Intelligent construction data analysis system based on digitization

Abstract

The invention relates to the technical field of intelligent construction, in particular to an intelligent construction data analysis system based on digitization. The data acquisition unit is used for acquiring various data of a construction site; by monitoring the working state of each construction device and the construction progress of each stage in real time, the latest data among the modules are obtained, the data of the modules are shared in one platform, management is carried out based on a digital driving mode, the data of the modules are not isolated any more, and the construction efficiency is improved. Instead, mutual interaction can be carried out through a digital system, the risk of data dependence is reduced, meanwhile, the complexity of data synchronization is reduced, the consistency and integrity of interactive data are ensured, and when data differences occur after analysis, the data are synchronized to related modules immediately, so that data delay is reduced, and the data synchronization efficiency is improved. The situation that maintenance is needed due to data asynchronization is avoided, and the accuracy and efficiency of a construction plan can be improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent construction technology, and in particular to an intelligent construction data analysis system based on digitization. Background Art

[0002] Smart construction refers to the application of information technology, automation technology and other advanced technologies in the construction process to improve project efficiency, reduce costs, reduce environmental impact, and improve the functionality and comfort of buildings.

[0003] The patent application number is CN202410700799.9, which states in the specification that "The present disclosure provides an intelligent construction integrated management platform, which belongs to the technical field of management systems. The intelligent construction integrated management platform includes: a construction progress management module, which is used to obtain the construction stage of the target construction area, determine multiple first time markers and multiple second time markers based on the construction stage, and the first construction equipment corresponding to each first time marker and the quality inspection items corresponding to each second time marker; a construction equipment management module, which is used to send equipment operation and maintenance information to the equipment management personnel based on the first time marker, and the equipment operation and maintenance information is used to instruct the equipment management personnel to operate the first construction equipment. Maintenance; construction quality management module, used to send quality inspection information to quality management personnel based on the second time mark, and the quality inspection information is used to instruct quality management personnel to inspect the construction quality according to the quality inspection items. The present disclosure can improve the synergy effect of the existing intelligent construction comprehensive management platform. Although the above technology can achieve efficient collaboration between modules such as construction progress management module, construction equipment management module and construction quality management module, which is conducive to speeding up the construction progress, when information is exchanged between multiple modules, one of the modules has information synchronization or sends out wrong information, which will affect the operation between multiple modules, resulting in the need for maintenance and affecting the accuracy and efficiency of the construction plan.

[0004] To sum up, the research and development of digital-based intelligent construction data analysis systems remains a key issue that needs to be urgently addressed in the field of intelligent construction technology. Summary of the invention

[0005] In order to solve the above-mentioned problems, the present invention provides a digital-based intelligent construction data analysis system to solve the problem that when information is interacted between multiple modules, one of the modules has information synchronization or sends erroneous information, which will affect the operation between multiple modules, resulting in the need for maintenance and affecting the accuracy and efficiency of the construction plan.

[0006] To achieve the above-mentioned object, the present invention provides a digital-based intelligent construction data analysis system, including a data acquisition unit for acquiring various data of a construction site;

[0007] A data concentration unit is used to centrally store the acquired data and synchronize the stored data to the data acquisition unit;

[0008] A prediction unit, used to predict problems that may occur during some routine construction processes at the construction site;

[0009] The repair unit is used to issue early warnings and provide repair suggestions when data confusion or data synchronization failure occurs.

[0010] Furthermore, the data acquisition unit includes:

[0011] Sensor modules are deployed on various construction equipment at the construction site to monitor the working status of various construction equipment at the construction site in real time;

[0012] 4D building information module, used to clearly display the time schedule and actual progress of construction tasks, and obtain the construction progress of each stage;

[0013] The data set contains:

[0014] A data storage module is used to receive data for storage and pre-process the received data to remove duplicate and invalid data;

[0015] The data analysis module is used to perform real-time analysis on the preprocessed data and display the working status of the data acquisition unit.

[0016] Furthermore, the prediction unit includes:

[0017] The historical data collection module is used to obtain the data of the units in the data set to provide a reliable data basis for subsequent predictions;

[0018] The data comparison module is used to obtain real-time data from the data concentration unit and then conduct multi-dimensional comparative analysis between the obtained real-time data and historical data;

[0019] The early warning module is used to receive data after multi-dimensional comparative analysis and further identify the data;

[0020] The repair unit includes:

[0021] The problem processing module is used to receive the warning problem data transmitted by the warning unit, and match the repair plan after analyzing and processing the data;

[0022] The self-learning module receives feedback data after the problem is solved and optimizes future repair plans.

