Intelligent construction platform full life cycle collaborative control system based on digital twin driving

Abstract

The present application relates to the field of intelligent construction and digital twin technology, in particular to a full life cycle collaborative control system of an intelligent construction platform based on digital twin driving, comprising: a data acquisition module, an acquired state vector and a design change flux; a semantic entropy quantification module, used for calculating semantic entropy values; a phase change decision module, used for determining an operation mode; when the semantic entropy value is greater than or equal to a preset high threshold value, the operation mode is a dimension reduction robust mode; when the semantic entropy value is less than or equal to a preset low threshold value, the operation mode is an efficient parallel mode; when the semantic entropy value is between the low threshold value and the high threshold value, the operation mode of the last moment is maintained; a self-healing collaborative control module, used for generating control instructions; activating a logic self-healing program to correct the component state vector; executing a parallel scheduling strategy; the present application avoids system resource consumption and logic jitter caused by high-frequency switching of the control strategy in the critical state, and ensures the robustness of the system in extreme environments.

Description

Technical Field

[0001] This invention relates to the fields of intelligent construction and digital twin technology, specifically to a full lifecycle collaborative control system for an intelligent construction platform driven by digital twins. Background Technology

[0002] The intelligent construction platform full lifecycle collaborative control system based on digital twins aims to solve the problems of data spatiotemporal disorder and model logic failure in high-concurrency construction environments.

[0003] With the application of modern intelligent construction technology and the advancement of high-concurrency construction, the construction environment is becoming increasingly complex. This complexity manifests in two main aspects: design change storms: drastic changes in upstream design intent lead to a surge in the throughput of design changes processed by the system. Increased latency; sensor delay superposition: Data acquisition from physical sensors in the field has time lags, especially in high-concurrency environments, which can lead to significant communication delays. Traditional control systems often struggle to effectively handle the combined effects of these two factors. When design changes are frequent and sensor delays are substantial, the component state vectors derived from the digital twin model... State vectors acquired by physical field sensors Significant discrepancies can occur between these discrepancies, known as data spatiotemporal distortion. This distortion can further lead to the logical failure of the digital twin model, causing the system to fail to accurately reflect the true state of the physical world, thereby triggering construction accidents or inefficiencies. Therefore, how to quantify the complex risks of spatiotemporal distortion into a single indicator and dynamically adjust control strategies accordingly to ensure the robustness and efficiency of the system has become a key issue that urgently needs to be addressed in this field. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a full lifecycle collaborative control system for an intelligent construction platform driven by digital twins. Specifically, the technical solution of this invention includes:

[0005] The data acquisition module is used to obtain component state vectors derived from digital twin models, state vectors collected by physical field sensors, and design change throughput from engineering collaboration platforms and IoT platforms.

[0006] The semantic entropy quantization module is used to calculate the semantic entropy value based on the component state vector, the state vector collected by physical field sensors, and the design change flux. The semantic entropy value is determined by weighted integration of historical state deviations within a preset observation time window and combined with the rate of change of the design change flux.

[0007] The phase transition decision module is used to determine the operating mode based on the comparison result between the semantic entropy value and the preset threshold. Specifically, when the semantic entropy value is greater than or equal to the preset high threshold, the operating mode is determined to be the dimensionality reduction robust mode; when the semantic entropy value is less than or equal to the preset low threshold, the operating mode is determined to be the efficient parallel mode; when the semantic entropy value is between the low threshold and the high threshold, the operating mode of the previous time step is maintained.

[0008] The self-healing collaborative control module is used to generate control commands in response to the operating mode; in response to the dimensionality reduction robust mode, it activates the logical self-healing program to correct the component state vector and executes the serial safety strategy; in response to the efficient parallel mode, it executes the parallel scheduling strategy.

[0009] Preferably, the semantic entropy quantization module calculates the semantic entropy value based on the component state vector, the state vector collected by physical field sensors, and the design change throughput, including:

[0010] Call the preset observation time window, memory decay factor and normalized frequency coefficient;

[0011] Calculate the squared Euclidean distance between the component state vector and the state vector acquired by the physical field sensors;

[0012] Divide the square of the Euclidean distance by the preset theoretical maximum magnitude of the state vector to obtain the normalized state deviation.

