Abnormity recognition system for monitoring water leakage of fabricated building

Abstract

The invention, which relates to the technical field of water leakage monitoring, discloses an abnormity identification system for monitoring water leakage of an assembled building body, and the system comprises a data transmission guarantee center which is in communication connection with a building body data acquisition module, a building body risk prediction module and a building body data analysis module. The building data acquisition module collects meteorological, hydrological and structural data, the risk prediction module predicts leakage risks by using an autoregression integral moving average model and a support vector machine algorithm to realize early warning, and the data analysis module draws a risk distribution map through a geographic information system, provides a visual interface and transmits data by using encryption and fault-tolerant technologies. According to the invention, comprehensive supervision of water leakage of the fabricated building is realized through the integrated system, the accuracy and timeliness of leakage risk identification are improved, a visual risk distribution map is provided, rapid response and processing are facilitated, the safety of data transmission is guaranteed, the leakage risk is effectively reduced, and the maintenance and management level of the building is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water leakage monitoring, and in particular to an abnormality recognition system for monitoring water leakage in assembled buildings. Background Art

[0002] Leakage monitoring in prefabricated buildings is a systematic process aimed at ensuring effective waterproofing measures are implemented throughout the design and construction process, thereby reducing or preventing leakage. This monitoring typically encompasses multiple aspects, including design review, material quality control, construction oversight, and subsequent quality inspection. During the design phase, regulatory authorities will require designers to fully consider waterproofing requirements, clarify waterproofing structures and joint practices, and provide specific design briefings during drawing review. During construction, regulatory authorities will urge construction companies to strictly adhere to the waterproofing plan, ensuring that the quality of waterproofing materials and construction processes meet standards. Furthermore, regulatory authorities will focus on monitoring key areas and processes prone to leakage, such as the sealing of prefabricated exterior wall joints and the connection between window frames and exterior walls. During the subsequent quality inspection phase, regulatory authorities will conduct water spray tests and other testing methods to verify the waterproofing effectiveness of prefabricated buildings. For any identified issues, regulatory authorities will require the responsible units to promptly rectify them and monitor the progress of these rectifications to ensure they are effectively resolved. In general, water leakage supervision of prefabricated buildings is a whole process supervision from design, construction to acceptance, aiming to ensure the waterproof quality of prefabricated buildings and the living safety of the people.

[0003] To address the issues of risk prediction and early warning in prefabricated building water leakage monitoring, existing technologies primarily rely on manual inspections and regular checks. However, this approach is limited by manpower and time resources, making it difficult to promptly detect and address potential leakage risks. This, in turn, leads to water leakage problems during the use of prefabricated buildings, which not only affects the building's functionality and comfort, but also causes more serious structural safety issues. Furthermore, existing technologies rely heavily on empirical judgment to identify and address leakage problems, lacking scientific data support and precise analytical methods, resulting in unsatisfactory monitoring results. To address this issue, an anomaly identification system for prefabricated building water leakage monitoring is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide an abnormality identification system for monitoring water leakage in prefabricated buildings to solve the problems raised in the above-mentioned background technology.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: an abnormality identification system for water leakage monitoring of prefabricated buildings, including a data transmission guarantee center, wherein the data transmission guarantee center is communicatively connected to a building data acquisition module, a building risk prediction module, and a building data analysis module; The building data acquisition module collects meteorological data, hydrological data and structural data of the prefabricated building through temperature and humidity sensors, rain gauges, groundwater level monitors, strain gauges and construction records; The building risk prediction module uses an autoregressive integral moving average model to predict the leakage risk of prefabricated buildings, and uses a support vector machine classification algorithm to train the model to distinguish between normal and abnormal states, thereby achieving early warning of potential leakage points in prefabricated buildings. The building data analysis module uses geographic information system technology and real-time data streams to draw a water leakage risk distribution map for prefabricated buildings, provides a visual interface to view the status of various parts of the building, and uses encryption technology and fault-tolerant mechanisms to transmit data.

