Large reinforced concrete building risk early warning system and method

Abstract

The invention belongs to the technical field of intelligent buildings, and discloses a large reinforced concrete building risk early warning system and method.The system comprises an embedded sensor module, a data acquisition module, a signal processing module, a risk assessment module and an early warning output module, and the health state of a large reinforced concrete building is monitored in real time; and the health state and the potential risk of the building are comprehensively analyzed and evaluated, and an early warning signal is sent in time to ensure that related personnel can quickly take countermeasures. According to the invention, real-time monitoring and risk early warning of large reinforced concrete buildings are realized.

Description

Technical Field

[0001] The present application belongs to the field of intelligent building technology, and in particular, relates to a risk early warning system and method for large reinforced concrete buildings. Background Art

[0002] In recent years, due to climate factors, aging of buildings and other reasons, large buildings have been damaged or even collapsed, causing many irreversible hazards.

[0003] Large reinforced concrete structures, such as dams, nuclear power plant main buildings, bridges, etc., may collapse due to external factors such as earthquakes and typhoons. If these large structures can monitor the changes in the reinforced concrete benchmark structure through sensors before they may collapse or be damaged, and carry out targeted preventive maintenance, it can greatly reduce the loss of life and property. Summary of the invention

[0005] The purpose of this application is to provide a risk warning system and method for large reinforced concrete buildings, which can monitor and evaluate the collapse risks caused by external factors such as earthquakes and typhoons, and issue warnings in time to reduce casualties and property losses and improve the safety and reliability of buildings.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] In a first aspect, the present application provides a large reinforced concrete construction risk early warning system, comprising:

[0008] Embedded sensor modules for real-time monitoring of the health of large reinforced concrete structures;

[0009] A data acquisition module for collecting data from the embedded sensor module;

[0010] A signal processing module is used to process and analyze the raw data transmitted from the data acquisition module;

[0011] The risk assessment module evaluates the health status and potential risks of the building by comprehensively analyzing the data provided by the signal processing module;

[0012] The early warning output module is used to issue early warning signals in a timely manner based on the analysis results provided by the risk assessment module to ensure that relevant personnel can take countermeasures quickly.

[0013] In some embodiments, the embedded sensor module is composed of a plurality of sensors, including an ultrasonic sensor chip, a steel bar strain gauge, and a linear variable differential transformer.

[0014] In some embodiments, the structure of the data acquisition module includes a multi-channel data acquisition unit, a preamplifier and an amplifier, a signal filter, a data storage unit, and a communication interface.

[0015] In some embodiments, the data acquisition module collects various types of sensor data in real time, and the data exported by the sensors on the steel bars constitute the initial database. When the internal damage intensifies, the data and the acquisition continue to change. The data is stored and preliminarily processed through the acquisition module to provide reliable data support for subsequent signal processing and risk assessment.

[0016] In some embodiments, the high quality data generated by the signal processing module is further used to evaluate the bending stiffness of the reinforced concrete beam under vertical loads that increase over time.

[0017] In some embodiments, the risk assessment module includes:

[0018] The data analysis unit receives and processes high-quality data from the signal processing module to identify structural anomalies and damage;

[0019] The random forest algorithm unit further processes these data. These data refer to: in the initial test state, there must first be a damage database. These data are mainly simulated data composed of changes in the internal stress of reinforced concrete when changes occur. As one of the databases for computer learning, in addition, it receives data during the learning process to determine which data exceeds the normal range, and then establishes a data set to improve the existing database.

[0020] The evaluation decision unit combines the results of data analysis and random forest algorithm to evaluate the health status and damage category of the building.

[0021] In some embodiments, the risk assessment module includes a baseline model and a determination of abnormal conditions, the baseline model being established by statistical analysis and feature extraction of data under normal operating conditions;

[0022] Assume that the stress, deflection and ultrasonic signal data under normal operating conditions are X0, which contains n features:

[0023] The normality and abnormality of the data are judged by the method of difference calculation and threshold setting, and the mean μ and standard deviation σ are used for setting:

[0024] ΔX=X t -X0

[0025] T=μ+kσ

[0026] Where, X tis the real-time data vector collected at a specific time, X0 is the data vector in the benchmark model, T is the threshold, μ is the mean of the benchmark model data, σ is the standard deviation of the benchmark model data, and k is the safety factor.