[0013] Calculate the time derivative of the design change flux;

[0014] An interference enhancement factor is constructed based on the derivative and a preset change sensitivity coefficient;

[0015] The product of the normalized state bias and the disturbance enhancement factor is weighted and integrated within the observation time window by introducing a memory decay factor.

[0016] Multiply the integral result by the normalized frequency coefficient to obtain the semantic entropy value.

[0017] Preferably, the phase transition decision module determines the operating mode based on the comparison result of the semantic entropy value and the preset threshold, including:

[0018] Call the high threshold, low threshold, and modal state variables from the previous time step;

[0019] If the semantic entropy value is greater than or equal to the high threshold, the modal state variable at the current moment is set as the first value representing the dimensionality reduction robust mode;

[0020] If the semantic entropy value is less than or equal to the low threshold, the modal state variable at the current moment is set to the second value representing the efficient parallel mode;

[0021] If the semantic entropy value is greater than the low threshold and less than the high threshold, the modal state variable at the current moment is set to the modal state variable at the previous moment.

[0022] Preferably, the self-healing collaborative control module generates control commands in response to the operating mode, including:

[0023] Construct a total control cost model that includes a parallel scheduling cost function and a serial security cost function;

[0024] Based on the determined operating mode, the parallel scheduling cost function and the serial security cost function are weighted and summed to obtain the objective function;

[0025] Find the minimum value of the objective function to obtain the optimal control command.

[0026] Preferably, the self-healing collaborative control module activates a logical self-healing procedure to correct the component state vector, including:

[0027] Construct a normalized topological energy equation; wherein the normalized topological energy equation includes a logical potential term characterizing the topological consistency between nodes, and an observation potential term characterizing the deviation between the inferred value and the observed value.

[0028] The modified state vector set is solved iteratively with the objective of minimizing the output value of the normalized topological energy equation.

[0029] The input of the digital twin model is overwritten using the modified state vector set.

[0030] Preferably, the logical potential term characterizing the topological consistency between nodes in the normalized topological energy equation includes:

[0031] Obtain the topological adjacency weights and preset component feature normalized distances in the BIM model;

[0032] Calculate the difference between the Euclidean distance between the state vectors of two interconnected nodes and the preset BIM theoretical distance;

[0033] Divide the square of the difference by the square of the feature normalized distance, and then perform a weighted summation in combination with the topological adjacency weights to obtain the logical potential term.

[0034] Preferably, the observation potential term in the normalized topological energy equation, which characterizes the deviation between the derived and observed values, includes:

[0035] Obtain the measurement variance of each sensor and the preset data fidelity weights;

[0036] Calculate the squared Euclidean distance between the state vector of a node and its corresponding original observation;

[0037] Divide the square of the Euclidean distance by the measurement variance and perform a weighted summation with data fidelity weights to obtain the observed potential term.

[0038] Preferably, the measurement variance is derived from the equipment's factory calibration parameters or the covariance matrix of the Kalman filter; the data fidelity weights are dimensionless coefficients dynamically set based on the on-site communication quality.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. This system solves the problems of data spatiotemporal disorder and model logic failure: By introducing a semantic entropy quantification module, the risk of data spatiotemporal disorder caused by design changes and sensor delays is quantified into a semantic entropy value; based on this entropy value, the phase change decision module can dynamically switch operating modes, thereby avoiding system logic failure caused by data deviation in high-concurrency construction environments, ensuring that the system can accurately reflect the real state of the physical world, and preventing construction accidents.

[0041] 2. This system achieves dynamic switching between safety and efficiency; through the dual-modal hysteresis control mechanism of the phase-change decision module, it switches to a high-efficiency parallel mode when the risk is low, minimizing the construction period and maximizing construction efficiency with a parallel scheduling strategy; when the risk is high, it switches to a dimension-reduction robust mode, prioritizing the execution of a serial safety strategy to minimize the risk of collisions; this mechanism achieves dynamic mathematical-level switching between safety and efficiency, ensuring the robustness of the system in extreme environments;

[0042] 3. This system possesses logical self-healing capabilities for data repair. Upon entering the dimensionality reduction robust mode, the self-healing collaborative control module activates the logical self-healing program. This program iteratively solves and corrects erroneous component state vectors by constructing and minimizing a normalized topological energy equation containing logical and observational potential terms. This mechanism can still deduce component states consistent with physical logic even in the event of sensor failure or data loss, achieving true logical self-healing.