[0006] A further improvement of the technical solution of the present invention is that: in the building data acquisition module, the process of collecting meteorological data, hydrological data, and structural data of the prefabricated building through temperature and humidity sensors, rain gauges, groundwater level monitors, strain gauges, and construction records includes: Temperature and humidity sensors are deployed in the basement, corridors, and exterior walls of the building. They measure the temperature and relative humidity of the environment using changes in resistance. The internal circuits of the temperature and humidity sensors convert the physical signals into electrical signals, which are then converted into digital signals using analog-to-digital converters to collect temperature and humidity data. Rain gauges are deployed on building roofs and open areas on the ground. They use a tipping bucket design. Rainwater flows through the funnel into a water collection container. A water collection threshold is set. When the water volume reaches the collection threshold, the tipping bucket flips and records the rainfall. Deploy groundwater level monitors in the groundwater around the building, using pressure sensors to determine the groundwater level by measuring the pressure of the water column on the sensor; Strain gauges are deployed on the beams, columns, and walls of the building. Based on the resistance strain effect of metal materials, the resistance of the material changes when it is deformed by force. The strain gauges are attached to the surface of the building. When the building deforms, the resistance value of the strain gauge changes. The resistance change is converted into a voltage signal for measurement through a bridge circuit. Key information during the construction process is recorded manually and regularly, including the location of expansion joints and post-casting strips and the quality of construction, as well as the construction time, process, materials used and test results.

[0007] A further improvement of the technical solution of the present invention is that: in the building data acquisition module, the process of preprocessing the collected temperature and humidity data, total precipitation, groundwater level, building strain value and construction records includes: Set the temperature threshold between -20°C and 50°C, and the relative humidity threshold between 0% and 100%, conduct a preliminary check on the collected temperature and humidity data, remove abnormal values, and perform standardization. The standardized temperature data , standardized humidity data ,in, is the original temperature data, is the relative humidity data; Remove the abnormal values of the bucket flipping time in succession in the dumping record, and accumulate the number of bucket flipping times to calculate the total precipitation P. , where N is the number of bucket flips, and X is the amount of precipitation corresponding to each flip, in millimeters; Remove the fluctuation abnormal value in the pressure sensor measurement value and convert the pressure value measured by the pressure sensor into the groundwater level height H. ,in is the pressure value measured by the pressure sensor, is the density of groundwater, g is the acceleration due to gravity; Remove the fluctuation abnormal value in the strain gauge resistance change and convert the strain gauge resistance change into strain value , , where k is the strain gauge sensitivity coefficient, is the building structure strain value, is the initial resistance value of the strain gauge.

[0008] A further improvement of the technical solution of the present invention is that: in the building data acquisition module, the process of extracting features from the collected temperature and humidity data, total precipitation, groundwater level, building strain value, and construction records includes: Calculate the daily average temperature and relative humidity change rate. The calculation process is as follows: ; ; in, and are the initial and final relative humidity values during the acquisition period, and They are the start time and end time of the collection respectively; The daily cumulative precipitation, hourly precipitation change rate, daily groundwater level change rate, and daily strain value change rate were calculated. The text description of the construction record was converted into a standard format, and the key features in the construction record were extracted, including the spacing of expansion joints, the location of post-cast strips, and the concrete strength grade.

[0009] A further improvement of the technical solution of the present invention is that: in the building risk prediction module, the process of training the autoregressive integrated moving average model includes: Constructing an autoregressive integrated moving average model, the autoregressive integrated moving average model includes an autoregressive term coefficient, a difference order, and a moving average term number. The number of autoregressive terms p is determined by a partial autocorrelation function graph, and the difference order d is determined by observing the stationarity of the time series data. If the original data is non-stationary, difference processing is performed until the data is stationary, and the number of moving average terms q is determined by an autocorrelation function graph. Using the collected historical data, the autoregressive integrated moving average model is trained. The calculation process is as follows: ; in, represents the penetration probability of the target building at time t, c is a constant term, is the white noise error term, and are the autoregressive and moving average coefficients, respectively.

[0010] A further improvement of the technical solution of the present invention is that: in the building risk prediction module, the process of predicting the leakage risk of prefabricated buildings using the trained autoregressive integral moving average model includes: The preprocessed temperature and humidity data, total precipitation, groundwater level and building strain value are used as input data of the autoregressive integral moving average model. According to the output results of the autoregressive integral moving average model, the leakage risk prediction value of the target building is obtained.

