acquisition unit, which then synchronously transmits them to the processor.

[0046] The processor acquires the tower crane's working status parameters in real time. Based on the current working height and the moment load calculated from the lifting capacity and luffing amplitude, it retrieves the additional tilt component caused by the tower's elastic deformation from the pre-stored tower elastic tilt angle and load curves. This component also contains components in two orthogonal directions. Before comparing the theoretical tilt angle change with the measured tilt angle change using vector difference, the processor first subtracts the additional tilt angle component from the measured tilt angle change to obtain the net measured tilt angle change after deducting elastic deformation. Then, it calculates the magnitude of the vector difference between the net measured tilt angle change and the theoretical tilt angle change, and compares this magnitude with a preset difference threshold.

[0047] By adopting this technical solution, the present invention can effectively separate the component caused by load elastic deformation in the tower tilt, eliminate the interference of tower crane load changes on the tilt monitoring results, and enable the monitoring results to accurately reflect the true tilt state caused by uneven foundation settlement, thus greatly improving the accuracy of tilt cause identification.

[0048] In another technical solution, the processor also stores a coupling correction model, which is used to calculate the amount of correction required to be applied to the elastic tilt angle and load curve of the tower body under different rigid body tilt angles. After the processor finds the additional tilt angle component from the elastic tilt angle and load curve of the tower, it further inputs the theoretical tilt angle change obtained by the current calculation as the rigid tilt angle of the tower into the coupled correction model to obtain the corrected additional tilt angle component. The processor then subtracts the corrected additional tilt component from the measured tilt angle change to obtain the net measured tilt angle change after deducting the effects of elastic deformation and rigid-elastic coupling.

[0049] In the above technical solution, the processor pre-stores a coupling correction model, which is obtained through finite element simulation analysis or field calibration tests. The model is stored in the processor in the form of a lookup table or a linear function expression. During finite element analysis, a three-dimensional solid model of the tower body with the same elastic tilt angle and load curve as the model used to establish the tower body is adopted. Within the range of rigid body tilt slope (the ratio of settlement difference to sensor spacing) from 0 to 1 / 500 (corresponding to an angle of approximately 0° to 0.115°), the same load combination as the elastic tilt angle curve calibration is applied. The rigid body tilt state is simulated by constraining the displacement of the foundation nodes. The elastic tilt angle at the bottom of the tower body in the corresponding state is calculated. The ratio of this elastic tilt angle to the elastic tilt angle under the same load without rigid body tilt is calculated to obtain the elastic tilt angle correction amount corresponding to different rigid body tilt angles (in this embodiment, the correction amount is expressed in the form of a product coefficient). During on-site calibration, a controlled horizontal thrust can be applied to the side of the tower crane foundation (e.g., using a combination of hydraulic jacks and tension sensors applied at different heights on the tower body) to simulate the additional force boundary changes caused by foundation tilting, or a correction scale can be established entirely based on finite element analysis results. To verify the model's accuracy, the small initial tilt angle of the foundation can be measured using a precision level and inclinometer under the tower crane's unloaded state, and the least squares fitting can be performed using the finite element simulation data to obtain the functional relationship between the correction amount and the rigid body tilt angle. The normal range for the correction amount is 0.9 to 1.1. If the calculated correction amount is within this range, subsequent calculations are performed based on the corrected additional tilt angle component; if it exceeds this range, the processor outputs a model anomaly warning signal, suggesting that the model be recalibrated. Before the recalibration is completed, the model should temporarily revert to the uncorrected state, i.e., the uncoupled additional tilt angle component should be directly deducted from the measured tilt angle change to ensure that the monitoring system can still output usable data during the calibration transition period.

[0050] After retrieving the additional tilt angle component from the tower's elastic tilt angle and load curve, the processor further inputs the currently calculated theoretical tilt angle change as the rigid body tilt angle of the tower into the coupled correction model, obtaining correction values ​​in two orthogonal directions. The retrieved additional tilt angle component is then multiplied by its corresponding correction value to obtain the corrected additional tilt angle component. Finally, the processor subtracts this corrected additional tilt angle component from the measured tilt angle change to obtain the net measured tilt angle change after deducting the effects of elastic deformation and rigid-elastic coupling.