[0027] In a second aspect, the present application provides a large reinforced concrete construction risk early warning method, comprising:

[0028] Real-time monitoring of the health status of large reinforced concrete structures, collecting stress, deflection and ultrasonic signals inside the structure;

[0029] Collect data from various sensors in real time, and the data derived from the sensors on the steel bars form the initial database;

[0030] Process and analyze the transmitted raw data, denoise, normalize and time-align the signals, and then use continuous wavelet transform to extract features and fuse information to generate high-quality data;

[0031] Comprehensively analyze data to assess the health status and potential risks of buildings, identify anomalies and damage in real time, and generate risk warning signals;

[0032] Based on the analysis results, an early warning signal is issued to ensure that relevant personnel can take countermeasures quickly.

[0033] In some embodiments, when internal damage intensifies, data and acquisition undergo continuous changes, and data is stored and preliminarily processed through the acquisition module to provide reliable data support for subsequent signal processing and risk assessment.

[0034] In some embodiments, signal preprocessing is performed in a signal processing module, including the following steps:

[0035] Perform denoising: Use a bandpass filter to filter the signal, remove low-frequency and high-frequency noise, and only retain the frequency band signal of interest;

[0036] Normalization: standardize the amplitude of the signal to eliminate the measurement differences between different sensors, and time-align the signal to ensure that the signals obtained at different measurement times have a consistent time reference;

[0037] Signal segmentation: Divide long-term signals into multiple short-time windows to facilitate subsequent feature extraction and analysis.

[0038] In some embodiments, when comprehensively analyzing the data to assess the health status and potential risks of a building, a random forest algorithm is used to aggregate and analyze the data to determine the final damaged or undamaged category, specifically including:

[0039] Data sampling: For a data set containing N samples, random forest extracts multiple subsets with replacement from the original data set through the bootstrap sampling method. The size of each subset is N, and a total of K selections are made;

[0040] Construct a decision tree: Based on the K sub-datasets, for each sampling subset, construct a decision tree. When each node splits, randomly select a subset of features and select the optimal split feature in the subset.

[0041] Integration results: For classification tasks, random forest integrates the classification results of each decision tree through majority voting;

[0042] The data set D = {(x1, y1), (x2, y2), ..., (x N ,y N )}, where x i is the feature vector, y i is the label; use the training data set D train Build a random forest model: RandomForestClassifier(n_estimators=35, max_depth=10), where n_estimators is the number of decision trees and max_depth is the maximum depth of each tree;

[0043] Using the test dataset D test Make predictions and get prediction results

[0044]

[0045] In the formula, h i (x) is the prediction result of the i-th decision tree for the input sample x, K is the total number of decision trees, and mode is majority voting.

[0046] Compared with the prior art, the large reinforced concrete building risk early warning system and method provided by this application has the following beneficial effects:

[0047] This application realizes real-time monitoring and risk warning of large reinforced concrete structures.

[0048] This application realizes real-time monitoring: it can monitor the health status of the building in real time, detect key parameters such as cracks inside the concrete, stress changes in the steel bars, and deflection of the structure, thereby improving the comprehensiveness and accuracy of monitoring and providing rich data support for the assessment of the health status of the building.

[0049] This application achieves accurate assessment: through multi-level information fusion and random forest algorithm, the health status of the building is accurately assessed to identify potential structural anomalies and damage.

[0050] This application achieves timely warning: when a major risk is detected in a building, the system can trigger a warning signal in a timely manner and pass the warning information to relevant personnel through multiple channels to ensure that response measures are taken quickly to minimize casualties and property losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for the technical description.

[0052] Figure 1 A schematic diagram of the structure of the risk early warning system for large reinforced concrete buildings provided for this application;

[0053] Figure 2 Schematic diagram of sensor arrangement provided for this application;

[0054] Figure 3 The signal processing flow chart provided for this application;

[0055] Figure 4 A flowchart of the decision tree generation provided for this application;

[0056] Figure 5 Flowchart of the random forest algorithm provided for this application

[0057] Figure 6 A flow chart of the large reinforced concrete construction risk warning method provided for this application. DETAILED DESCRIPTION

[0058] The following is further explained in detail through specific implementation methods.

