High-formwork multi-parameter real-time automatic monitoring system and application method

Abstract

The invention discloses a high-formwork multi-parameter real-time automatic monitoring system and an application method, and relates to the technical field of monitoring analysis, and the high-formwork multi-parameter real-time automatic monitoring system comprises a data stream receiving unit, a stiffness matrix determination unit, a function construction unit, a model calibration unit, a full field set setting unit, an abnormal region determination unit, a root cause analysis unit and an intervention measure output unit. Through multi-source data fusion and physical intelligence, technical optimization of high-formwork safety monitoring from passive alarm to active prevention and control is achieved, multi-source heterogeneous data such as sensors, construction and environments are aligned through precise space-time synchronization, a unified data stream is formed, meanwhile, sparse actual measurement data are utilized, and the safety of the high-formwork is improved. According to the method, parameters of a finite element digital twinborn model are dynamically calibrated, a high-fidelity model synchronous with a real structural mechanical state is generated, forward simulation is performed on the basis of the model, a complete stress field and a force flow path of an area without a sensor are perspective, and jump from point monitoring to field cognition is realized.

Description

High formwork multi-parameter real-time automated monitoring system and application method Technical Field

[0001] This invention relates to the field of monitoring and analysis technology, specifically to a real-time automated monitoring system and application method for high formwork multi-parameter monitoring. Background Technology

[0002] A high-formwork multi-parameter real-time automated monitoring system constructs a dynamic model through data fusion and physical inversion, enabling causal tracing and risk prediction, upgrading from passive alarm to proactive intelligent prevention and control. The invention patent application number 202511306692.7 discloses "A high-formwork optimization design method and system based on BIM and finite element analysis, the method including: S100, establishing a three-dimensional BIM model of the main structure, creating an intelligent parametric frame family mapped to the main structure family and with self-verification; S200, importing the parametric family and linking it to the main model, generating multiple sets of frame arrangement methods through spatial topology analysis." The process involves several steps: S300, S400, S500, S600, S700, S800, S9 ...

[0003] The aforementioned existing technologies have solved the problem of the inability to achieve coordinated linkage of geometric data, mechanical data, and construction stage data throughout the entire process. However, during use, the system cannot provide early warning of hidden risks in unmeasured areas, leading to sudden accidents with unclear causes. At the same time, the load analysis is detached from the actual construction dynamics, the model is severely distorted, the calculation results cannot guide practice, and the disposal decision lacks quantitative simulation, making it impossible to evaluate the merits of the solution and resulting in a waste of resources. Summary of the Invention

[0004] The purpose of this invention is to provide a real-time automated monitoring system and application method for high formwork multi-parameter monitoring to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a high-formwork multi-parameter real-time automated monitoring system, comprising a root cause analysis unit and an intervention measure output unit; a data stream receiving unit for reading the original asynchronous data stream, analyzing it, and outputting a primary event set; a stiffness matrix determination unit for extracting the connection relationship matrix of the high-formwork system from the construction BIM model, constructing a finite element model based on the relationship matrix, and calculating the parameterized stiffness matrix; a function construction unit for analyzing the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and generating a likelihood function; a model calibration unit for determining the optimal estimate of each control parameter based on the likelihood function, substituting it into the finite element model, and thus outputting the effective coefficients of the model; a full-field set setting unit for obtaining the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment, determining the nodal displacement response vector and the material parameters of the elements in the full model, and then outputting the full-field set; and an abnormal region determination unit for extracting the actual force flow path point set in the full-field set and the ideal force flow path point set in the finite element model, calculating the distance between the two point sets, and marking the abnormal region based on the distance value.

[0006] Preferably, the data stream receiving unit includes a data acquisition module, a data processing module, and a primary event generation module. The data acquisition module reads the original asynchronous data stream, including sensor monitoring data, construction activity data, and environmental data, and performs spatiotemporal encoding on each data point in the data stream using a time protocol and coordinate mapping with the construction BIM model. The data processing module determines the spatiotemporally encoded data point sequence, divides a sliding time window according to a fixed duration, resamples data of different frequencies within the window to a unified time series, and generates a multimodal data tensor. The primary event generation module acquires the aligned multimodal data tensor, analyzes each data channel, identifies statistical mutation points, over-limit alarm points, and construction behavior characteristic points, and outputs a primary event set containing start and end times, spatial locations, event types, and change amplitudes.

[0007] Preferably, the stiffness matrix determination unit includes a model building module, an actual stiffness analysis module, a stiffness coefficient analysis module, an initial displacement analysis module, and a stiffness matrix optimization module. The model building module extracts the connection relationship matrix of the high-support formwork system from the construction BIM model. This matrix records the starting node number, ending node number, and element type of all members. The element types include vertical members, horizontal members, and diagonal braces. An initial stiffness matrix is ​​established based on the member cross-sectional properties and material parameters from the design drawings. This matrix is ​​generated by assembling and superimposing the element stiffness matrices of each member according to their position in the overall structure. A finite element model is generated using the connection relationship matrix, member cross-sectional properties, and material parameters. The actual stiffness analysis module determines the bottom position of all vertical members in the high-support formwork system, using these as support points. Assign corresponding control parameters For the first Support points ,Sure Support stiffness reference value ,according to and Calculate the actual stiffness ,in , Indicates the first One support point, express The control parameters, This represents the total number of support points, and the process is repeated sequentially. Output the corresponding actual stiffness ,in Indicates the first The actual stiffness of each support point; the stiffness coefficient analysis module obtains the actual stiffness of all connection nodes in the high-support formwork system. ,for Assign corresponding control parameters For the first Connecting nodes ,Sure Rotational stiffness reference value and relative angle ,according to and Calculate stiffness coefficient ,in , Indicates the first One connection node, express The control parameters, express The control parameters, This represents the total number of connected nodes, and is traversed sequentially. Output the corresponding stiffness coefficient ,in express The stiffness coefficient; the initial displacement analysis module reads all key members in the high formwork system. ,for Assign corresponding control coefficients For the first Key components ,Sure length and along The position vector of the axis from the top to the bottom ,in , Indicates the first Location coordinates, This indicates the total number of position coordinates. Indicates the first A key component, Indicates the total number of critical members. express The control parameters, express Control parameters, according to and Calculate the initial lateral displacement vector at different coordinate positions. ,in traverse sequentially Output the corresponding initial lateral displacement vector. ,in express The initial lateral displacement vector; the stiffness matrix optimization module calculates each control parameter. Corresponding stiffness contribution matrix According to the initial stiffness matrix and Analyze the parameterized stiffness matrix ,in , , Indicates the first One control parameter, Indicates the first A stiffness contribution matrix, Indicates the serial number. Indicates the first One control parameter, Indicates the total number of control parameters. express The corresponding stiffness contribution matrix.