[0043] 4. This system improves the stability and accuracy of the control strategy: The calculation of semantic entropy not only considers instantaneous physical measurement errors, but also reflects the accumulation of historical errors by introducing a memory decay factor for weighted integration. In addition, it constructs an interference enhancement factor by combining the change rate of the design change flux, which sensitively captures potential logical breaks caused by design changes. At the same time, the hysteresis control mechanism adopted by the phase change decision module simulates the inertia principle in physical phase transitions, avoiding system resource consumption and logic jitter caused by high-frequency switching of control strategies in critical states. Attached Figure Description

[0044] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0045] Figure 1This is a structural diagram of the system of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0047] Example 1:

[0048] Please see Figure 1 A digital twin-driven intelligent construction platform-based full lifecycle collaborative control system includes:

[0049] The data acquisition module is used to obtain component state vectors derived from digital twin models, state vectors collected by physical field sensors, and design change throughput from engineering collaboration platforms and IoT platforms.

[0050] The semantic entropy quantization module is used to calculate the semantic entropy value based on the component state vector, the state vector collected by physical field sensors, and the design change flux. The semantic entropy value is determined by weighted integration of historical state deviations within a preset observation time window and combined with the rate of change of the design change flux.

[0051] The phase transition decision module is used to determine the operating mode based on the comparison result between the semantic entropy value and the preset threshold. Specifically, when the semantic entropy value is greater than or equal to the preset high threshold, the operating mode is determined to be the dimensionality reduction robust mode; when the semantic entropy value is less than or equal to the preset low threshold, the operating mode is determined to be the efficient parallel mode; when the semantic entropy value is between the low threshold and the high threshold, the operating mode of the previous time step is maintained.

[0052] The self-healing collaborative control module is used to generate control commands in response to the operating mode; in response to the dimensionality reduction robust mode, it activates the logical self-healing program to correct the component state vector and executes the serial safety strategy; in response to the efficient parallel mode, it executes the parallel scheduling strategy.

[0053] The semantic entropy quantization module calculates semantic entropy values ​​based on component state vectors, state vectors collected by physical field sensors, and design change throughput, including:

[0054] Call the preset observation time window, memory decay factor and normalized frequency coefficient;

[0055] Calculate the squared Euclidean distance between the component state vector and the state vector acquired by the physical field sensors;

[0056] Divide the square of the Euclidean distance by the preset theoretical maximum magnitude of the state vector to obtain the normalized state deviation.

[0057] Calculate the time derivative of the design change flux;

[0058] An interference enhancement factor is constructed based on the derivative and a preset change sensitivity coefficient;

[0059] The product of the normalized state bias and the disturbance enhancement factor is weighted and integrated within the observation time window by introducing a memory decay factor.

[0060] Multiply the integral result by the normalized frequency coefficient to obtain the semantic entropy value;

[0061] The phase transition decision module determines the operating mode based on the comparison between the semantic entropy value and a preset threshold, including:

[0062] Call the high threshold, low threshold, and modal state variables from the previous time step;

[0063] If the semantic entropy value is greater than or equal to the high threshold, the modal state variable at the current moment is set as the first value representing the dimensionality reduction robust mode;

[0064] If the semantic entropy value is less than or equal to the low threshold, the modal state variable at the current moment is set to the second value representing the efficient parallel mode;

[0065] If the semantic entropy value is greater than the low threshold and less than the high threshold, the modal state variable at the current moment is set to the modal state variable at the previous moment.

[0066] The self-healing collaborative control module responds to the operating mode and generates control commands, including:

[0067] Construct a total control cost model that includes a parallel scheduling cost function and a serial security cost function;

[0068] Based on the determined operating mode, the parallel scheduling cost function and the serial security cost function are weighted and summed to obtain the objective function;

[0069] Find the minimum value of the objective function to obtain the optimal control command.