[0023] Furthermore, the data acquisition unit operation process includes:

[0024] The sensor module acquires and updates the data of various construction equipment in real time, and then builds an IoT network through the sensor module and connects to the data concentration unit to form an IoT architecture;

[0025] The 4D building information module obtains real-time and past data of construction progress;

[0026] The unit operation processes in the data set include;

[0027] After receiving the data from the data acquisition unit, the data storage module stores the data, removes duplicate and invalid data, and classifies and marks the pre-processed data;

[0028] The data analysis module accesses the data of the data storage module and performs real-time analysis to visualize the working status of the data acquisition unit. The analyzed data will be synchronized to the data acquisition unit to eliminate data silos.

[0029] Furthermore, the prediction unit operation process includes:

[0030] The historical data collection module accesses the historical data of the data center unit, extracts and summarizes past construction records, equipment status and key data of resource use;

[0031] The data comparison module compares and analyzes real-time data with historical data, and then makes a comprehensive dynamic forecast;

[0032] The early warning module identifies common risks or abnormal conditions based on historical data and combines them with dynamic prediction data, and issues an early warning after identifying similar characteristics.

[0033] The repair unit operation process includes:

[0034] After receiving the data of the warning problem, the problem handling module further analyzes it in combination with the historical maintenance records, finds out the root cause, and provides a repair plan;

[0035] After the problem is solved, the autonomous learning module collects feedback data from the data acquisition unit, verifies the repair results, and updates the system model.

[0036] Furthermore, the prediction unit includes:

[0037] The process of comprehensive dynamic prediction by the data comparison module: C t =S real_t -L hist_t , used to compare the difference between real-time data and historical data under the same conditions, where C t is the difference in data at time t, S real_t is the real-time data of time, L hist_t is the historical data of time; A t=f(C t )×α, used to adjust the weight according to the size and importance of the contrast difference, where A t is the difference weight adjusted value, f(C t ) is the absolute value or square used to calculate the size of the difference, α is the adjustment factor used to give different weights to the differences of different types of data;

[0038] S pred_t+k =φ1S real_t +φ2S real_t-1 +···+θ1C t +θ2C t-1 +···, which means predicting future trends based on the difference between real-time data and historical data, where S pred_t+k is the data value predicted at the future time point t+k, φ i is the autoregressive coefficient of the historical data in the model, θ j is the moving average coefficient of the difference data; S final_pred_t+k =S perd_t+k +λ×(S real_t+k -S perd_t+k ), dynamically adjust the prediction results according to the new real-time data, where S final_pred_t+k is the final forecast value after adjustment, λ is the dynamic adjustment coefficient, which indicates the adjustment strength of the forecast deviation.

[0039] (S real_t+k -S perd_t+k ) is the error between actual data and predicted data; the entire comprehensive dynamic prediction process is based on the comparison of real-time data with historical data, which requires a combination of difference analysis, weight adjustment, dynamic prediction model and real-time feedback.

[0040] Furthermore, the prediction unit includes:

[0041] The process of issuing an early warning after the early warning module identifies similar features: First, the early warning module extracts relevant features from both real-time data and historical data. The feature vector is represented as: Among them G real_t is the real-time data feature vector at time t, G hist_i is the feature vector at a specific time point i in the historical data, and is the value of the nth feature at time t and i. The similarity between the current feature vector and the historical feature vector is measured by Euclidean distance and cosine similarity. Euclidean distance: where f(G real_t , G hist_i ) Euclidean distance and cosine similarity between current data and historical data: Where yxd(G real_t , G hist_i) is the cosine similarity between current data and historical data; Early warning judgment:

[0042] in is the preset threshold of Euclidean distance, τ is the preset threshold of cosine similarity, and the distance between real-time data and its corresponding historical risk data is less than Or if the familiarity is greater than τ, the early warning module determines that the real-time data is similar to the historical risk data and triggers an early warning; the expression formula for issuing early warning information is: Where S t Is the flag of whether the warning is triggered at time t, 1 means triggering the warning, 0 means not triggering. t =1, the early warning module sends an early warning message to the manager.