[0059] The present application provides a risk early warning system for large reinforced concrete buildings. Several ultrasonic sensor chips are embedded in the main building parts of large reinforced concrete, such as the reinforced concrete walls of nuclear island plants. When large-scale benchmark reinforced concrete structures are cast, they are installed on the vertical or horizontal stirrups of reinforced concrete. Two steel bar stress gauges (installed on the top and bottom of the steel bars) are embedded in the concrete. A linear variable differential transformer is installed outside to monitor deflection.

[0060] like Figures 1 to 5 As shown, the system includes an embedded sensor module, a data acquisition module, a signal processing module, a risk assessment module, and a warning output module.

[0061] Embedded sensor modules are used to monitor the health status of large reinforced concrete structures in real time. By collecting stress, deflection and ultrasonic signals inside the structure, accurate data support is provided to assess the risk status of the building.

[0062] The data acquisition module is used to collect data from the embedded sensor module, including stress, deflection and ultrasonic signals. These data are an important basis for evaluating the health status and potential risks of large reinforced concrete structures. The data acquisition module is responsible for real-time acquisition of various sensor data. The data derived from the sensors on the steel bars form the initial database. When the internal damage increases, the data and acquisition continue to change. The data is stored and preliminarily processed through the acquisition module to provide reliable data support for subsequent signal processing and risk assessment.

[0063] The signal processing module is used to process and analyze the raw data transmitted from the data acquisition module. The signal processing module first denoises, normalizes and time-aligns the signal through the preprocessing unit, and then uses continuous wavelet transform (CWT) to extract features and fuse information to generate high-quality data.

[0064] The risk assessment module evaluates the health status and potential risks of the building by comprehensively analyzing the data provided by the signal processing module. The risk assessment module identifies anomalies and damage in real time and generates risk warning signals.

[0065] The risk assessment module includes the benchmark model and the identification of abnormal conditions, where the benchmark model is established by statistical analysis and feature extraction of data under normal operating conditions. The stress, deflection and ultrasonic signal data under normal operating conditions are set as X0, which contains n features:

[0066] The normality and abnormality of data are judged by difference calculation and threshold setting. The threshold can be determined based on the statistical characteristics of historical data and experimental data. Generally, the mean μ and standard deviation σ are used for setting:

[0067] ΔX=X t -X0

[0068] T=μ+kσ

[0069] Where, X t is the real-time data vector collected at a specific time, X0 is the data vector in the benchmark model, representing the stress, deflection, and ultrasonic signal data under normal operating conditions. T represents the threshold, μ represents the mean of the benchmark model data, σ represents the standard deviation of the benchmark model data, and k is a safety factor used to adjust the sensitivity of the threshold.

[0070] The early warning output module is used to issue early warning signals in a timely manner based on the analysis results provided by the risk assessment module to ensure that relevant personnel can take countermeasures quickly. The early warning output module comprehensively analyzes sensor data and assessment results, and immediately triggers the early warning mechanism when a potential collapse risk of a building is detected. The early warning information includes the structure name, location, monitoring time, risk assessment results and recommended measures, and is communicated through multiple channels (sound and light alarm, SMS notification, email and mobile push).

[0071] In one embodiment, the embedded sensor module is composed of a plurality of sensors, including an ultrasonic sensor chip, a steel bar stress gauge, and a linear variable differential transformer. Figure 2 As shown, four sensors are set, namely the first sensor 1, the second sensor 2, the third sensor 3, and the fourth sensor 4. The ultrasonic sensor chip is installed on the vertical and horizontal stirrups of reinforced concrete to monitor the quality and cracks inside the concrete. The steel bar stress gauges are installed on the top and bottom of the steel bars to measure the stress changes of the steel bars. The linear variable differential transformer is installed outside the structure to monitor the deflection changes of the structure. The ultrasonic sensor chip adopts the ultrasonic pulse velocity method, which is most commonly used to detect the quality of concrete, the location or depth of cracks inside steel bars and masonry structures. The initiation of cracks will cause an increase in ultrasonic velocity and attenuation, and can be used as a damage indicator.