[0051] By adopting this technical solution, the present invention takes into account the influence of the tower's rigid tilt state on the elastic deformation force boundary conditions, which can eliminate the systematic errors caused by changes in the force boundary, further improve the calculation accuracy of the net measured tilt angle change, and make the coupling verification results more reliable.

[0052] In another technical solution, the processor also stores an initial rigid body tilt angle, which is calculated from the initial settlement of the four displacement sensors in the initial state. Before inputting the currently calculated theoretical tilt angle change into the coupled correction model, the processor first adds the initial rigid body tilt angle to the theoretical tilt angle change to obtain the current absolute rigid body tilt angle. The processor then inputs the current absolute rigid body tilt angle into the coupled correction model to obtain the corrected additional tilt angle component.

[0053] In the above technical solution, the processor pre-stores the initial rigid body tilt angle. This angle is calculated from the initial settlement of the four displacement sensors when the tower crane is stationary under no-load conditions after the system installation and commissioning are completed. The calculation method is consistent with the calculation method of the theoretical tilt angle change. The initial rigid body tilt angle also includes components in two orthogonal directions, reflecting the initial tilt state of the foundation when the tower crane is installed.

[0054] Before inputting the currently calculated theoretical tilt angle change into the coupled correction model, the processor first adds the initial rigid body tilt angle and the theoretical tilt angle change in two orthogonal directions to obtain the current absolute rigid body tilt angle. This absolute rigid body tilt angle fully reflects the cumulative rigid body tilt state of the tower crane from installation completion to the current moment. The processor then inputs the current absolute rigid body tilt angle into the coupled correction model to obtain the corrected additional tilt angle component.

[0055] By adopting this technical solution, the present invention can fully reflect the actual rigid body tilt state of the tower, provide accurate input parameters for the coupled correction model, ensure the effectiveness of the correction effect throughout the entire life cycle of the tower crane, and avoid correction errors caused by ignoring the initial foundation tilt.

[0056] In another technical solution, the processor also stores the initial elastic deformation tilt angle, which is obtained from the tower elastic tilt angle and load curve based on the tower working height, initial load and initial rigid body tilt angle in the initial state, and then corrected by the coupling correction model. The processor subtracts the initial elastic deformation tilt angle from the corrected additional tilt angle component to obtain the elastic deformation change. Then, it subtracts the elastic deformation change from the measured tilt angle change to obtain the net measured tilt angle change after deducting the elastic deformation change.

[0057] In the above technical solution, the processor pre-stores the initial elastic deformation tilt angle. This tilt angle is obtained from the tower's elastic tilt angle and load curve based on the tower's working height, initial load, and initial rigid body tilt angle in the initial unloaded static state after system installation and commissioning, and then corrected by a coupled correction model. The initial elastic deformation tilt angle contains components in two orthogonal directions, reflecting the elastic deformation caused by the tower crane's own weight in the initial state.

[0058] After obtaining the corrected additional tilt component, the processor subtracts the pre-stored initial elastic deformation tilt angle from the corrected additional tilt component to obtain the change in elastic deformation from the initial state to the current time. Then, this change in elastic deformation is subtracted from the measured tilt change to obtain the net measured tilt change after deducting the elastic deformation change. This net measured tilt change only reflects the tower tilt change caused by uneven foundation settlement from the initial state to the current time.

[0059] By adopting this technical solution, the present invention can eliminate the influence of the elastic deformation of the tower body in the initial state on the monitoring results, so that the net measured tilt angle change can accurately reflect the tilt change caused by foundation settlement during the current monitoring period, thereby improving the stability and comparability of monitoring data during the long-term operation of the tower crane.

[0060] In another technical solution, the processor also stores a preset non-coplanar error threshold; The processor uses the least squares method to fit an optimal fitting plane based on the settlement changes at the four corners and the horizontal distance data between the four displacement sensors. It also calculates the absolute value of the residual settlement difference at the four corners relative to the optimal fitting plane and takes the maximum value of the absolute value of the four residual settlement differences as the non-coplanar error. When the non-coplanar error exceeds the preset non-coplanar error threshold, the processor outputs a basic torsion alarm signal and skips the operation of comparing the vector difference between the theoretical tilt angle change and the measured tilt angle change at the current sampling time. At the same time, it skips the alarm judgment operation of the difference exceeding the threshold corresponding to the vector difference comparison and continues to execute the operation of comparing the maximum value of the settlement change with the preset settlement threshold, and continues to execute the operation of comparing the magnitude of the measured tilt angle change with the preset tilt angle threshold. When the non-coplanar error does not exceed the preset non-coplanar error threshold, the processor extracts the theoretical tilt angle change from the slope of the two directions of the best-fit plane, and uses the theoretical tilt angle change and the measured tilt angle change as the modulus of the vector difference.