[0070] This embodiment uses a convex optimization algorithm based on gradient projection to solve for the optimal control command. The specific steps include: calculating the objective function. Regarding control commands gradient Iterative updates using gradient projection : ,in For learning rate, To control the feasible region constraints of instructions, such as resource upper limit constraints; when the gradient magnitude Less than the preset threshold Alternatively, when the maximum number of iterations is reached, stop iteration and output the optimal control command. ;

[0071] This embodiment provides a full lifecycle collaborative control system for an intelligent construction platform driven by digital twins. The system aims to solve the problems of data spatiotemporal disorder and model logic failure caused by the superposition of design change storms and sensor delays in high-concurrency construction environments.

[0072] The system mainly includes a data acquisition module, a semantic entropy quantization module, a phase transition decision module, and a self-healing collaborative control module;

[0073] The main function of the data acquisition module is to build the foundation for the mapping between the physical world and the digital world. In this embodiment, the module acquires three types of core data in real time from the engineering collaboration platform CDE and the Internet of Things platform IoT:

[0074] Component state vectors derived from digital twin models and the state vectors collected by physical field sensors Both are defined as composite heterogeneous vectors, containing geometric and physical components; that is... ;in, The spatial coordinate vector of the component, in meters. The stress or strain vector at the critical point of the component, in MPa or dimensionless; in this embodiment, the component state vector. for dimensional vector; where, Includes the three-dimensional coordinates of the component's centroid and four vertices, with dimensions of... ; Maximum stress including three key points and maximum strain , dimension ;therefore Total dimensions for Component state vectors derived from digital twin models The theoretical state of the component, such as theoretical coordinates and stress values, is output by the DEVS simulation engine.

[0075] State vectors acquired by physical field sensors The physical state transmitted back in real time by the field sensor network;

[0076] Design change throughput The change command count, derived from the engineering collaboration platform, represents the degree of drastic change in the upstream design intent.

[0077] The semantic entropy quantization module is the core sensing unit of this system. Its purpose is to quantify complex spatiotemporal disorder risks into a single dimensionless scalar. In this embodiment, the module calculates the semantic entropy value based on the component state vector, the state vector collected by physical field sensors, and the design change flux. ;

[0078] To accurately capture the coupling effect of historical error accumulation and change storms, this embodiment uses the following semantic entropy evolution equation for calculation:

[0079]

[0080] in, : The semantic entropy value at time t, with a dimension of 1, serves as a measure of the system's logical disorder.

[0081] Unit of normalized frequency coefficient Its value is the reciprocal of the preset observation time window. This is used to ensure that the entropy value is dimensionless;

[0082] Preset observation time window unit It is set to 1.5 times the longest communication delay period on site, and is used to define the backtracking range of historical errors;

[0083] Units of time lag variables , as the integration variable;

[0084] Memory decay factor unit Based on historical data calibration, it is usually taken as 0.5 times the communication jitter frequency to reflect that the closer to the current time, the greater the error weight; Based on history Group communication delay data Perform calibration; calculate communication delay. Standard deviation As a measure of jitter; the communication jitter frequency is defined as ;final The value is ,in The calibration constant is usually taken as... ;

[0085] The maximum modulus of the state vector theory originates from the physical boundary constraints defined in the BIM model and is used to normalize spatial deviations; due to It is a heterogeneous vector, and its magnitude is... Defined as weighted composite modulus To perform normalization, the maximum magnitude of the state vector theory is... Set as ;in, and These are the preset dimensionless weighting coefficients. These are the preset BIM geometric boundary constants. These are the preset material physical limit constants;

[0086] : Geometric boundary constant, taken as the maximum diagonal length of the BIM model bounding box, used to normalize geometric deviations;

[0087] Physical limit constant, taken as the yield strength value of the component material, used to normalize stress deviation;

[0088] : Weighting coefficients used to balance dimensions and coordinate the contributions of geometric and physical errors to the system's entropy, and satisfying ;

[0089] The design change flux continuous function after moving average filtering ensures its derivative. It is stable in numerical calculations; this embodiment uses a window size of [missing value]. The exponentially weighted moving average (EWMA) filter is used for processing; the EWMA formula is... ,in For the original change count, the attenuation factor ,and Set to twice the average communication delay period;

[0090] Change the unit of sensitivity coefficient The frequency of design institute change announcements is set according to the average frequency of such announcements. Average time interval between the issuance of historical design change orders Proportional; the specific formula is as follows ,in For dimensionless calibration coefficients, such as , By statistical analysis of the past The average time interval between design change releases within a month is used to obtain the result.