[0043] Further, the repair unit comprises:

[0044] The problem handling module receives the warning information and analyzes the root cause and provides a repair solution: Use multiple regression analysis to find the most relevant features. The following is the expression formula: u = δ1z1 + δ2z2 + ··· + δ n z n +ω, where u is the severity of the anomaly, z n are the characteristic values, such as equipment temperature, schedule lag time, etc., δ n is the weight of each feature on the abnormal phenomenon, ω is the error term, which represents the unexplained part of the model, and is expressed by δ n The value is used to find out the characteristics that have the greatest impact on the current risk phenomenon, that is, the root cause. The problem is classified according to the root cause and matched with a specific repair plan. The following is the classification expression formula: V = f (F anomaly ), where V is the problem classification result, indicating which type of problem it belongs to, and f is the classification function, which predicts the problem type based on the characteristics of risk data.

[0045] Further, the repair unit comprises:

[0046] The problem handling module selects a repair solution from the repair library or optimization system according to the problem classification result. The following is the expression formula: A = solution (C), where A is the repair solution, indicating the action to be taken. Among them, FA opt is the optimized repair solution, Ω i is the weight of each loss, M(A, G anomaly ) is the loss function of the repair solution under specific abnormal characteristics, which represents the cost of the repair solution; the expression formula for tracking and feedback of the execution results after executing the repair solution is: where ψ newis the updated regression coefficient, which is used for the next prediction, μ is the learning rate, which is used to adjust the coefficient,

[0047] It is the gradient of the loss function with respect to the repair solution, indicating the feedback of the repair effect, and dynamically adjusting the model parameters according to the actual execution effect.

[0048] Further, the repair unit comprises:

[0049] The process of updating the system model by the self-learning module: the real-time data, abnormal data, feedback data, and the execution result data of the repair plan are collected centrally, and the collected data is pre-processed to set the model parameters. Then the expression formula is: in is the current model parameter The predicted value obtained, M is the loss function, which measures the error between the model prediction value and the true target value. It is the optimal model parameter. Not only can the model be updated when new data appears, but the model structure or hyperparameters can also be adjusted through adaptive mechanisms. The expression formula is adaptively adjusted: where υ new is the updated learning rate, M prev , Mcurrent is the verification loss of the previous round and the current round. When the verification set loss continues to decrease, the system increases the learning rate accordingly to speed up the convergence; otherwise, the learning rate is reduced to prevent the model from diverging.

[0050] Beneficial Effects

[0051] Compared with the known public technology, the technical solution provided by the present invention has the following advantages:

[0052] Beneficial effects:

[0053] The present invention monitors the working status of various construction equipment and the construction progress of each stage in real time, so as to obtain the latest data between various modules, share the data of various modules on a platform, and manage them based on a digital drive, so that the data of each module no longer exists in isolation, but can interact with each other through a digital system, thereby reducing the risk of data dependence and the complexity of data synchronization, ensuring the consistency and integrity of interactive data, and immediately synchronizing the data to related modules when the analyzed data is different, so as to reduce data delay and avoid the need for maintenance due to data asynchrony, which is beneficial to improving the accuracy and efficiency of construction plans, and introduces prediction and repair functions to predict and plan possible routine construction problems in advance, and propose recommended repair plans, and autonomously learn and optimize the repair plans in combination with the feedback of repair results, so as to improve the work efficiency of solving construction problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a system diagram of the digital-based intelligent construction data analysis system of the present invention. DETAILED DESCRIPTION

[0055] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0056] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0057] The present invention is further described in detail with reference to the accompanying drawings:

[0058] Example:

[0059] like Figure 1 As shown, the present invention provides a digital-based intelligent construction data analysis system, including: a data acquisition unit, used to acquire various data of the construction site;

[0060] The data acquisition unit includes:

[0061] Sensor modules are deployed on various construction equipment at the construction site to monitor the working status of various construction equipment at the construction site in real time;

[0062] 4D building information module, used to clearly display the time schedule and actual progress of construction tasks, and obtain the construction progress of each stage

[0063] The data acquisition unit operation process includes:

[0064] The sensor module acquires and updates the data of various construction equipment in real time, and then builds an IoT network through the sensor module and connects to the data concentration unit to form an IoT architecture;

[0065] The 4D building information module obtains real-time and past data of construction progress.

[0066] Specifically, through the data acquisition unit, the Internet of Things architecture formed by the sensor module is utilized, and in conjunction with the 4D building information module, various data and construction progress of each module can be acquired in real time, ensuring data synchronization between different modules, greatly reducing data dependency issues, and providing a more accurate data basis for data analysis.