[0061] In the above technical solution, the processor pre-stores a preset non-coplanar error threshold of 2mm. The placement positions of the four displacement sensors form a rectangular plane. The processor uses the least squares method to perform plane fitting on the settlement changes at the four corners. Assuming that the plane equation contains three coefficients to be determined, the three coefficients are obtained by minimizing the sum of the squared residuals between the actual settlement values ​​at the four corners and the calculated values ​​on the plane, thereby determining the best fitting plane for the foundation settlement.

[0062] The processor calculates the difference between the actual settlement changes at the four corners and the settlement values ​​at the corresponding points on the best-fit plane, obtaining four residual settlement differences. The maximum absolute value is taken as the non-coplanar error. When the non-coplanar error exceeds a preset non-coplanar error threshold, the processor outputs a foundation torsion alarm signal, discards the theoretical tilt angle change data calculated in this sampling, does not store it in the processor for any subsequent calculations, and skips the operation of comparing the vector difference between the theoretical and measured tilt angle changes at the current sampling time. It also skips the corresponding threshold-exceeding alarm operation and continues to compare the maximum settlement change value with the measured tilt angle change magnitude. When the non-coplanar error does not exceed the threshold, the processor extracts the theoretical tilt angle change from the slopes of the two directions of the best-fit plane and then performs subsequent vector difference calculations.

[0063] By adopting this technical solution, the present invention can accurately identify the non-coplanar torsional deformation condition of tower crane foundation. When the foundation torsion causes the conventional planar conversion model to fail, it automatically skips the invalid vector comparison operation, avoiding false alarms caused by model failure. At the same time, it retains the independent alarm functions for foundation settlement and tower tilt angle, ensuring the safety of the monitoring system.

[0064] In another technical solution, a hardware trigger synchronization circuit is also set between the data acquisition unit and the processor; at the beginning of each data acquisition cycle, the hardware trigger synchronization circuit simultaneously sends an edge trigger pulse to the four displacement sensors and the dual-axis tilt sensor. Four displacement sensors and a dual-axis tilt sensor are synchronously sampled and latched on the rising edge of the same pulse. The latched signal initiates analog-to-digital conversion on the rising edge of the same pulse or the falling edge that follows it. The digital values ​​output by the four displacement sensors and the dual-axis tilt sensor after analog-to-digital conversion are all associated with the period identifier of the current trigger pulse; The processor combines four settlement measurements and one tilt measurement collected under the same trigger pulse into a set of synchronous measurement data based on the period identifier, so that the maximum acquisition time difference between the settlement measurements of the four displacement sensors and the tilt measurements of the dual-axis tilt sensor is less than 1 microsecond.

[0065] In the above technical solution, a hardware trigger synchronization circuit is set between the data acquisition unit and the processor. This function can be implemented using an FPGA synchronization trigger circuit. The output of the hardware trigger synchronization circuit is connected to the external trigger input of the four displacement sensors and the dual-axis tilt sensor, respectively. The width of the edge trigger pulse can be set to 1μs, and the rising edge serves as the sampling latch signal. All four displacement sensors and the dual-axis tilt sensor support the external trigger sampling latch function. Each sensor completes the latching of the analog signal instantaneously on the rising edge (the latching response time is on the nanosecond level, and the latching time difference between the sensors can be controlled within 1 microsecond through equal-length wiring and high-speed drive circuits). After latching is completed, analog-to-digital conversion is started. In this embodiment, the single analog-to-digital conversion time does not exceed 10μs. The "acquisition time difference" mentioned in this invention specifically refers to the difference in the sampling latching time and does not include the subsequent analog-to-digital conversion time.