[0008] Preferably, the function construction unit includes a displacement vector calculation module, a prediction vector calculation module, and a matrix calculation module. The displacement vector calculation module sets a time window, arranges all aligned geometric monitoring data and mechanical monitoring data within the current time window in sequence to form an observation vector, maps the construction activity data within the current time window to obtain a load vector, and calculates the displacement vectors of all degrees of freedom of the finite element model using the parameterized stiffness matrix and load vector. The prediction vector calculation module reads the sensor position mapping matrix, and calculates the prediction vectors of the displacement, tilt, and axial force sensors based on the sensor position mapping matrix and the displacement vectors of all degrees of freedom through selective extraction and internal force conversion. It then concatenates the prediction vectors of all sensors in the same order as the observation vectors to output the prediction vectors of the finite element model. The matrix calculation module determines the measurement accuracy of each sensor, generates an error covariance matrix based on the measurement accuracy, and constructs a likelihood function using the observation vectors, prediction vectors, and error covariance matrix.

[0009] Preferably, the model calibration unit includes a prior probability analysis module, a gradient value calculation module, a sampling execution module, a sample output module, an estimate generation module, and a model validation module; the prior probability analysis module determines each control parameter. After determining the corresponding physical properties and their reasonable value ranges, a corresponding prior probability distribution is set for them. ,according to Calculate the joint prior probability density ,in , Indicates the first One control parameter, Indicates the total number of control parameters. express The prior probability distribution, Indicates the first One control parameter The prior probability distribution, The sequence number is represented by the following: the gradient value calculation module sets up the control parameter space and calculates the gradient value of the logarithmic posterior probability with respect to each control parameter using the adjoint variable method; the sampling execution module establishes a set of auxiliary coefficient variables. ,use and control parameter vector Construct the main function ,in , Represents the observation vector. Represents the identity matrix. Representing the logarithmic posterior probability, the main function is discretized and solved to simulate the evolution trajectory of the control parameters and auxiliary coefficient variables in virtual time, thus obtaining candidate samples. And calculate the change in the principal function value at the beginning and end of the trajectory. ,according to Calculate the acceptance probability of the current candidate sample. ,in , According to the probability of acceptance Determine whether to accept the candidate sample. If accepted, then the control parameter vector in the candidate sample... Used as the control parameter vector for the next sampling, and auxiliary coefficient variables in the candidate samples are discarded. Conversely, As the control parameter vector for the next sampling, discard This completes one sampling iteration; the sample output module repeats the operation and runs multiple independent sampling chains in parallel, removing the first sample from each chain. Sub-samples, where After setting a target value and confirming chain convergence, collect samples from all affected links to form an approximate representation sample set of the posterior distribution. , Represents the total number of samples. Indicates the first The sample set includes a set of samples. The estimation generation module calculates the statistical moments that approximate all samples in the sample set. Based on the statistical moments, it constructs the posterior expectation estimate and covariance matrix of the control parameter vector. The optimal estimate of the control parameters and the quantitative index of the parameter estimate are output through the posterior expectation and covariance matrix. The model verification module substitutes the optimal estimate into the finite element model to obtain the model prediction value. It calculates the deviation value based on the observation vector and the prediction vector, analyzes the measurement standard deviation of each sensor according to the error covariance matrix, and uses the deviation value and measurement standard deviation to deduce the residual of each sensor. The model is then verified, and the effective coefficient of the model is output.

[0010] Preferably, the overall set unit includes a nodal load analysis module, a displacement response module, a force flow plotting module, and a route storage module. The nodal load analysis module acquires the three-dimensional region of the currently poured concrete, its real-time thickness, and the spatial position and operating parameters of the construction equipment. After determining the concrete unit weight and construction load vector, it calculates the distributed pressure and concentrated load acting on the pouring formwork. The distributed pressure is equivalently distributed to the corresponding nodes using the element shape functions in the finite element model, and the equipment impact load is applied to the node closest to the spatial position of the equipment. The nodal forces generated by all static loads are superimposed with the dynamic impact load vector to construct the nodal load vector. The displacement response module... The module reads the stiffness matrix and nodal load vectors of the finite element model, selects a solver based on the dynamic characteristics of the load, constructs and analyzes the corresponding equations, and outputs the displacement response vectors of all nodes in the entire model. The force flow plotting module determines the nodal displacement response vectors and material parameters of the elements in the entire model, calculates the stress and internal force components of all elements, compares the internal forces with the design bearing capacity of the members, calculates the real-time bearing capacity utilization rate of each element, and generates a three-dimensional isosurface based on the internal forces and stress components of all elements. The force flow path is traced and plotted based on the principal stress directions in the three-dimensional isosurface. The route storage module stores the stress, internal forces, real-time bearing capacity utilization rate, and force flow path point set of all elements into the full field set.