[0067] A data concentration unit is used to centrally store the acquired data and synchronize the stored data to the data acquisition unit;

[0068] The data set contains:

[0069] A data storage module is used to receive data for storage and pre-process the received data to remove duplicate and invalid data;

[0070] The data analysis module is used to analyze the preprocessed data in real time and display the working status of the data acquisition unit;

[0071] The unit operation processes in the data set include;

[0072] After receiving the data from the data acquisition unit, the data storage module stores the data, removes duplicate and invalid data, and classifies and marks the pre-processed data;

[0073] The data analysis module accesses the data of the data storage module and performs real-time analysis to visualize the working status of the data acquisition unit. The analyzed data will be synchronized to the data acquisition unit to eliminate data silos.

[0074] Specifically, the data concentration unit helps to build a unified platform and share data in real time by uniformly storing the data of each module. It also uses the data analysis module to automatically process and analyze the data of each module, and reversely synchronizes the analyzed data to each module to reduce the probability of errors caused by human intervention. It automatically analyzes data anomalies and performs automatic corrections, further reducing problems caused by data dependence.

[0075] A prediction unit, used to predict problems that may occur during some routine construction processes at the construction site;

[0076] The prediction unit includes:

[0077] The historical data collection module is used to obtain the data of the units in the data set to provide a reliable data basis for subsequent predictions;

[0078] The data comparison module is used to obtain real-time data from the data concentration unit and then conduct multi-dimensional comparative analysis between the obtained real-time data and historical data;

[0079] The early warning module is used to receive data after multi-dimensional comparative analysis and further identify the data;

[0080] The prediction unit operation process includes:

[0081] The historical data collection module accesses the historical data of the data center unit, extracts and summarizes past construction records, equipment status and key data of resource use;

[0082] The data comparison module compares and analyzes real-time data with historical data, and then makes a comprehensive dynamic forecast;

[0083] The early warning module identifies common risks or abnormal conditions based on historical data and combines them with dynamic prediction data, and issues an early warning after identifying similar characteristics.

[0084] Specifically, the prediction unit uses the data comparison module to search for risk data in real time, and compares and analyzes the risk data with the historical data of the historical data collection module to conduct a comprehensive dynamic prediction. After the early warning module identifies the characteristics of the risk data, it combines similar characteristics in common risks or abnormal situations to make data synchronization more flexible. The delayed synchronization mechanism is used to update data without affecting the construction progress, ensuring that data synchronization is coordinated with the construction rhythm.

[0085] The repair unit is used to issue warnings and provide repair suggestions when data is disordered or cannot be synchronized;

[0086] The repair unit includes:

[0087] The problem processing module is used to receive the warning problem data transmitted by the warning unit, and match the repair plan after analyzing and processing the data;

[0088] Self-learning module, which receives feedback data after the problem is solved and optimizes future repair solutions;

[0089] The repair unit operation process includes:

[0090] After receiving the data of the warning problem, the problem handling module further analyzes it in combination with the historical maintenance records, finds out the root cause, and provides a repair plan;

[0091] After the problem is solved, the autonomous learning module collects feedback data from the data acquisition unit, verifies the repair results, and updates the system model.

[0092] Specifically, the repair unit analyzes the warning problem data and historical maintenance records to identify the features that have the greatest impact on the current risk data, so that the system can automatically retry or provide repair plans based on data analysis, reduce the impact of data synchronization failure on the overall system, and track the repair result data and feed it back to the system model to update the repair plan.

[0093] The prediction unit includes:

[0094] The process of comprehensive dynamic prediction by the data comparison module: C t =S real_t -L hist_t , used to compare the difference between real-time data and historical data under the same conditions, where C t is the difference in data at time t, S real_t is the real-time data of time, L hist_t is the historical data of time; A t =f(C t )×α, used to adjust the weight according to the size and importance of the contrast difference, where A t is the difference weight adjusted value, f(C t ) is the absolute value or square used to calculate the size of the difference, α is the adjustment factor used to give different weights to the differences of different types of data;

[0095] S pred_t+k =φ1S real_t +φ2S real_t-1 +···+θ1C t +θ2C t-1 +···, which means predicting future trends based on the difference between real-time data and historical data, where S pred_t+k is the data value predicted at the future time point t+k, φ i is the autoregressive coefficient of the historical data in the model, θ j is the moving average coefficient of the difference data; S final_pred_t+k =S perd_t+k +λ×(S real_t+k -S perd_t+k ), dynamically adjust the prediction results according to the new real-time data, where S final_pred_t+k is the final forecast value after adjustment, λ is the dynamic adjustment coefficient, which indicates the adjustment strength of the forecast deviation.