[0011] A further improvement of the technical solution of the present invention is that: in the building risk prediction module, the process of training the support vector machine classification model includes: From the processed data, standard values of temperature and relative humidity, total precipitation, groundwater level, building strain, expansion joint spacing, post-casting strip location, and concrete strength grade were selected as input features for the support vector machine classification model. 70% of the data was used as the training set, and 30% of the data was used as the test set. The support vector machine classification model is trained using the training set data, the radial basis function kernel is selected, and the optimal hyperparameter combination is determined by the cross-validation method. The hyperparameters include the regularization parameter C and the kernel function parameter , the training process is as follows: ; in, is the output of the support vector machine classification model, is the Lagrange multiplier, is the sample label, 1 indicates normal state, -1 indicates abnormal state, is the radial basis function kernel, b is the bias term; The performance of the support vector machine classification model is evaluated using the test set. The accuracy of the support vector machine classification model is calculated by comparing the prediction results of the support vector machine classification model with the actual labels. The calculation process is as follows: ; Generate a confusion matrix and analyze the classification effect of the support vector machine classification model. The confusion matrix consists of four parts: true positives, false positives, true negatives, and false negatives. Among them, true positives are the number of samples that are actually abnormal and correctly predicted to be abnormal, false positives are the number of samples that are actually normal but incorrectly predicted to be abnormal, true negatives are the number of samples that are actually normal and correctly predicted to be normal, and false negatives are the number of samples that are actually abnormal but incorrectly predicted to be normal. Use precision, recall, and F1 score as evaluation indicators to comprehensively evaluate the performance of the support vector machine classification model.

[0012] A further improvement of the technical solution of the present invention is that: in the building risk prediction module, based on the support vector machine classification model, the process of distinguishing normal and abnormal states and achieving early warning of potential leakage points in prefabricated buildings includes: The preprocessed temperature and humidity data, total precipitation, groundwater level and building strain value are used as input data of the support vector machine classification model. According to the output results of the support vector machine classification model, the target building state is obtained. The target building state includes normal state and abnormal state. If the prediction result is an abnormal state, the corresponding early warning mechanism is triggered. If the prediction result is a normal state, no early warning prompt is issued.

[0013] A further improvement of the technical solution of the present invention is that: in the building data analysis module, a process of drawing a water leakage risk distribution map of the prefabricated building by using geographic information system technology in combination with real-time data streams and providing a visual interface for viewing the status of various parts of the building includes: Integrate and convert temperature and humidity data, total precipitation data, groundwater level data, building strain data, and construction record data into a format suitable for geographic information systems. Associate the spatial coordinate data of deployed sensors with the corresponding sensor data, and arrange the data at each time point in chronological order. Using the QGIS platform, the pre-processed data was imported into the geographic information system in CSV format. Utilizing the spatial analysis capabilities of GIS, combined with sensor data and the spatial structure of the building, a water leakage risk distribution map was generated. Thresholds for temperature, humidity, total precipitation, groundwater level, and strain were set. If any detected area exceeded the set threshold, it was marked as a high-risk area. If no detected area exceeded the set threshold, it was marked as a normal area. The three-dimensional model of the building is displayed in the geographic information system, the location of each sensor and the data it collects are marked and updated, the changes in the water leakage risk in the target building are dynamically displayed on the map, and color coding is used to divide the risk levels of different areas, with green coding indicating low risk, yellow coding indicating medium risk, and red coding indicating high risk.

[0014] A further improvement of the technical solution of the present invention is that: in the building data analysis module, the process of transmitting data using encryption technology and fault tolerance mechanism includes: The pre-processed temperature and humidity, total precipitation, groundwater level, building strain value and construction record data are integrated and converted into a format suitable for transmission. The AES symmetric encryption algorithm is used to generate a random key and encrypt the data to generate an encrypted data block. The encrypted data block is split into small data packets containing data and CRC checksum to achieve data fragmentation and verification. The data packets are transmitted via the TCP / IP protocol and checked and reassembled at the receiving end. The original data is decrypted using the key and a retransmission mechanism is designed to deal with data loss and corruption. The RESTful API interface allows external systems to query and obtain the building status data.