[0011] Preferably, the abnormal region determination unit includes a minimum distance analysis module and an abnormal region marking module; the minimum distance analysis module extracts the actual force flow path point set in the entire field. With the ideal force flow path point set in the finite element model ,for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in , express The Middle 1 node express The total number of nodes in the system express The Middle 1 node express The total number of nodes in the system express With all nodes The minimum distance for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in , express With all nodes The minimum distance, express With all nodes The minimum distance; the abnormal region marking module will and The maximum value in the value is used as the distance between the two point sets. If the distance value is greater than the preset deviation threshold, the spatial region corresponding to the continuous sub-path segment that is farthest from the ideal force flow path point set in the actual force flow path point set is marked as an abnormal region.

[0012] Preferably, the root cause analysis unit includes an anomaly type identification module, a joint event merging module, a cause-effect graph generation module, a basic score calculation module, and an event sorting module. The anomaly type identification module receives the primary event set and the anomaly region, extracts key mechanical features based on the unit stress components, internal force components, and force flow paths within the anomaly region, and identifies the dominant anomaly type and force flow deviation index of each anomaly region through these key mechanical features. The joint event merging module analyzes the corresponding severity coefficients based on the maximum load-bearing capacity and force flow deviation index of the units within the anomaly region. Using the dominant anomaly type, spatial coordinates, force flow paths, and severity coefficients of the anomaly region, it generates anomaly events, performs linear fitting on the same type of primary events in continuous time to obtain trend events, and merges the original primary event set, anomaly events, and trend events to form a joint event set. The cause-effect graph generation module statistically analyzes the joint event set. For all related pairs, the weight coefficient of each related edge is calculated using the spatial distance between events within the related pair. A directed weighted causal graph is constructed according to the related pairs and weight coefficients. After the basic score calculation module obtains a real-time event, it traverses the graph in the reverse direction along the related edges within the causal graph, starting from that event, to obtain a set of causal paths. For each causal path, the generation time difference between the candidate root cause event on the path and the current event is determined. The time decay factor is calculated based on the time difference and combined with the weights of the related edges on all paths to obtain the basic score. The event sorting module reads knowledge base entries, which store the typical causal strength between different types of events. It matches the knowledge base entries according to the event types in the path, corrects the basic score, and obtains the actual credibility score of the path. The path to which each candidate root cause event belongs is sorted according to the highest credibility score, and the root cause sorting list of the current event is output.

[0013] Preferably, the intervention output unit includes a model deduction module and a measure recommendation module; the model deduction module modifies the local parameters of the finite element model according to the root cause event, performs local deduction through nonlinear finite element analysis, and predicts and quantifies the force flow path affected by the root cause event; the measure recommendation module reads the treatment measure knowledge base, matches the corresponding adjustment scheme according to the knowledge base, performs rapid simulation of the adjustment scheme using the finite element model, and selects the optimal intervention measure by comparing the mechanical indicators before and after the simulation.

[0014] The high-formwork multi-parameter real-time automated monitoring application method includes the following steps: S1. Read the raw asynchronous data stream, analyze it, and output a primary event set; S2. Extract the connection relationship matrix of the high-formwork system from the construction BIM model, construct a finite element model based on the relationship matrix, and calculate the parameterized stiffness matrix; S3. Analyze the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and generate a likelihood function; S4. Determine the optimal estimate of each control parameter based on the likelihood function, substitute it into the finite element model, and output the effective coefficients of the model; S5. Obtain... S6. After determining the nodal displacement response vector and element material parameters of the entire model by taking the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment, output the entire field set; S7. Extract the actual force flow path point set from the entire field set and the ideal force flow path point set from the finite element model, calculate the distance between the two point sets, and mark the abnormal areas according to the distance values; S8. Generate abnormal events according to the abnormal areas and form a cause-effect graph. After obtaining the real-time events, select the root cause ranking list of the real-time events according to the cause-effect graph; S9. Use the root cause ranking list to determine the adjustment scheme, and select the optimal intervention measures in combination with the simulation results.

[0015] Compared with existing technologies, the beneficial effects of this invention are: 1. This invention achieves technological optimization of high-support formwork safety monitoring from "passive alarm" to "active prevention and control" through multi-source data fusion and physical intelligence. Utilizing precise spatiotemporal synchronization, it aligns heterogeneous data from multiple sources such as sensors, construction, and the environment to form a unified data stream. Simultaneously, it uses sparse measured data to dynamically calibrate the parameters of the finite element digital twin model, generating a high-fidelity model synchronized with the actual structural mechanical state. Based on this model, forward simulation is performed, revealing the complete stress field and force flow path in areas without sensor deployment, achieving a leap from "point monitoring" to "field cognition"; 2. This invention abstracts threshold alarms, construction actions, and mechanical diagnosis results into events, and uses a calibration model to verify the physical transmission path between events, constructing a causal graph. When an alarm triggers and generates a new event, the system can trace back to the physically plausible root cause, and simulate the root cause deterioration and the effect of reinforcement measures in a digital twin, quantifying risk evolution and recommending the optimal solution. It integrates data-driven deep learning, physical-driven finite element analysis, and knowledge-driven approaches to form a high-formwork monitoring system with perception, diagnosis, prediction, and decision-making capabilities, elevating construction safety management to a new level of explainability, predictability, and optimizability. Attached Figure Description

[0016] Figure 1 is a schematic diagram of the overall system flow provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please refer to Figure 1. This invention provides a technical solution: a high-formwork multi-parameter real-time automated monitoring system, including a root cause analysis unit and an intervention measure output unit; a data stream receiving unit, used to read the original asynchronous data stream, analyze it, and output a primary event set; a stiffness matrix determination unit, used to extract the connection relationship matrix of the high-formwork system from the construction BIM model, construct a finite element model based on the relationship matrix, and calculate the parameterized stiffness matrix; a function construction unit, used to analyze the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and generate a likelihood function; a model calibration unit, used to determine the optimal estimate of each control parameter based on the likelihood function, substitute it into the finite element model, and thus output the effective coefficients of the model; a full-field set setting unit, used to obtain the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment, determine the nodal displacement response vector and the material parameters of the elements in the full model, and output the full-field set; and an abnormal region determination unit, used to extract the actual force flow path point set in the full-field set and the ideal force flow path point set in the finite element model, calculate the distance value between the two point sets, and mark the abnormal region based on the distance value.