[0096] (S real_t+k -S perd_t+k) is the error between actual data and predicted data; the entire comprehensive dynamic prediction process is based on the comparison of real-time data with historical data, which requires a combination of difference analysis, weight adjustment, dynamic prediction model and real-time feedback.

[0097] The prediction unit includes:

[0098] The process of issuing an early warning after the early warning module identifies similar features: First, the early warning module extracts relevant features from both real-time data and historical data. The feature vector is represented as: Among them G real_t is the real-time data feature vector at time t, G hist_i is the feature vector at a specific time point i in the historical data, and is the value of the nth feature at time t and i; the similarity between the current feature vector and the historical feature vector is measured by Euclidean distance and cosine similarity. Euclidean distance: where f(G real_t , G hist_i ) Euclidean distance and cosine similarity between current data and historical data: Where yxd(G real_t , G hist_i ) is the cosine similarity between current data and historical data; Early warning judgment:

[0099] in is the preset threshold of Euclidean distance, τ is the preset threshold of cosine similarity, and the distance between real-time data and its corresponding historical risk data is less than Or if the familiarity is greater than τ, the early warning module determines that the real-time data is similar to the historical risk data and triggers an early warning; the expression formula for issuing early warning information is: Where S t Is the flag of whether the warning is triggered at time t, 1 means triggering the warning, 0 means not triggering. t =1, the early warning module sends an early warning message to the manager.

[0100] Based on historical data, real-time data and analysis models, it predicts possible future situations, such as construction progress, equipment failure, material requirements or cost deviations. The fundamental purpose is to help managers discover potential problems in advance, optimize the construction process, and make optimization suggestions. Through external tools such as display screens, it generates forecast reports, conducts visual displays, and assists managers in making dynamic adjustments to avoid construction stagnation and affect the accuracy and efficiency of the construction plan due to the lack of synchronization of data in various modules, ensuring that the construction plan, various equipment and construction material resource allocation are more coordinated and reasonable, thereby improving work efficiency.

[0101] The repair unit includes:

[0102] The problem handling module receives the warning information and analyzes the root cause and provides a repair solution: Use multiple regression analysis to find the most relevant features. The following is the expression formula: u = δ1z1 + δ2z2 + ··· + δ n z n +ω, where u is the severity of the anomaly, z n are the characteristic values, such as equipment temperature, schedule lag time, etc., δ n is the weight of each feature on the abnormal phenomenon, ω is the error term, which represents the unexplained part of the model, and is expressed by δ n The value is used to find out the characteristics that have the greatest impact on the current risk phenomenon, that is, the root cause. The problem is classified according to the root cause and matched with a specific repair plan. The following is the classification expression formula: V = f (F anomaly ), where V is the problem classification result, indicating which type of problem it belongs to, and f is the classification function, which predicts the problem type based on the characteristics of risk data.

[0103] The repair unit includes:

[0104] The problem handling module selects a repair solution from the repair library or optimization system according to the problem classification result. The following is the expression formula: A = solution (C), where A is the repair solution, indicating the action to be taken. Among them, FA opt is the optimized repair solution, Ω i is the weight of each loss, M(A, G anomaly ) is the loss function of the repair solution under specific abnormal characteristics, which represents the cost of the repair solution; the expression formula for tracking and feedback of the execution results after executing the repair solution is: where ψ new is the updated regression coefficient, which is used for the next prediction, μ is the learning rate, which is used to adjust the coefficient,

[0105] It is the gradient of the loss function with respect to the repair solution, indicating the feedback of the repair effect, and dynamically adjusting the model parameters according to the actual execution effect.