[0091] The time derivative of the design change flux characterizes the instantaneous intensity of the change storm.

[0092] This module calculates and Square the Euclidean distance between them and divide by The normalized state deviation is obtained; simultaneously, the derivative of the design change flux is calculated and combined with... Constructing the interference enhancement factor is the formula in Term; the product of these two in the window Internal attenuation term Perform a weighted integral and finally multiply by Obtain the semantic entropy value;

[0093] This technique, by introducing a memory forgetting curve and an interference enhancement factor, not only considers instantaneous physical measurement errors but also keenly captures potential logical breaks caused by design changes. Even if the physical error is temporarily small, if the design change is drastic, the entropy value will increase significantly, thus providing an early warning of the risk of system collapse.

[0094] The phase transition decision module is used to resolve the conflict between computational efficiency and system robustness; in this embodiment, the module is based on semantic entropy values. The operating mode of the system is determined by comparing the result with a preset threshold.

[0095] To prevent high-frequency oscillations in the system under critical conditions, this embodiment introduces a dual-mode hysteresis control mechanism, the logic of which is defined as follows:

[0096]

[0097] in, Modal state variables, 1 represents a dimension-reduced robust mode, and 0 represents an efficient parallel mode;

[0098] High thresholds, such as 0.8, are set to 90% of the critical entropy value for system collapse.

[0099] A low threshold, such as 0.5, is set as the upper limit of the system's normal operating entropy value;

[0100] When semantic entropy value Soaring Breakthrough At this time, the system is forced to enter dimensionality reduction robust mode. Only when the entropy value falls back to Only when the following conditions are met will the efficient parallel mode be restored. When in between, maintain the modal state from the previous moment. constant;

[0101] This mechanism simulates the inertia principle in physical phase transitions, maintaining the stability of the strategy in the gray area of ​​data quality fluctuations, and avoiding system resource consumption and logic jitter caused by frequent switching of control strategies.

[0102] The self-healing collaborative control module is used to execute specific control strategies. In this embodiment, the module responds to the operating mode and generates optimal control commands. The module makes decisions based on the following total control cost model:

[0103]

[0104] in: The parallel scheduling cost function is defined as the estimated total project duration; its expression is: ;in, This refers to the set of terminal nodes in the project network diagram. In control commands Schedule task The earliest completion time, which is derived through the forward computation logic of the Critical Path Method (CPM); control instructions. Defined as a resource allocation vector ,in Representative assigned to the first The quantity or priority weight of key resources; tasks Duration Its resource allocation They are inversely proportional and can be modeled as ,in As the baseline construction period, It is a time adjustment function based on resource allocation, for example... , For resources For the task The weights; That is, the calculation is based on this dynamic duration;

[0105] The serial safety cost function is defined as a collision risk penalty term based on the potential field method; its expression is: ;in, In control commands Components under action The predicted location, For the first The location of a static obstacle or adjacent component The preset safe distance threshold, To prevent tiny positive numbers with a denominator of zero;

[0106] Responding to dimensionality reduction robust mode The system determines that the current data has low credibility, prioritizes activating the logical self-healing procedure (see the correction status below for details), and executes the serial security policy minimization. The primary goal is to reduce the risk of accidents.

[0107] Responding to efficient parallel mode The system determines that the current environment is safe and executes a parallel scheduling strategy to minimize [the impact of the environment]. The primary goal is to shorten the construction period;

[0108] By using a weighted summation objective function, the system can achieve a dynamic, mathematical-level switch between safety and efficiency, ensuring that the system does not crash under extreme conditions and maximizes construction efficiency under normal conditions.

[0109] Example 2:

[0110] The self-healing collaborative control module activates a logical self-healing procedure to correct the component state vector, including:

[0111] Construct a normalized topological energy equation; wherein the normalized topological energy equation includes a logical potential term characterizing the topological consistency between nodes, and an observation potential term characterizing the deviation between the inferred value and the observed value.

[0112] The modified state vector set is solved iteratively with the objective of minimizing the output value of the normalized topological energy equation.