[0019] The data stream receiving unit includes a data acquisition module, a data processing module, and a primary event generation module. The data acquisition module reads the raw asynchronous data stream, including sensor monitoring data, construction activity data, and environmental data, and uses a time protocol and coordinate mapping with the construction BIM model to perform spatiotemporal encoding on each data point in the data stream. The data processing module determines the spatiotemporally encoded data point sequence, divides the sliding time window according to a fixed duration, resamples data of different frequencies within the window to a unified time series, and generates a multimodal data tensor. The primary event generation module acquires the aligned multimodal data tensor, analyzes each data channel, identifies statistical mutation points, over-limit alarm points, and construction behavior characteristic points, and outputs a primary event set containing start and end times, spatial locations, event types, and change amplitudes. The stiffness matrix determination module includes a model building module, an actual stiffness analysis module, a stiffness coefficient analysis module, an initial displacement analysis module, and a stiffness matrix optimization module. The model building module extracts the connection relationship matrix of the high-support formwork system from the construction BIM model. This matrix records the start node number, end node number, and element type of all members. Element types include vertical members, horizontal members, and diagonal braces. An initial stiffness matrix is ​​established based on the member cross-sectional properties and material parameters from the design drawings. This matrix is ​​created by assembling and superimposing the element stiffness matrices of each member according to their position in the overall structure. A finite element model is generated using the connection relationship matrix, member cross-sectional properties, and material parameters. The actual stiffness analysis module determines the bottom position of all vertical members in the high-support formwork system, using these as support points. Assign corresponding control parameters For the first Support points ,Sure Support stiffness reference value ,according to and Calculate the actual stiffness ,in , Indicates the first One support point, express The control parameters, This represents the total number of support points, and the process is repeated sequentially. Output the corresponding actual stiffness ,in Indicates the first The actual stiffness of each support point; the stiffness coefficient analysis module obtains the actual stiffness of all connection nodes in the high formwork system. ,for Assign corresponding control parameters For the first Connecting nodes ,Sure Rotational stiffness reference value and relative angle ,according to and Calculate stiffness coefficient ,in , Indicates the first One connection node, express The control parameters, express The control parameters, This represents the total number of connected nodes, and is traversed sequentially. Output the corresponding stiffness coefficient ,in express The stiffness coefficient; the initial displacement analysis module reads all key members in the high formwork system. ,for Assign corresponding control coefficients For the first Key components ,Sure length and along The position vector of the axis from the top to the bottom ,in , Indicates the first Location coordinates, This indicates the total number of position coordinates. Indicates the first A key component, Indicates the total number of critical members. express The control parameters, express Control parameters, according to and Calculate the initial lateral displacement vector at different coordinate positions. ,in traverse sequentially Output the corresponding initial lateral displacement vector. ,in express The initial lateral displacement vector; the stiffness matrix optimization module calculates the values ​​for each control parameter. Corresponding stiffness contribution matrix According to the initial stiffness matrix and Analyze the parameterized stiffness matrix ,in , , Indicates the first One control parameter, Indicates the first A stiffness contribution matrix, Indicates the serial number. Indicates the first One control parameter, Indicates the total number of control parameters. express The corresponding stiffness contribution matrix; the function construction unit includes a displacement vector calculation module, a prediction vector calculation module, and a matrix calculation module; the displacement vector calculation module sets a time window, arranges all aligned geometric monitoring data and mechanical monitoring data within the current time window in sequence to form an observation vector, maps the load vector using the construction activity data within the current time window, and calculates the displacement vectors of all degrees of freedom of the finite element model using the parameterized stiffness matrix and load vector; the prediction vector calculation module reads the sensor position mapping matrix, and calculates the prediction vectors of the displacement, tilt, and axial force sensors based on the sensor position mapping matrix and the displacement vectors of all degrees of freedom through selective extraction and internal force conversion, and concatenates the prediction vectors of all sensors in the same order as the observation vectors to output the prediction vectors of the finite element model; the matrix calculation module determines the measurement accuracy of each sensor, generates an error covariance matrix based on the measurement accuracy, and constructs a likelihood function using the observation vectors, prediction vectors, and error covariance matrix; the likelihood function is specifically:

[0020] in, Represents the observation vector. Represents the control parameter vector. Represents the prediction vector. Represents the error covariance matrix. Indicates the total number of sensors. This indicates that the control parameter vector is Under the condition that the observation vector is The model calibration unit includes a prior probability analysis module, a gradient value calculation module, a sampling execution module, a sample output module, an estimate generation module, and a model validation module; the prior probability analysis module determines each control parameter. After determining the corresponding physical properties and their reasonable value ranges, a corresponding prior probability distribution is set for them. ,according to Calculate the joint prior probability density ,in , Indicates the first One control parameter, Indicates the total number of control parameters. express The prior probability distribution, Indicates the first One control parameter The prior probability distribution, The sequence number is represented by the following: The gradient value calculation module sets up the control parameter space and uses the adjoint variable method to calculate the gradient value of the logarithmic posterior probability with respect to each control parameter; The sampling execution module establishes a set of auxiliary coefficient variables. ,use and control parameter vector Construct the main function ,in , Represents the observation vector. Represents the identity matrix. Representing the logarithmic posterior probability, the main function is discretized and solved to simulate the evolution trajectory of the control parameters and auxiliary coefficient variables in virtual time, thus obtaining candidate samples. And calculate the change in the principal function value at the beginning and end of the trajectory. ,according to Calculate the acceptance probability of the current candidate sample. ,in , According to the probability of acceptance Determine whether to accept the candidate sample. If accepted, then the control parameter vector in the candidate sample... Used as the control parameter vector for the next sampling, and auxiliary coefficient variables in the candidate samples are discarded. Conversely, As the control parameter vector for the next sampling, discard This completes one sampling iteration; the sample output module repeats the operation and runs multiple independent sampling chains in parallel, removing the first sample from each chain. Sub-samples, where After setting a target value and confirming chain convergence, collect samples from all affected links to form an approximate representation sample set of the posterior distribution. , Represents the total number of samples. Indicates the first The sample set is divided into several modules. The estimation module calculates the statistical moments that approximate all samples in the sample set. Based on these statistical moments, it constructs the posterior expectation estimate and covariance matrix of the control parameter vector. The optimal estimate of the control parameters and the quantitative index of the parameter estimate are output through the posterior expectation and covariance matrix. The model verification module substitutes the optimal estimate into the finite element model to obtain the model prediction value. It calculates the deviation value based on the observed vector and the predicted vector, analyzes the measurement standard deviation of each sensor according to the error covariance matrix, and uses the deviation value and measurement standard deviation to calculate the residual of each sensor. This residual is then verified, and the effective coefficients of the model are output. The verification process is as follows: if the absolute value of all residuals is less than or equal to the preset safety factor, the effective coefficient is set to 1, indicating that the model prediction matches the observed data. The calibrated model is accepted, and subsequent automated analysis proceeds. Conversely, if residuals exceeding the safety factor occur, the effective coefficient is set to 0, triggering a tiered diagnosis. Specifically, if the residual of a single sensor exceeds the limit, the sensor is flagged as suspicious, and robust estimation is attempted. If the residuals of multiple sensors exceed the limit, the model is deemed to be potentially faulty. Automated decision-making is suspended, a high-level alarm is issued, and the load input and model estimation are reviewed retrospectively. Manual intervention is also recommended. (Full-field centralized setup) The finite element model includes a nodal load analysis module, a displacement response module, a force flow plotting module, and a route storage module. The nodal load analysis module acquires the three-dimensional region of the currently poured concrete, its real-time thickness, and the spatial location and operating parameters of the construction equipment. After determining the concrete unit weight and construction load vector, it calculates the distributed pressure and concentrated load acting on the pouring formwork. Through the element shape functions in the finite element model, the distributed pressure is equivalently distributed to the corresponding nodes, and the equipment impact load is applied to the node closest to the spatial location of the equipment. The nodal forces generated by all static loads are superimposed with the dynamic impact load vector to construct... The module outputs the nodal load vectors; the displacement response module reads the stiffness matrix and nodal load vectors of the finite element model, selects a solver based on the dynamic characteristics of the load, constructs and analyzes the corresponding equations, and outputs the displacement response vectors of all nodes in the entire model; specifically, the selection of a solver based on the dynamic characteristics of the load, and the construction and analysis of the corresponding equations, are as follows: if the load component is not significant, a quasi-static solver is used to analyze the corresponding equilibrium equations; if the load component is significant, a dynamic time history solver is used to determine the mass matrix, damping matrix, and stiffness matrix, and then constructs and analyzes the corresponding motion equations; the equilibrium equations are:

[0021] in, Represents the stiffness matrix. Represents the nodal displacement vector. Represents the nodal load vector; the equation of motion is:

[0022] in, Represents the mass matrix, Represents the damping matrix. Represents the nodal acceleration vector. The module represents the nodal velocity vector; after determining the nodal displacement response vector and element material parameters of the entire model, the force flow plotting module calculates the stress and internal force components of all elements, compares the internal forces with the design bearing capacity of the members, calculates the real-time bearing capacity utilization rate of each element, and generates a three-dimensional isosurface based on the internal forces and stress components of all elements. The force flow path is then traced and plotted based on the principal stress directions in the three-dimensional isosurface. The route storage module stores the stress, internal forces, real-time bearing capacity utilization rate, and force flow path point set of all elements into the full field set. The abnormal region determination module includes a minimum distance analysis module and an abnormal region marking module; the minimum distance analysis module extracts the actual force flow path point set from the full field set. With the ideal force flow path point set in the finite element model ,for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in , express The Middle 1 node express The total number of nodes in the system express The Middle 1 node express The total number of nodes in the system express With all nodes The minimum distance for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in , express With all nodes The minimum distance, express With all nodes The minimum distance; the anomaly region marking module will and The maximum value in the value is used as the distance between two point sets. If the distance value is greater than the preset deviation threshold, the spatial region corresponding to the continuous sub-path segment in the actual force flow path point set that is farthest from the ideal force flow path point set is marked as an abnormal region. The root cause analysis unit includes an anomaly type identification module, a joint event merging module, a cause-effect graph generation module, a basic score calculation module, and an event sorting module. After receiving the primary event set and the abnormal region, the anomaly type identification module extracts key mechanical features based on the unit stress components, internal force components, and force flow paths within the abnormal region. It then identifies the dominant anomaly type and force flow deviation index of each abnormal region through these key mechanical features. The joint event merging module sorts the abnormal regions according to the single... The maximum load-bearing capacity utilization rate and force flow deviation index of the element are analyzed to obtain the corresponding severity coefficient. Using the dominant anomaly type, spatial coordinates, force flow path, and severity coefficient of the abnormal region, abnormal events are generated. Linear fitting is performed on continuous primary events of the same type to obtain trend events. The original set of primary events, abnormal events, and trend events are merged to form a joint event set. The causal graph generation module counts all related pairs within the joint event set, calculates the weight coefficient of each related edge using the spatial distance between events within the related pair, and constructs a directed weighted causal graph according to the related pairs and weight coefficients. Specifically, the counting of all related pairs within the joint event set involves: S01, determining each event in the set. Source element and neighboring elements in the finite element model; S02, arbitrary selection event. Read events and Spatial location coordinates, analysis and If the spatial distance between source units is less than a preset value, then it is determined that there is a related event pair. Conversely, no operation is performed; S03, will and The spatial location is mapped to the corresponding node in the finite element model, and the force flow path is traversed to search for the path that passes through. Check the path of the corresponding node. Does the corresponding node pass through this path? If so, The index number of the corresponding node in the path is less than If so, then it is determined that there is an associated pair. Conversely, no operation is performed.