[0015] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art: 1. The present invention provides an abnormality identification system for water leakage monitoring in prefabricated buildings. Through an integrated building data acquisition module, it can comprehensively and in real time collect meteorological, hydrological and structural data of prefabricated buildings, providing a solid data foundation for the accurate prediction of leakage risks and significantly improving the accuracy and timeliness of risk identification.

[0016] 2. The present invention provides an abnormality identification system for monitoring water leakage in prefabricated buildings. The building risk prediction module uses an advanced autoregressive integral sliding average model and a support vector machine classification algorithm to achieve intelligent prediction of leakage risks and effective distinction between abnormal states. It can not only provide early warning of potential leakage points, but also greatly reduce the probability of false alarms and missed alarms, thereby enhancing the pertinence and efficiency of supervision work.

[0017] 3. The present invention provides an abnormality identification system for monitoring water leakage in prefabricated buildings. The building data analysis module uses geographic information system technology to intuitively display the spatial distribution of water leakage risks. The visual interface greatly facilitates management personnel to monitor and manage the status of various parts of the building. At the same time, encryption technology and fault-tolerant mechanisms are used to ensure the security and reliability of data transmission, further ensuring the continuity of supervision work and the integrity of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0019] Figure 1 A block diagram of the present invention. DETAILED DESCRIPTION