[0113] The input of the digital twin model is overwritten using the modified state vector set;

[0114] Construct the logical potential term representing the topological consistency between nodes in the normalized topological energy equation, including:

[0115] Obtain the topological adjacency weights and preset component feature normalized distances in the BIM model;

[0116] Calculate the difference between the Euclidean distance between the state vectors of two interconnected nodes and the preset BIM theoretical distance;

[0117] Divide the square of the difference by the square of the feature normalized distance, and then perform a weighted summation in combination with the topological adjacency weights to obtain the logical potential term.

[0118] The observation potential term, which characterizes the deviation between the derived and observed values, is constructed in the normalized topological energy equation, including:

[0119] Obtain the measurement variance of each sensor and the preset data fidelity weights;

[0120] Calculate the squared Euclidean distance between the state vector of a node and its corresponding original observation;

[0121] Divide the square of the Euclidean distance by the measurement variance and perform a weighted summation with data fidelity weights to obtain the observation potential term;

[0122] Measurement variance is derived from the equipment's factory calibration parameters or the covariance matrix of the Kalman filter; data fidelity weights are dimensionless coefficients dynamically set based on the on-site communication quality.

[0123] This embodiment details the implementation of the logical self-healing program in the aforementioned self-healing collaborative control module. This program only operates when the system is in dimensionality reduction robust mode. It is activated at that time;

[0124] Construction of the normalized topological energy equation

[0125] To reconstruct the true state of components even with corrupted data, this embodiment introduces a normalized topological energy equation. The equation transforms logic repair into an energy minimization problem; during the activation of the logic self-healing procedure ( Logical potential term Rigid body constraints must be enforced. Therefore, the normalized topological energy equation... The final form adopted is the topological deformation energy based on theoretical distance deviation:

[0126]

[0127] Logical potential term is used to characterize the topological consistency between nodes, that is, whether the relative positions between components meet physical constraints;

[0128] Topological adjacency weights are derived from hard logical dependencies in the BIM model, such as beam-column support relationships, which are 1 if they are not, and 0 otherwise.

[0129] : Feature normalized distance, which is the preset average physical size of the components, used to eliminate the influence of physical units;

[0130] : Characterizes topological deformation energy; where Nodes extracted from the BIM model With nodes The theoretical distance between them; the physical meaning of this term is the actual distance between the two nodes after penalty correction. Theoretical distance The deviation ensures that the system can deduce a state that conforms to rigid body geometric constraints when reliable observation data is unavailable; therefore, in The logical potential term should be replaced with ;

[0131] The observation potential term is used to characterize the deviation between the inferred value and the observed value, that is, the calculated state should be as close as possible to the reliable sensor data;

[0132] Raw observations, from IoT data streams;

[0133] Measurement variance originates from the factory calibration parameters of the equipment or the covariance matrix of the Kalman filter; the larger the variance, the less reliable the sensor is, and its contribution to energy penalty is weakened by the divisor effect.

[0134] Data fidelity weight is a dimensionless coefficient dynamically set based on the quality of on-site communication. The dynamic setting formula is ,in To maximize fidelity, This represents the packet loss rate for the current time period. This is the preset packet loss rate threshold. Function ensures The range of values ​​is within between;

[0135] Self-healing execution process: Based on the current BIM topology and sensor data, instantiate the above... To minimize The output value is the objective; using gradient descent or other optimization methods, a set of corrected state vectors is iteratively solved. ; using the modified state vector set By directly overwriting the input of the digital twin model, logical deviations caused by data corruption can be forcibly corrected.

[0136] This self-healing mechanism cleverly utilizes the complementarity of the first term of physical constraints and the second term of data statistical properties; when the sensor fails... When the network is large, the system will automatically rely on topology logic to perform deductions; when the communication quality is good... In large-scale systems, the system relies more on measured data; this mechanism ensures that even in the event of partial sensor failure or data loss, the system can still deduce the physical logic of the component states, achieving true logical self-healing.