[0066] At the start of each data acquisition cycle, the hardware trigger synchronization circuit simultaneously sends an edge trigger pulse to all four displacement sensors and the dual-axis tilt sensor. The four displacement sensors and the dual-axis tilt sensor are synchronously sampled and latched on the rising edge of the same pulse, and the latched signals initiate analog-to-digital conversion on the rising edge of the same pulse. The digital outputs from the four displacement sensors and the dual-axis tilt sensor after analog-to-digital conversion are all associated with the cycle identifier of the current trigger pulse. The cycle identifier is an incrementing 32-bit integer, and a counter automatically increments by 1 with each trigger pulse. Based on the cycle identifier, the processor combines the four settlement measurements and one tilt measurement collected under the same trigger pulse into a set of synchronous measurement data, ensuring that the maximum acquisition time difference between the settlement measurements from the four displacement sensors and the tilt measurement from the dual-axis tilt sensor is less than 1 μs.

[0067] By adopting this technical solution, the present invention can achieve microsecond-level synchronous sampling, completely eliminate the clock drift error present in the software timestamp alignment method, ensure the strict correspondence between settlement data and tilt angle data at the same time, and provide a reliable data foundation for high-precision coupled calculation.

[0068] In another technical solution, the processor, according to a preset periodic upload cycle, packages the raw settlement data measured by the four displacement sensors, the raw tilt data measured by the dual-axis tilt sensor, the theoretical tilt change calculated by the processor, the measured tilt change, and the vector difference magnitude into an encrypted data packet, and sends it to the cloud server via a wireless network. The cloud server runs a time-series-based deep learning model to predict the basic settlement trend and tilt development trend in the next 24 hours based on the current data characteristics, and sends the prediction results and corresponding three-level decision suggestions back to the processor display screen on site for display. Furthermore, each time the processor outputs an alarm signal, in addition to triggering the audible and visual alarm, it also packages the following data from 30 seconds before the alarm to 10 seconds after the alarm into an encrypted event data packet: raw settlement data measured by four displacement sensors, raw tilt data measured by dual-axis tilt sensors, theoretical tilt change calculated by the processor, measured tilt change, and vector difference magnitude. The processor sends encrypted event data packets to the cloud server via a wireless network; the cloud server runs the time-series-based deep learning model, which uses historical normal data and known accident cases as training sets, predicts the foundation settlement trend and tilt angle development trend in the next 24 hours based on the data characteristics of this alarm event, and provides three-level decision suggestions, including continuing to observe, recommending reduced load and immediate suspension of work and reinforcement. The cloud server transmits the three-level decision recommendations back to the on-site processor display screen.

[0069] In the above technical solution, the processor also packages the raw settlement data measured by the four displacement sensors, the raw tilt data measured by the dual-axis tilt sensor, the theoretical tilt change calculated by the processor, the measured tilt change, and the vector difference magnitude into an encrypted data packet according to a preset periodic upload cycle (e.g., once every 4 hours), and sends it to the cloud server via a wireless network. The cloud server runs the same deep learning model to predict the foundation settlement trend and tilt development trend in the next 24 hours, and sends the prediction results and corresponding three-level decision suggestions back to the processor display screen on site for display, for reference by on-site management personnel during daily inspections.

[0070] Each time the processor outputs an alarm signal, in addition to triggering an audible and visual alarm, it automatically packages all relevant data from 30 seconds before the alarm to 10 seconds after the alarm into an event data packet. This packet includes raw settlement data measured by four displacement sensors, raw tilt data measured by a dual-axis tilt sensor, the theoretical tilt change calculated by the processor, the measured tilt change, and the vector difference magnitude. The data packet is encrypted and sent to a cloud server via a wireless network. Based on the data characteristics of this event, the cloud server's deep learning model predicts the foundation settlement and tilt development trends over the next 24 hours and provides three levels of decision recommendations based on the predicted rate of change: continued observation, reduced load, and immediate work stoppage and reinforcement. The cloud server then transmits these three levels of decision recommendations back to the processor's display screen on-site for display.

[0071] By employing this technical solution, the present invention can completely preserve the original and calculated data before and after an abnormal event, achieving traceability of the abnormal event. Simultaneously, it utilizes a cloud-based deep learning model to predict future risk trends, providing on-site management personnel with scientific emergency response decision-making suggestions and improving the intelligent level of tower crane safety management.