[0072] In one embodiment, the structure of the data acquisition module includes a multi-channel data acquisition unit, a preamplifier and an amplifier, a signal filter, a data storage unit, and a communication interface. The ultrasonic signal is collected seven times per minute and stored in a data recorder. Considering the appropriate resolution of the acquired signal, the center frequency of the measured ultrasonic signal is about 60kHz and the sampling frequency is 1MHz. The recording duration of each measurement is 5 milliseconds, allowing multiple reflections to occur; the steel bar stress gauge and the linear variable differential transformer data are automatically collected every 200 milliseconds, and the data is transmitted to the signal processing module and the risk assessment module in real time through the communication interface. At the same time, from the ultrasonic features, the formation of microcracks inside the beam can be detected and located. By applying a static load test on the reinforced concrete beam to test the deformation process of the crack, four embedded ultrasonic sensors are arranged in four points, and different features are extracted according to the transmitted ultrasonic signals. A benchmark model is established through historical data and simulated experimental data, including stress, deflection and ultrasonic signal data under normal operating conditions.

[0073] In one embodiment, the purpose of the signal preprocessing in the signal processing module is to perform preliminary processing on the original signal to remove noise and interference and improve the quality and reliability of the signal. The preprocessing step includes the following steps:

[0074] De-noising is performed by filtering the signal with a bandpass filter to remove low-frequency and high-frequency noise, and only retaining the frequency band signal of interest; de-noising can be performed by wavelet transform denoising, that is, performing wavelet decomposition, threshold processing and signal reconstruction on the pre-processed signal to eliminate the noise component. The specific process is as follows:

[0075] Wavelet decomposition: normalized radar signal X norm (t) Perform multi-scale wavelet decomposition to decompose the signal into detail coefficients d in different frequency bands j,k and the approximation coefficient c j,k :

[0076]

[0077] Normalization processing standardizes the amplitude of the signal to eliminate the measurement differences between different sensors and time-aligns the signal to ensure that the signals obtained at different measurement times have a consistent time reference;

[0078] Finally, the signal is segmented to separate the long-term signal into multiple short-time windows to facilitate subsequent feature extraction and analysis.

[0079] The signal processing process involves extracting load / crack sensitive features from the data collected during data acquisition to diagnose the deformation of the structure. Feature extraction is done by applying signal processing techniques to the sensor data. Continuous wavelet transform (CWT) is an effective signal processing method that can be used as a change / damage feature indicator. Wavelet transform is an estimation method used to quantify the energy dispersion value from the acquired signal and is able to analyze the signal simultaneously in two dimensions: time and frequency.

[0080] Preprocessing signals of ultrasonic sensors in concrete structures It may include changes caused by cracks or other defects; ψ(t) is the mother wavelet function, then The continuous wavelet transform CWT(a, b) is defined as:

[0081]

[0082] In the formula, a is the scale parameter, which controls the expansion and contraction of the wavelet and reflects the frequency information of the signal; b is the translation parameter, which controls the position of the wavelet and reflects the time information of the signal; Indicates signal Convolution with a wavelet function of scale a and centered at b.

[0083] The mother wavelet function ψ(t) is the Morlet wavelet, which is a complex-valued wavelet and is defined as follows:

[0084]

[0085] Where w0 is the center frequency, which is usually selected as 5 or 6 to ensure that the wavelet has good frequency resolution and time resolution.

[0086] By analyzing the wavelet coefficients of CWT (a, b) at different specific scales a and positions b, the features in the signal can be extracted, such as spectrum peaks, mutation points, areas where signal energy is concentrated, such as obvious high-frequency components; these features can be used to identify cracks, voids or other defects in concrete structures.

[0087] Information fusion is the comparison of information from multiple sensors to improve overall decision making and positioning. Data fusion can be divided into three levels: signal-level fusion, feature-level fusion, and decision-level fusion. Signal-level fusion is the lowest level of fusion, which combines the original signals from multiple sensors to produce a new original signal, which is expected to contain more useful information; feature-level fusion, also known as mid-level fusion, involves calculating feature values ​​from each sensor pair individually or comparing the features of different sensors to select the most relevant features for decision making. In decision-level fusion, each sensor can provide an independent decision based on its own features, and then the results of all features are compared to obtain an overall decision and positioning. This fusion is called the highest level of fusion. Through these three levels of information fusion, useful information can be extracted more accurately and the assessment of structural health status can be improved.