[0137] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A full lifecycle collaborative control system for an intelligent construction platform driven by digital twins, characterized in that: include: The data acquisition module is used to obtain component state vectors derived from digital twin models, state vectors collected by physical field sensors, and design change throughput from engineering collaboration platforms and IoT platforms. The semantic entropy quantization module is used to calculate the semantic entropy value based on the component state vector, the state vector collected by physical field sensors, and the design change flux. The semantic entropy value is determined by weighted integration of historical state deviations within a preset observation time window and combined with the rate of change of the design change flux. The phase transition decision module is used to determine the operating mode based on the comparison result between the semantic entropy value and the preset threshold. Specifically, when the semantic entropy value is greater than or equal to the preset high threshold, the operating mode is determined to be the dimensionality reduction robust mode; when the semantic entropy value is less than or equal to the preset low threshold, the operating mode is determined to be the efficient parallel mode; when the semantic entropy value is between the low threshold and the high threshold, the operating mode of the previous time step is maintained. The self-healing collaborative control module is used to generate control commands in response to the operating mode; in response to the dimensionality reduction robust mode, it activates the logical self-healing program to correct the component state vector and executes the serial safety strategy; in response to the efficient parallel mode, it executes the parallel scheduling strategy.

2. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive as described in claim 1, characterized in that, The semantic entropy quantization module calculates semantic entropy values ​​based on component state vectors, state vectors collected by physical field sensors, and design change throughput, including: Call the preset observation time window, memory decay factor and normalized frequency coefficient; Calculate the squared Euclidean distance between the component state vector and the state vector acquired by the physical field sensors; Divide the square of the Euclidean distance by the preset theoretical maximum magnitude of the state vector to obtain the normalized state deviation. Calculate the time derivative of the design change flux; An interference enhancement factor is constructed based on the derivative and a preset change sensitivity coefficient; The product of the normalized state bias and the disturbance enhancement factor is weighted and integrated within the observation time window by introducing a memory decay factor. Multiply the integral result by the normalized frequency coefficient to obtain the semantic entropy value.

3. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive according to claim 1, characterized in that, The phase transition decision module determines the operating mode based on the comparison between the semantic entropy value and a preset threshold, including: Call the high threshold, low threshold, and modal state variables from the previous time step; If the semantic entropy value is greater than or equal to the high threshold, the modal state variable at the current moment is set as the first value representing the dimensionality reduction robust mode; If the semantic entropy value is less than or equal to the low threshold, the modal state variable at the current moment is set to the second value representing the efficient parallel mode; If the semantic entropy value is greater than the low threshold and less than the high threshold, the modal state variable at the current moment is set to the modal state variable at the previous moment.

4. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive as described in claim 1, characterized in that, The self-healing collaborative control module responds to the operating mode and generates control commands, including: Construct a total control cost model that includes a parallel scheduling cost function and a serial security cost function; Based on the determined operating mode, the parallel scheduling cost function and the serial security cost function are weighted and summed to obtain the objective function; Find the minimum value of the objective function to obtain the optimal control command.

5. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive according to claim 1, characterized in that, The self-healing collaborative control module activates a logical self-healing procedure to correct the component state vector, including: Construct a normalized topological energy equation; wherein the normalized topological energy equation includes a logical potential term characterizing the topological consistency between nodes, and an observation potential term characterizing the deviation between the inferred value and the observed value. The modified state vector set is solved iteratively with the objective of minimizing the output value of the normalized topological energy equation. The input of the digital twin model is overwritten using the modified state vector set.

6. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive according to claim 5, characterized in that, Construct the logical potential term representing the topological consistency between nodes in the normalized topological energy equation, including: Obtain the topological adjacency weights and preset component feature normalized distances in the BIM model; Calculate the difference between the Euclidean distance between the state vectors of two interconnected nodes and the preset BIM theoretical distance; Divide the square of the difference by the square of the feature normalized distance, and then perform a weighted summation in combination with the topological adjacency weights to obtain the logical potential term.

7. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive according to claim 5, characterized in that, The observation potential term, which characterizes the deviation between the derived and observed values, is constructed in the normalized topological energy equation, including: Obtain the measurement variance of each sensor and the preset data fidelity weights; Calculate the squared Euclidean distance between the state vector of a node and its corresponding original observation; Divide the square of the Euclidean distance by the measurement variance and perform a weighted summation with data fidelity weights to obtain the observed potential term.

8. The full lifecycle collaborative control system for an intelligent construction platform based on digital twin drive according to claim 7, characterized in that, Measurement variance is derived from the equipment's factory calibration parameters or the covariance matrix of the Kalman filter; data fidelity weights are dimensionless coefficients dynamically set based on the on-site communication quality.

Images