[0106] The repair unit includes:

[0107] The process of updating the system model by the self-learning module: the real-time data, abnormal data, feedback data, and the execution result data of the repair plan are collected centrally, and the collected data is pre-processed to set the model parameters. Then the expression formula is: in is the current model parameter The predicted value obtained, M is the loss function, which measures the error between the model prediction value and the true target value. It is the optimal model parameter. Not only can the model be updated when new data appears, but the model structure or hyperparameters can also be adjusted through adaptive mechanisms. The expression formula is adaptively adjusted: where υ new is the updated learning rate, M prev , Mcurrent is the verification loss of the previous round and the current round. When the verification set loss continues to decrease, the system increases the learning rate accordingly to speed up the convergence; otherwise, the learning rate is reduced to prevent the model from diverging. Through the continuous replacement of new and old data and under the guidance of feedback information, it is convenient to continuously optimize and improve the prediction and decision-making capabilities of the system.

[0108] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. The digital-based intelligent construction data analysis system is characterized by: include: A data acquisition unit, used to acquire various data of the construction site; A data concentration unit is used to centrally store the acquired data and synchronize the stored data to the data acquisition unit; A prediction unit, used to predict problems that may occur during some routine construction processes at the construction site; The repair unit is used to issue early warnings and provide repair suggestions when data confusion or data synchronization failure occurs.

2. The digital-based intelligent construction data analysis system according to claim 1 is characterized in that: The data acquisition unit includes: Sensor modules are deployed on various construction equipment at the construction site to monitor the working status of various construction equipment at the construction site in real time; 4D building information module, used to clearly display the time schedule and actual progress of construction tasks, and obtain the construction progress of each stage; The data set contains: A data storage module is used to receive data for storage and pre-process the received data to remove duplicate and invalid data; The data analysis module is used to perform real-time analysis on the preprocessed data and display the working status of the data acquisition unit.

3. The digital-based intelligent construction data analysis system according to claim 2 is characterized in that: The prediction unit includes: The historical data collection module is used to obtain the data of the units in the data set to provide a reliable data basis for subsequent predictions; The data comparison module is used to obtain real-time data from the data concentration unit and then conduct multi-dimensional comparative analysis between the obtained real-time data and historical data; The early warning module is used to receive data after multi-dimensional comparative analysis and further identify the data; The repair unit includes: The problem processing module is used to receive the warning problem data transmitted by the warning unit, and match the repair plan after analyzing and processing the data; The self-learning module receives feedback data after the problem is solved and optimizes future repair plans.

4. The digital-based intelligent construction data analysis system according to claim 3 is characterized in that: The data acquisition unit operation process includes: The sensor module acquires and updates the data of various construction equipment in real time, and then builds an IoT network through the sensor module and connects to the data concentration unit to form an IoT architecture; The 4D building information module obtains real-time and past data of construction progress; The unit operation processes in the data set include; After receiving the data from the data acquisition unit, the data storage module stores the data, removes duplicate and invalid data, and classifies and marks the pre-processed data; The data analysis module accesses the data of the data storage module and performs real-time analysis to visualize the working status of the data acquisition unit. The analyzed data will be synchronized to the data acquisition unit to eliminate data silos.

5. The digital-based intelligent construction data analysis system according to claim 4 is characterized in that: The prediction unit operation process includes: The historical data collection module accesses the historical data of the data center unit, extracts and summarizes past construction records, equipment status and key data of resource use; The data comparison module compares and analyzes real-time data with historical data, and then makes a comprehensive dynamic forecast; The early warning module identifies common risks or abnormal conditions based on historical data and combines them with dynamic prediction data, and issues an early warning after identifying similar characteristics. The repair unit operation process includes: After receiving the data of the warning problem, the problem handling module further analyzes it in combination with the historical maintenance records, finds out the root cause, and provides a repair plan; After the problem is solved, the autonomous learning module collects feedback data from the data acquisition unit, verifies the repair results, and updates the system model.

6. The digital-based intelligent construction data analysis system according to claim 5 is characterized in that: The prediction unit includes: The process of comprehensive dynamic prediction by the data comparison module: C t =S real_t -L hist_t , used to compare the difference between real-time data and historical data under the same conditions, where C t is the difference in data at time t, S real_t is the real-time data of time, L hist_t It is the historical data of time; A t =f(C t )×α, used to adjust the weight according to the size and importance of the contrast difference, where A t is the difference weight adjusted value, f(C t ) is the absolute value or square used to calculate the size of the difference, α is the adjustment factor used to give different weights to the differences of different types of data; S pred_t+k =φ1S real_t +φ2S real_t-1 +···+θ1C t +θ2C t-1 +···, which means predicting future trends based on the difference between real-time data and historical data, where S pred_t+k is the data value predicted at the future time point t+k, φ i is the autoregressive coefficient of the historical data in the model, θ j is the moving average coefficient of the difference data; S final_pred_t+k =S perd_t+k +λ×(S real_t+k -S perd_t+k ), dynamically adjust the prediction results according to the new real-time data, where S final_pred_t+k is the final forecast value after adjustment, λ is the dynamic adjustment coefficient, which indicates the adjustment strength of the forecast deviation. (S real_t+k -S perd_t+k ) is the error between actual data and predicted data; the entire comprehensive dynamic prediction process is based on the comparison of real-time data with historical data, which requires a combination of difference analysis, weight adjustment, dynamic prediction model and real-time feedback.