[0088] At the same time, the high-quality data generated by the signal processing module is further used to evaluate the bending stiffness of reinforced concrete beams under vertical loads that increase over time, specifically by measuring deflection as a performance indicator. The evaluation stages of bending stiffness can be divided into four stages: (I) crack-free stage, (II) crack formation stage, (III) crack stabilization stage, and (IV) collapse stage. Collapse usually manifests itself as bending of steel bars or crushing of concrete. During the test, a hydraulic jack press is used to slowly and proportionally increase the load force. While measuring the load force, the data of all sensors and measuring devices are recorded.

[0089] In one embodiment, the risk assessment module consists of a data analysis unit, a random forest algorithm unit, and an evaluation and decision unit. The data analysis unit is responsible for receiving and processing high-quality data from the signal processing module to identify abnormalities and damage of the structure. The random forest algorithm unit further processes these data. The specific process includes: extracting key features from ultrasonic signals, such as crack location, depth, and expansion speed; using historical data to train the random forest model, and constructing multiple decision trees to classify and judge the features; classifying the input features through each decision tree, weighing the weights of the undamaged state and the loaded or damaged state; using the receiver operating characteristic (ROC) curve to statistically evaluate the model evaluation results to visualize the overall decision process; and summarizing the voting results of all decision trees through the random forest algorithm to determine the final damaged or undamaged category. Finally, the evaluation and decision unit combines the results of data analysis and the random forest algorithm to evaluate the health status and damage category of the building.

[0090] A random forest consists of multiple decision trees, each of which is built on a different subset of the training data. The specific steps are as follows:

[0091] Data sampling: For a data set containing N samples, Random Forest uses the Bootstrap Sampling method to extract multiple subsets with replacement from the original data set. The size of each subset is N, and a total of K subsets are selected.

[0092] Construct a decision tree: Based on the K subsets of data, construct a decision tree for each sampling subset. When each node splits, randomly select a subset of features and select the optimal split feature in the subset.

[0093] Integration results: For classification tasks, Random Forest integrates the classification results of each decision tree through majority voting.

[0094] The data set D = {(x1, y1), (x2, y2), ..., (x N ,y N )}, where x i is the eigenvector, y i is the label (damaged or not damaged); using the training dataset D train Build a random forest model: RandomForestClassifier(n_estimators=35, max_depth=10)

[0095] Where n_estimators is the number of decision trees and max_depth is the maximum depth of each tree.

[0096] Using the test dataset D test Make predictions and get prediction results

[0097]

[0098] In the formula, h i (x) is the prediction result of the i-th decision tree for the input sample x, K is the total number of decision trees, and mode represents majority voting. The performance of the model on the test data is evaluated by calculating the accuracy, recall and F1 score of the model, so as to effectively evaluate the damaged or undamaged category of the concrete structure and provide reliable technical support for structural health monitoring and risk assessment.

[0099] In one embodiment, when the indication is a damage category, the crack stabilization stage is entered. The early warning trigger is based on the output results of the risk assessment module. When the concrete structure is detected to be at risk of collapse, the corresponding early warning mechanism is triggered. The early warning information can be transmitted to relevant personnel in a variety of ways to ensure that the information can be conveyed in a timely manner and attract attention. The basic content of the early warning includes: structure name, location, monitoring time, risk assessment results, recommended measures, etc. The information transmission method includes but is not limited to: sound and light alarm, sound and light alarm device at the monitoring site to issue an alarm; SMS notification, SMS notification to relevant persons in charge and maintenance personnel; email, send detailed risk assessment reports and recommended measures through the email system; push warning information in the mobile terminal program. As the risk assessment results trend to a higher risk level, the push frequency will be increased, or as the risk assessment results trend to a lower risk level, the push frequency will be reduced.