[0072] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and embodiments shown and described herein.

Claims

1. An integrated monitoring system for tower crane foundation settlement and verticality, characterized in that, include: Four displacement sensors are fixed at the four corners of the tower crane foundation. Each displacement sensor is used to measure the settlement of the corresponding corner relative to a reference point. A dual-axis tilt sensor is fixed to the outer side of the standard section at the bottom of the tower crane tower body to measure the actual tilt angle of the tower body along two orthogonal directions; A data acquisition unit is electrically connected to four displacement sensors and a dual-axis tilt sensor to synchronously acquire various settlement amounts and measured tilt angles. A processor is connected in communication with the data acquisition unit. The processor stores the initial settlement of each displacement sensor, the initial measured tilt angle of the dual-axis tilt sensor, the horizontal distance data between the four displacement sensors, a preset settlement threshold, a preset tilt angle threshold, and a preset difference threshold. The horizontal distance data between the four displacement sensors includes the distance between adjacent sensors along two orthogonal directions. The processor calculates the settlement change at each corner based on the settlement values ​​collected from the four displacement sensors and the initial settlement value corresponding to each displacement sensor, and selects the maximum value among the settlement changes to compare with the preset settlement threshold. The processor subtracts the measured tilt angle of the currently acquired dual-axis tilt sensor from the initial measured tilt angle to obtain the measured tilt angle change, which includes components in two orthogonal directions. The processor calculates the magnitude of the measured tilt angle change and compares it with a preset tilt angle threshold. The processor calculates the theoretical change in tilt angle caused by uneven settlement based on the settlement changes at each corner and the horizontal distance data between the four displacement sensors. The theoretical change in tilt angle includes components in two orthogonal directions. The processor uses the theoretical change in tilt angle and the measured change in tilt angle as vectors, calculates the magnitude of the vector difference, uses the magnitude as the difference value, and compares the difference value with a preset difference threshold. When the maximum value of the settlement change exceeds the preset settlement threshold, or the magnitude of the measured tilt angle change exceeds the preset tilt angle threshold, or the difference exceeds the preset difference threshold, the processor outputs an alarm signal.

2. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 1, characterized in that, When the processor calculates the theoretical change in dip angle caused by uneven settlement, it calculates the ratio of the settlement difference to the corresponding horizontal distance along two orthogonal directions to obtain the theoretical dip angle components in the two directions.

3. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 2, characterized in that, The processor also stores the current working height of the tower crane, as well as the tower body elastic tilt angle and load curve obtained in advance through finite element analysis or on-site calibration; The processor acquires the tower crane's working status in real time, including lifting capacity, luffing range, and slewing angle. Based on the current working height and load status, the processor retrieves the additional tilt component caused by the elastic deformation of the tower body from the tower body elastic tilt angle and load curve; Before comparing the theoretical tilt angle change with the measured tilt angle change using vector difference, the processor first subtracts the additional tilt angle component from the measured tilt angle change to obtain the net measured tilt angle change after deducting elastic deformation. The processor calculates the magnitude of the vector difference between the net measured change in tilt angle and the theoretical change in tilt angle, uses this magnitude as the difference value, and compares it with a preset difference threshold.

4. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 3, characterized in that, The processor also stores a coupling correction model, which is used to calculate the amount of correction required to be applied to the elastic tilt angle and load curve of the tower body under different rigid body tilt angles. After the processor finds the additional tilt angle component from the elastic tilt angle and load curve of the tower, it further inputs the theoretical tilt angle change obtained by the current calculation as the rigid tilt angle of the tower into the coupled correction model to obtain the corrected additional tilt angle component. The processor then subtracts the corrected additional tilt component from the measured tilt angle change to obtain the net measured tilt angle change after deducting the effects of elastic deformation and rigid-elastic coupling.

5. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 4, characterized in that, The processor also stores an initial rigid body tilt angle, which is calculated from the initial settlement of the four displacement sensors in the initial state. Before inputting the currently calculated theoretical tilt angle change into the coupled correction model, the processor first adds the initial rigid body tilt angle to the theoretical tilt angle change to obtain the current absolute rigid body tilt angle. The processor then inputs the current absolute rigid body tilt angle into the coupled correction model to obtain the corrected additional tilt angle component.

6. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 5, characterized in that, The processor also stores the initial elastic deformation tilt angle, which is obtained from the tower elastic tilt angle and load curve based on the tower working height, initial load and initial rigid body tilt angle in the initial state, and then corrected by the coupled correction model. The processor subtracts the initial elastic deformation tilt angle from the corrected additional tilt angle component to obtain the elastic deformation change. Then, it subtracts the elastic deformation change from the measured tilt angle change to obtain the net measured tilt angle change after deducting the elastic deformation change.

7. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 1, characterized in that, The processor also stores a preset non-coplanar error threshold; The processor uses the least squares method to fit an optimal fitting plane based on the settlement changes at the four corners and the horizontal distance data between the four displacement sensors. It also calculates the absolute value of the residual settlement difference at the four corners relative to the optimal fitting plane and takes the maximum value of the absolute value of the four residual settlement differences as the non-coplanar error. When the non-coplanar error exceeds the preset non-coplanar error threshold, the processor outputs a basic torsion alarm signal and skips the operation of comparing the vector difference between the theoretical tilt angle change and the measured tilt angle change at the current sampling time. At the same time, it skips the alarm judgment operation of the difference exceeding the threshold corresponding to the vector difference comparison and continues to execute the operation of comparing the maximum value of the settlement change with the preset settlement threshold, and continues to execute the operation of comparing the magnitude of the measured tilt angle change with the preset tilt angle threshold. When the non-coplanar error does not exceed the preset non-coplanar error threshold, the processor extracts the theoretical tilt angle change from the slope of the two directions of the best-fit plane, and uses the theoretical tilt angle change and the measured tilt angle change as the modulus of the vector difference.

8. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 1, characterized in that, A hardware trigger synchronization circuit is also set between the data acquisition unit and the processor; at the beginning of each data acquisition cycle, the hardware trigger synchronization circuit simultaneously sends an edge trigger pulse to the four displacement sensors and the dual-axis tilt sensor. Four displacement sensors and a dual-axis tilt sensor are synchronously sampled and latched on the rising edge of the same pulse. The latched signal initiates analog-to-digital conversion on the rising edge of the same pulse or the falling edge that follows it. The digital values ​​output by the four displacement sensors and the dual-axis tilt sensor after analog-to-digital conversion are all associated with the period identifier of the current trigger pulse; The processor combines four settlement measurements and one tilt measurement collected under the same trigger pulse into a set of synchronous measurement data based on the period identifier, so that the maximum acquisition time difference between the settlement measurements of the four displacement sensors and the tilt measurements of the dual-axis tilt sensor is less than 1 microsecond.

9. The integrated monitoring system for tower crane foundation settlement and verticality as described in claim 1, characterized in that, The processor packages the raw settlement data measured by the four displacement sensors, the raw tilt data measured by the dual-axis tilt sensor, the theoretical tilt change calculated by the processor, the measured tilt change, and the vector difference magnitude into an encrypted data packet according to the preset periodic upload cycle, and sends it to the cloud server via wireless network. The cloud server runs a time-series-based deep learning model to predict the foundation settlement trend and dip angle development trend in the next 24 hours based on the current data characteristics, and sends the prediction results and corresponding Level 3 decision-making suggestions back to the on-site processor display screen for display. Furthermore, each time the processor outputs an alarm signal, in addition to triggering the audible and visual alarm, it also packages the following data from 30 seconds before the alarm to 10 seconds after the alarm into an encrypted event data packet: raw settlement data measured by four displacement sensors, raw tilt data measured by dual-axis tilt sensors, theoretical tilt change calculated by the processor, measured tilt change, and vector difference magnitude. The processor sends encrypted event data packets to the cloud server via a wireless network; The cloud server runs the time-series-based deep learning model, which uses historical normal data and known accident cases as training sets. Based on the data characteristics of this alarm event, it predicts the foundation settlement trend and tilt angle development trend in the next 24 hours and provides three-level decision suggestions, including continuing to observe, suggesting reduced load, and immediately stopping work and reinforcing. The cloud server transmits the three-level decision recommendations back to the on-site processor display screen.

Images