7. The digital-based intelligent construction data analysis system according to claim 6 is characterized in that: The prediction unit includes: The process of issuing an early warning after the early warning module identifies similar features: First, the early warning module extracts relevant features from both real-time data and historical data. The feature vector is represented as: Among them G real_t is the real-time data feature vector at time t, G hist_i is the feature vector at a specific time point i in the historical data, and is the value of the nth feature at time t and i. The similarity between the current feature vector and the historical feature vector is measured by Euclidean distance and cosine similarity. Euclidean distance: where f(G real_t , G hist_i ) Euclidean distance and cosine similarity between current data and historical data: Where yxd(G real_t , G hist_i ) is the cosine similarity between current data and historical data; Early warning judgment: in is the preset threshold of Euclidean distance, τ is the preset threshold of cosine similarity, and the distance between real-time data and its corresponding historical risk data is less than Or if the familiarity is greater than τ, the early warning module determines that the real-time data is similar to the historical risk data and triggers an early warning; the expression formula for issuing early warning information is: Where S t Is the flag of whether the warning is triggered at time t, 1 means triggering the warning, 0 means not triggering. t =1, the early warning module sends an early warning message to the manager.

8. The digital-based intelligent construction data analysis system according to claim 7 is characterized in that: The repair unit includes: The problem handling module receives the warning information and analyzes the root cause and provides a repair solution: Use multiple regression analysis to find the most relevant features. The following is the expression formula: u = δ1z1 + δ2z2 + ··· + δ n z n +ω, where u is the severity of the anomaly, z n are the characteristic values, such as equipment temperature, schedule lag time, etc., δ n is the weight of each feature on the abnormal phenomenon, ω is the error term, which represents the unexplained part of the model, and is expressed by δ n The value is used to find out the characteristics that have the greatest impact on the current risk phenomenon, that is, the root cause. The problem is classified according to the root cause and matched with a specific repair plan. The following is the classification expression formula: V=f(F anomaly ), where V is the problem classification result, indicating which type of problem it belongs to, and f is the classification function, which predicts the problem type based on the characteristics of risk data.

9. The digital-based intelligent construction data analysis system according to claim 8, characterized in that: The repair unit includes: The problem handling module selects a repair solution from the repair library or optimization system according to the problem classification result. The following is the expression formula: A = solution (C), where A is the repair solution, indicating the action to be taken. Among them, FA opt is the optimized repair solution, Ω i is the weight of each loss, M(A, G anomaly ) is the loss function of the repair solution under specific abnormal characteristics, which represents the cost of the repair solution; the expression formula for tracking and feedback of the execution results after executing the repair solution is: where ψ new is the updated regression coefficient, which is used for the next prediction, μ is the learning rate, which is used to adjust the coefficient, It is the gradient of the loss function with respect to the repair solution, indicating the feedback of the repair effect, and dynamically adjusting the model parameters according to the actual execution effect.

10. The digital-based intelligent construction data analysis system according to claim 8, characterized in that: The repair unit includes: The process of updating the system model by the self-learning module: the real-time data, abnormal data, feedback data, and the execution result data of the repair plan are collected centrally, and the collected data is pre-processed to set the model parameters. Then the expression formula is: in is the current model parameter The predicted value obtained, M is the loss function, which measures the error between the model prediction value and the true target value. It is the optimal model parameter. Not only can the model be updated when new data appears, but the model structure or hyperparameters can also be adjusted through adaptive mechanisms. The expression formula is adaptively adjusted: where υ new is the updated learning rate, M prev , Mcurrent is the verification loss of the previous round and the current round. When the verification set loss continues to decrease, the system increases the learning rate accordingly to speed up the convergence; otherwise, the learning rate is reduced to prevent the model from diverging.

Images