[0100] The process of the early warning output module is first initiated by the early warning trigger unit. When the risk assessment module detects that the building has a potential risk of collapse, the early warning trigger unit receives the risk signal and determines whether to trigger an early warning based on the preset threshold. Next, the information processing unit processes and formats the early warning information to generate a detailed report containing the structure name, location, monitoring time, risk assessment results and recommended measures. Then, the multi-channel notification system is activated to transmit the early warning information in a variety of ways: the sound and light alarm device sounds an alarm at the monitoring site to remind the on-site personnel to pay attention; the SMS notification system sends early warning SMS to the relevant person in charge and maintenance personnel; the email system sends a detailed risk assessment report and recommended measures; the mobile push application adjusts the push frequency according to the danger level of the risk assessment results to ensure timely delivery of information. Through these steps, the early warning output module ensures that the early warning information can be accurately and quickly transmitted to all relevant personnel, so that they can quickly take countermeasures to minimize casualties and property losses and ensure the safety and stability of large reinforced concrete structures.

[0101] In addition, based on the above-mentioned large reinforced concrete construction risk warning system, the present application also provides a large reinforced concrete construction risk warning method, such as Figure 6 As shown, specifically including:

[0102] Through embedded sensor modules, the health status of large reinforced concrete structures is monitored in real time, and stress, deflection and ultrasonic signals inside the structure are collected to provide accurate data support and assess the risk status of the building;

[0103] The data acquisition module collects data from the embedded sensor module, including stress, deflection and ultrasonic signals. The data acquisition module is responsible for real-time acquisition of various sensor data. The data derived from the sensors on the steel bars form the initial database. When the internal damage increases, the data and acquisition will continue to change. The data acquisition module is used to store and preliminarily process the data, providing reliable data support for subsequent signal processing and risk assessment.

[0104] The signal processing module processes and analyzes the raw data transmitted from the data acquisition module, and the preprocessing unit performs denoising, normalization and time alignment on the signal. Then, continuous wavelet transform (CWT) is used to extract features and fuse information to generate high-quality data.

[0105] The risk assessment module comprehensively analyzes the data provided by the signal processing module to evaluate the health status and potential risks of the building, identifies abnormalities and damage in real time, and generates risk warning signals;

[0106] According to the analysis results provided by the risk assessment module, early warning signals are issued in a timely manner through the early warning output module to ensure that relevant personnel can take countermeasures quickly.

[0107] In order to effectively carry out risk warning in an environment where the building is intact and static, this application also includes the establishment of a benchmark model and a data comparison mechanism. In an environment where the building is intact and static, no new monitoring data will be generated within the system, so data comparison and learning are required through modeling and historical data analysis.

[0108] In one embodiment, the signal preprocessing is performed in the signal processing module to perform preliminary processing on the original signal to remove noise and interference and improve the quality and reliability of the signal. The preprocessing step includes the following steps:

[0109] Perform denoising processing, use a bandpass filter to filter the signal, remove low-frequency and high-frequency noise, and only retain the frequency band signal of interest;

[0110] Normalization processing standardizes the amplitude of the signal to eliminate the measurement differences between different sensors and time-aligns the signal to ensure that the signals obtained at different measurement times have a consistent time reference;

[0111] Finally, the signal is segmented to separate the long-term signal into multiple short-time windows to facilitate subsequent feature extraction and analysis.

[0112] This application provides an effective risk warning method for large reinforced concrete structures (such as dams, nuclear power plant main buildings, bridges, etc.) through sensors and data analysis technology. This method can monitor and evaluate the collapse risk caused by external factors such as earthquakes and typhoons, and issue early warnings in a timely manner to reduce casualties and property losses and improve the safety and reliability of buildings.

[0113] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.

Claims

1. A large reinforced concrete construction risk early warning system, characterized in that: include: Embedded sensor modules for real-time monitoring of the health of large reinforced concrete structures; A data acquisition module for collecting data from the embedded sensor module; A signal processing module is used to process and analyze the raw data transmitted from the data acquisition module; The risk assessment module evaluates the health status and potential risks of the building by comprehensively analyzing the data provided by the signal processing module; The early warning output module is used to issue early warning signals in a timely manner based on the analysis results provided by the risk assessment module to ensure that relevant personnel can take countermeasures quickly.

2. The large reinforced concrete construction risk early warning system according to claim 1 is characterized in that: The embedded sensor module consists of multiple sensors, including ultrasonic sensor chips, steel bar strain gauges, and linear variable differential transformers.