[0023] S04. In the calibration model, for events Apply virtual perturbation to the source unit and calculate If the response change value of a neighboring unit exceeds a preset value, it is determined that there is a related event pair. Conversely, no operation is performed.

[0024] The response change value is specifically the change in mechanical parameters calculated by finite element post-processing based on the element displacement vectors before and after the application of virtual disturbance.

[0025] S05, Traverse sequentially ,Sure After the corresponding association pairs, repeat the operation until all events are selected, where Indicates the first One event, The total number of events is represented by the base score calculation module. After acquiring real-time events, the module traverses the graph in reverse order along the edges of the causal graph, starting from each event. This yields a set of causal paths. For each path, the module determines the time difference between the candidate root cause events and the current event, calculates the time decay factor based on this difference, and combines it with the weights of all edges to obtain the base score. The event sorting module reads knowledge base entries, which store typical causal strengths between different types of events. It matches the knowledge base entries with the event types in the path, adjusts the base score accordingly, and obtains the actual score of the path. The system calculates the credibility score of each candidate root cause event, ranking them according to the highest credibility score of the path to which they belong, and outputs a root cause ranking list for the current event. The intervention output unit includes a model deduction module and a measure recommendation module. The model deduction module modifies the local parameters of the finite element model based on the root cause event, performs local deduction through nonlinear finite element analysis, and predicts and quantifies the force flow path affected by the root cause event. The measure recommendation module reads the intervention measure knowledge base, matches corresponding adjustment schemes based on the knowledge base, performs rapid simulation of the adjustment schemes using the finite element model, and selects the optimal intervention by comparing the mechanical indices before and after the simulation. Preventive measures; a real-time automated monitoring method for high-formwork multi-parameter applications, including the following steps: S1. Read the raw asynchronous data stream, analyze it, and output a primary event set; S2. Extract the connection relationship matrix of the high-formwork system from the construction BIM model, construct a finite element model based on the relationship matrix, and calculate the parameterized stiffness matrix; S3. Analyze the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and generate a likelihood function; S4. Determine the optimal estimate of each control parameter based on the likelihood function, substitute it into the finite element model, and output the effective coefficients of the model; S5. After obtaining the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment, and determining the nodal displacement response vector and material parameters of the elements in the full model, the full field set is output; S6, the actual force flow path point set in the full field set and the ideal force flow path point set in the finite element model are extracted, the distance between the two point sets is calculated, and the abnormal areas are marked according to the distance values; S7, abnormal events are generated according to the abnormal areas, and a cause-effect graph is formed. After obtaining the real-time events, the root cause ranking list of the real-time events is selected according to the cause-effect graph; S8, the adjustment scheme is determined using the root cause ranking list, and the optimal intervention measures are selected in combination with the simulation results.

[0026] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A high-support formwork multi-parameter real-time automated monitoring system, characterized in that, Includes a root cause analysis unit and an intervention output unit; The data stream receiving unit is used to read the raw asynchronous data stream, analyze it, and output a primary event set; the stiffness matrix determination unit is used to extract the connection relationship matrix of the high formwork system from the construction BIM model, construct a finite element model based on the relationship matrix, and calculate the parameterized stiffness matrix. The function building unit is used to analyze the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and to generate the likelihood function; The model calibration unit is used to determine the optimal estimate of each control parameter based on the likelihood function, substitute it into the finite element model, and thus output the effective coefficient of the model; the full field set setting unit is used to obtain the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment, determine the nodal displacement response vector of the full model and the material parameters of the elements, and then output the full field set. The abnormal region determination unit is used to extract the actual force flow path point set in the whole field concentration and the ideal force flow path point set in the finite element model, calculate the distance value between the two point sets, and mark the abnormal region based on the distance value.

2. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The data stream receiving unit includes a data acquisition module, a data processing module, and a primary event generation module. The data acquisition module reads the raw asynchronous data stream, including sensor monitoring data, construction activity data, and environmental data, and performs spatiotemporal encoding on each data point in the data stream using a time protocol and coordinate mapping with the construction BIM model. The data processing module determines the spatiotemporally encoded data point sequence, divides a sliding time window according to a fixed duration, resamples data of different frequencies within the window to a unified time series, and generates a multimodal data tensor. The primary event generation module acquires the aligned multimodal data tensor, analyzes each data channel, identifies statistical mutation points, over-limit alarm points, and construction behavior characteristic points, and outputs a primary event set containing start and end times, spatial locations, event types, and change amplitudes.

3. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The stiffness matrix determination unit includes a model building module, an actual stiffness analysis module, a stiffness coefficient analysis module, an initial displacement analysis module, and a stiffness matrix optimization module. The model building module extracts the connection relationship matrix of the high-support formwork system from the construction BIM model. This matrix records the start node number, end node number, and element type of all members. An initial stiffness matrix is ​​established based on the member cross-sectional properties and material parameters from the design drawings. A finite element model is generated using the connection relationship matrix, member cross-sectional properties, and material parameters. The actual stiffness analysis module determines the bottom position of all uprights in the high-support formwork system, using these as support points. Assign corresponding control parameters For the first Support points ,Sure Support stiffness reference value ,according to and Calculate the actual stiffness ,in ， Indicates the first One support point, express The control parameters, This represents the total number of support points, and the process is repeated sequentially. Output the corresponding actual stiffness ,in Indicates the first The actual stiffness of each support point; the stiffness coefficient analysis module obtains the actual stiffness of all connection nodes in the high-support formwork system. ,for Assign corresponding control parameters For the first Connecting nodes ,Sure Rotational stiffness reference value and relative angle ,according to and Calculate stiffness coefficient ,in ， Indicates the first One connection node, express The control parameters, express The control parameters, This represents the total number of connected nodes, and is traversed sequentially. Output the corresponding stiffness coefficient ,in express The stiffness coefficient; the initial displacement analysis module reads all key members in the high formwork system. ,for Assign corresponding control coefficients For the first Key components ,Sure length and along The position vector of the axis from the top to the bottom ,in ， Indicates the first Location coordinates, This indicates the total number of position coordinates. Indicates the first A key component, Indicates the total number of critical members. express The control parameters, express Control parameters, according to and Calculate the initial lateral displacement vector at different coordinate positions. ,in traverse sequentially Output the corresponding initial lateral displacement vector ,in express The initial lateral displacement vector; the stiffness matrix optimization module calculates each control parameter. Corresponding stiffness contribution matrix According to the initial stiffness matrix and Analyze the parameterized stiffness matrix ,in ， ， Indicates the first One control parameter, Indicates the first A stiffness contribution matrix, Indicates the serial number. Indicates the first One control parameter, Indicates the total number of control parameters. express The corresponding stiffness contribution matrix.

4. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The function construction unit includes a displacement vector calculation module, a prediction vector calculation module, and a matrix calculation module. The displacement vector calculation module sets a time window, arranges all aligned geometric and mechanical monitoring data within the current time window sequentially to form an observation vector, maps the construction activity data within the current time window to obtain a load vector, and calculates the displacement vectors of all degrees of freedom of the finite element model using a parameterized stiffness matrix and load vectors. The prediction vector calculation module reads the sensor position mapping matrix, and calculates the prediction vectors of the displacement, tilt, and axial force sensors based on the sensor position mapping matrix and the displacement vectors of all degrees of freedom through selective extraction and internal force conversion. It then concatenates the prediction vectors of all sensors in the same order as the observation vectors to output the prediction vectors of the finite element model. The matrix calculation module determines the measurement accuracy of each sensor, generates an error covariance matrix based on the measurement accuracy, and constructs a likelihood function using the observation vectors, prediction vectors, and error covariance matrix.

5. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The model calibration unit includes a prior probability analysis module, a gradient value calculation module, a sampling execution module, a sample output module, an estimate generation module, and a model validation module; the prior probability analysis module determines each control parameter. After determining the corresponding physical properties and their reasonable value ranges, a corresponding prior probability distribution is set for them. ,according to Calculate the joint prior probability density ,in ， Indicates the first One control parameter, Indicates the total number of control parameters. express The prior probability distribution, Indicates the first One control parameter The prior probability distribution, The sequence number is represented by the following: the gradient value calculation module sets up the control parameter space and calculates the gradient value of the logarithmic posterior probability with respect to each control parameter using the adjoint variable method; the sampling execution module establishes a set of auxiliary coefficient variables. ,use and control parameter vector Construct the main function ,in ， Represents the observation vector. Represents the identity matrix. Representing the logarithmic posterior probability, the main function is discretized and solved to simulate the evolution trajectory of the control parameters and auxiliary coefficient variables in virtual time, thus obtaining candidate samples. And calculate the change in the principal function value at the beginning and end of the trajectory. ,according to Calculate the acceptance probability of the current candidate sample. ,in ， According to the probability of acceptance Determine whether to accept the candidate sample. If accepted, then the control parameter vector in the candidate sample... Used as the control parameter vector for the next sampling, and auxiliary coefficient variables in the candidate samples are discarded. Conversely, As the control parameter vector for the next sampling, discard This completes one sampling iteration; the sample output module repeats the operation and runs multiple independent sampling chains in parallel, removing the first sample from each chain. Sub-samples, where After setting a target value and confirming chain convergence, collect samples from all affected links to form an approximate representation sample set of the posterior distribution. ， Represents the total number of samples. Indicates the first The sample set includes a set of samples. The estimation generation module calculates the statistical moments that approximate all samples in the sample set. Based on the statistical moments, it constructs the posterior expectation estimate and covariance matrix of the control parameter vector. The optimal estimate of the control parameters and the quantitative index of the parameter estimate are output through the posterior expectation and covariance matrix. The model verification module substitutes the optimal estimate into the finite element model to obtain the model prediction value. It calculates the deviation value based on the observation vector and the prediction vector, analyzes the measurement standard deviation of each sensor according to the error covariance matrix, and uses the deviation value and measurement standard deviation to deduce the residual of each sensor. The model is then verified, and the effective coefficient of the model is output.

6. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The overall setup unit includes a nodal load analysis module, a displacement response module, a force flow plotting module, and a route storage module. The nodal load analysis module acquires the three-dimensional region of the currently poured concrete, its real-time thickness, and the spatial location and operating parameters of the construction equipment. After determining the concrete unit weight and construction load vector, it calculates the distributed pressure and concentrated load acting on the pouring formwork. Using element shape functions in the finite element model, it equivalently distributes the distributed pressure to the corresponding nodes and applies the equipment impact load to the node closest to the equipment's spatial location. It then superimposes the nodal forces generated by all static loads with the dynamic impact load vector to construct the nodal load vector. The displacement response module reads... The stiffness matrix and nodal load vectors of the finite element model are taken, and the solver is selected according to the dynamic characteristics of the load. The corresponding equations are constructed and analyzed, thereby outputting the displacement response vectors of all nodes in the entire model. After determining the nodal displacement response vectors and material parameters of the elements in the entire model, the force flow plotting module calculates the stress and internal force components of all elements. The internal forces are compared with the design bearing capacity of the members to calculate the real-time bearing capacity utilization rate of each element. A three-dimensional isosurface is generated by the internal forces and stress components of all elements. The force flow path is traced and plotted according to the principal stress direction in the three-dimensional isosurface. The route storage module stores the stress, internal forces, real-time bearing capacity utilization rate and force flow path point set of all elements into the full field set.

7. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The abnormal region determination unit includes a minimum distance analysis module and an abnormal region marking module; the minimum distance analysis module extracts the actual force flow path point set in the entire field. With the ideal force flow path point set in the finite element model ,for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in ， express The Middle 1 node express The total number of nodes in the system express The Middle 1 node express The total number of nodes in the system express With all nodes The minimum distance for Each node Calculate its relationship with All nodes minimum distance Filter out The maximum value in ,in ， express With all nodes The minimum distance, express With all nodes The minimum distance; the abnormal region marking module will and The maximum value in the value is used as the distance between the two point sets. If the distance value is greater than the preset deviation threshold, the spatial region corresponding to the continuous sub-path segment that is farthest from the ideal force flow path point set in the actual force flow path point set is marked as an abnormal region.

8. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The root cause analysis unit includes an anomaly type identification module, a joint event merging module, a cause-effect graph generation module, a basic score calculation module, and an event sorting module. The anomaly type identification module receives the primary event set and anomaly regions, extracts key mechanical features based on the unit stress components, internal force components, and force flow paths within the anomaly regions, and identifies the dominant anomaly type and force flow deviation index of each anomaly region through these key mechanical features. The joint event merging module analyzes the corresponding severity coefficients based on the maximum load-bearing capacity and force flow deviation index of the units within the anomaly regions. Using the dominant anomaly type, spatial coordinates, force flow paths, and severity coefficients of the anomaly regions, it generates anomaly events, performs linear fitting on continuous primary events of the same type to obtain trend events, and merges the original primary event set, anomaly events, and trend events to form a joint event set. The cause-effect graph generation module statistically analyzes the events within the joint event set. For each associated pair, the weight coefficient of each associated edge is calculated using the spatial distance between events within the associated pair. A directed weighted causal graph is constructed based on the associated pairs and weight coefficients. After acquiring a real-time event, the basic score calculation module traverses the graph in the reverse direction along the associated edges within the causal graph, starting from that event. After obtaining a set of causal paths, for each causal path, the generation time difference between the candidate root cause event on the path and the current event is determined. The time decay factor is calculated based on the time difference and combined with the weights of the associated edges on all paths to obtain the basic score. The event sorting module reads knowledge base entries, which store the typical causal strength between different types of events. The basic score is corrected by matching the knowledge base entries according to the event type in the path to obtain the actual credibility score of the path. The paths to which each candidate root cause event belongs are sorted according to the highest credibility score, and the root cause sorting list of the current event is output.

9. The high-support formwork multi-parameter real-time automated monitoring system according to claim 1, characterized in that, The intervention output unit includes a model deduction module and a measure recommendation module. The model deduction module modifies the local parameters of the finite element model based on the root cause event, performs local deduction through nonlinear finite element analysis, and predicts and quantifies the force flow path affected by the root cause event. The measure recommendation module reads the treatment measure knowledge base, matches the corresponding adjustment scheme according to the knowledge base, performs rapid simulation of the adjustment scheme using the finite element model, and selects the optimal intervention measure by comparing the mechanical indicators before and after the simulation.

10. A method for real-time automated monitoring of multiple parameters in high-support formwork, characterized in that, The monitoring application method described herein is applicable to the high-formwork multi-parameter real-time automated monitoring system described in any one of claims 1-9, and includes the following steps: S1. Read the original asynchronous data stream, analyze it, and output a primary event set; S2. Extract the connection relationship matrix of the high-formwork system from the construction BIM model, construct a finite element model based on the relationship matrix, and calculate the parameterized stiffness matrix; S3. Analyze the observation vector, prediction vector, and error covariance matrix using the parameterized stiffness matrix, and generate a likelihood function; S4. Determine the optimal estimate of each control parameter based on the likelihood function, substitute it into the finite element model, and thus output the model. S5. Obtain the real-time thickness of the currently poured concrete and the operating parameters of the construction equipment. After determining the nodal displacement response vector and material parameters of the elements in the full model, output the full field set. S6. Extract the actual force flow path point set in the full field set and the ideal force flow path point set in the finite element model. Calculate the distance between the two point sets and mark the abnormal areas based on the distance. S7. Generate abnormal events according to the abnormal areas and form a cause-effect graph. After obtaining the real-time events, select the root cause ranking list of the real-time events based on the cause-effect graph. S8. Use the root cause ranking list to determine the adjustment scheme and select the optimal intervention measures based on the simulation results.

Images