3. The large reinforced concrete construction risk early warning system according to claim 1 is characterized in that: The structure of the data acquisition module includes a multi-channel data acquisition unit, a preamplifier and an amplifier, a signal filter, a data storage unit, and a communication interface.

4. The large reinforced concrete construction risk early warning system according to claim 1 is characterized in that: The data acquisition module collects data from various sensors in real time. The data exported by the sensors on the steel bars form the initial database. When the internal damage intensifies, the data and acquisition continue to change. The data is stored and preliminarily processed by the acquisition module to provide reliable data support for subsequent signal processing and risk assessment.

5. The large reinforced concrete construction risk early warning system according to claim 1 is characterized in that: The risk assessment modules include: The data analysis unit receives and processes high-quality data from the signal processing module to identify structural anomalies and damage; The random forest algorithm unit further processes these data. The evaluation decision unit combines the results of data analysis and random forest algorithm to evaluate the health status and damage category of the building.

6. The large reinforced concrete construction risk early warning system according to claim 1 is characterized in that: The risk assessment module includes a baseline model and the identification of abnormal situations. The baseline model is established by statistical analysis and feature extraction of data under normal operating conditions; Assume that the stress, deflection and ultrasonic signal data under normal operating conditions are X0, which contains n features: The normality and abnormality of the data are judged by the method of difference calculation and threshold setting, and the mean μ and standard deviation σ are used for setting: ΔX=X t -X0 T=μ+kσ Where, X t is the real-time data vector collected at a specific time, X0 is the data vector in the benchmark model, T is the threshold, μ is the mean of the benchmark model data, σ is the standard deviation of the benchmark model data, and k is the safety factor.

7. A risk warning method for large reinforced concrete construction, characterized in that: include: Real-time monitoring of the health status of large reinforced concrete structures, collecting stress, deflection and ultrasonic signals inside the structure; Collect data from various sensors in real time, and the data derived from the sensors on the steel bars form the initial database; Process and analyze the transmitted raw data, denoise, normalize and time-align the signals, and then use continuous wavelet transform to extract features and fuse information to generate high-quality data; Comprehensively analyze data to assess the health status and potential risks of buildings, identify anomalies and damage in real time, and generate risk warning signals; Based on the analysis results, an early warning signal is issued to ensure that relevant personnel can take countermeasures quickly.

8. The large reinforced concrete construction risk early warning method according to claim 1 is characterized in that: When internal damage intensifies, data and acquisition undergo continuous changes, and data storage and preliminary processing are performed through the acquisition module to provide reliable data support for subsequent signal processing and risk assessment.

9. The large reinforced concrete building risk early warning method according to claim 1 is characterized in that: Signal preprocessing is performed in the signal processing module, including the following steps: Perform denoising: Use a bandpass filter to filter the signal, remove low-frequency and high-frequency noise, and only retain the frequency band signal of interest; Normalization: standardize the amplitude of the signal to eliminate the measurement differences between different sensors, and time-align the signal to ensure that the signals obtained at different measurement times have a consistent time reference; Signal segmentation: Divide long-term signals into multiple short-time windows to facilitate subsequent feature extraction and analysis.

10. The large reinforced concrete construction risk early warning method according to claim 1, characterized in that: When comprehensively analyzing the data to assess the health status and potential risks of the building, the random forest algorithm is used to summarize and analyze the data to determine the final damaged or non-damaged categories, including: Data sampling: For a data set containing N samples, random forest extracts multiple subsets with replacement from the original data set through the bootstrap sampling method. The size of each subset is N, and a total of K selections are made; Construct a decision tree: Based on the K sub-datasets, for each sampling subset, construct a decision tree. When each node splits, randomly select a subset of features and select the optimal split feature in the subset. Integration results: For classification tasks, random forest integrates the classification results of each decision tree through majority voting; The data set D = {(x1, y1), (x2, y2), ..., (x N ,y N )}, where x i is the feature vector, y i is the label; use the training data set D train Build a random forest model: RandomForestClassifier(n_estimators=35, max_depth=10), where n_estimators is the number of decision trees and max_depth is the maximum depth of each tree; Using the test dataset D test Make predictions and get prediction results In the formula, h i (x) is the prediction result of the i-th decision tree for the input sample x, K is the total number of decision trees, and mode is majority voting.

Images