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] Example 1, as Figure 1 As shown, the present invention provides an abnormality identification system for monitoring water leakage in assembled buildings, including a data transmission guarantee center, wherein the data transmission guarantee center is communicatively connected to a building data acquisition module, a building risk prediction module, and a building data analysis module; The building data acquisition module collects meteorological data, hydrological data and structural data of prefabricated buildings through temperature and humidity sensors, rain gauges, groundwater level monitors, strain gauges and construction records. Temperature and humidity sensors are deployed in the basement, corridors and exterior walls of the building to measure the temperature and relative humidity in the environment using resistance changes. The physical signal is converted into an electrical signal using the internal circuit of the temperature and humidity sensor, and the analog signal is converted into a digital signal through an analog-to-digital converter to collect temperature and humidity data. Rain gauges are deployed on the roof of the building and in open areas on the ground. A tipping bucket design is adopted. Rainwater flows into the water collection container through the funnel. A water collection threshold is set. When the water volume reaches the water collection threshold, the tipping bucket flips to record the rainfall. A groundwater level monitor is deployed in the groundwater around the building. Pressure sensors are used to measure The pressure of the water column on the sensor is measured to determine the height of the groundwater level. Strain gauges are deployed on the beams, columns and walls of the building. Based on the resistance strain effect of metal materials, the resistance of the material changes when it is deformed by force. The strain gauge is attached to the surface of the building. When the building is deformed, the resistance value of the strain gauge changes. The resistance change is converted into a voltage signal for measurement through a bridge circuit. Key information of the construction process is recorded regularly by humans. The key information includes the location of deformation joints and post-pouring strips and the construction quality. The construction time, process, materials used and test results are recorded. The temperature threshold is set between -20°C and 50°C, and the relative humidity threshold is set between 0% and 100%. The collected temperature and humidity data are preliminarily checked, abnormal values are removed, and standardized. The standardized temperature data , standardized humidity data ,in, is the original temperature data, For relative humidity data, remove the abnormal values of multiple consecutive flipping times in the bucket flipping records, and accumulate the number of bucket flipping times to calculate the total precipitation P. , where N is the number of bucket flips, X is the precipitation corresponding to each flip, in millimeters. Remove the fluctuation abnormal value in the pressure sensor measurement value and convert the pressure value measured by the pressure sensor into the groundwater level height H. ,in is the pressure value measured by the pressure sensor, is the density of groundwater, g is the acceleration of gravity, and the fluctuation abnormal value in the strain gauge resistance change is removed, and the strain gauge resistance change is converted into the strain value. , , where k is the strain gauge sensitivity coefficient, is the building structure strain value, is the initial resistance value of the strain gauge, and the daily average temperature and relative humidity change rate are calculated as follows: ; ; in, and are the initial and final relative humidity values during the acquisition period, and The daily cumulative precipitation, hourly precipitation change rate, daily groundwater level change rate, and daily strain value change rate were calculated for the start and end time of the collection. The text description of the construction record was converted into a standard format, and key features in the construction record were extracted, including the spacing of expansion joints, the location of the post-cast strip, and the concrete strength grade. The building risk prediction module uses an autoregressive integral moving average model to predict the leakage risk of prefabricated buildings. The support vector machine classification algorithm is used to train the model to distinguish between normal and abnormal states, thereby achieving early warning of potential leakage points in prefabricated buildings. The autoregressive integral moving average model is constructed. The autoregressive integral moving average model includes an autoregressive term coefficient, a difference order, and a moving average term number. The number of autoregressive terms p is determined by the partial autocorrelation function graph, and the difference order d is determined by observing the stationarity of the time series data. If the original data is non-stationary, differential processing is performed until the data is stationary. The number of moving average terms q is determined by the autocorrelation function graph. The collected historical data is used to train the autoregressive integral moving average model. The calculation process is as follows: ; in, represents the penetration probability of the target building at time t, c is a constant term, is the white noise error term, and are the autoregressive and moving average coefficients respectively. The preprocessed temperature and humidity data, total precipitation, groundwater level and building strain value are used as the input data of the autoregressive integral moving average model. According to the output results of the autoregressive integral moving average model, the leakage risk prediction value of the target building is obtained. The standard values of temperature and relative humidity, total precipitation, groundwater level, building strain value, deformation joint spacing, post-casting strip setting location and concrete strength grade are selected from the processed data as the input features of the support vector machine classification model. 70% of the data are used as the training set and 30% of the data are used as the test set. The support vector machine classification model is trained using the training set data, the radial basis function kernel is selected, and the optimal hyperparameter combination is determined by the cross-validation method. The hyperparameters include the regularization parameter C and the kernel function parameter , the training process is as follows: ; in, is the output of the support vector machine classification model, is the Lagrange multiplier, is the sample label, 1 indicates normal state, -1 indicates abnormal state, is the radial basis function kernel, b is the bias term, and the performance of the support vector machine classification model is evaluated using the test set. The accuracy of the support vector machine classification model is calculated by comparing the prediction results of the support vector machine classification model with the actual labels. The calculation process is as follows: ; Generate a confusion matrix and analyze the classification effect of the support vector machine classification model. The confusion matrix consists of four parts: true positives, false positives, true negatives, and false negatives. Among them, true positives are the number of samples that are actually abnormal and correctly predicted to be abnormal, false positives are the number of samples that are actually normal but incorrectly predicted to be abnormal, true negatives are the number of samples that are actually normal and correctly predicted to be normal, and false negatives are the number of samples that are actually abnormal but incorrectly predicted to be normal. Use precision, recall, and F1 score as evaluation indicators to comprehensively evaluate the performance of the support vector machine classification model. Use the preprocessed temperature and humidity data, total precipitation, groundwater level, and building strain value as input data of the support vector machine classification model. According to the output results of the support vector machine classification model, obtain the target building state, which includes normal state and abnormal state. If the prediction result is an abnormal state, the corresponding early warning mechanism is triggered. If the prediction result is a normal state, no early warning prompt is issued; The building data analysis module uses geographic information system technology and real-time data streams to draw a water leakage risk distribution map for prefabricated buildings, provides a visual interface to view the status of various parts of the building, and uses encryption technology and fault-tolerant mechanisms to transmit data. It integrates and converts temperature and humidity data, total precipitation data, groundwater level data, building strain value data, and construction record data into a format suitable for the geographic information system. It associates the spatial coordinate data of the deployed sensors with the corresponding sensor data, arranges the data at each time point in chronological order, selects the QGIS platform, and imports the pre-processed data into the geographic information system in CSV format. Using the spatial analysis function of GIS, combined with the sensor collection data and the spatial structure of the building, a water leakage risk distribution map is generated. The temperature and humidity, total precipitation, groundwater level, and strain value thresholds are set. If there is a detection area that exceeds the set threshold, it is marked as a high-risk area. If there is no detection area that exceeds the set threshold, it is marked as a normal area. The 3D model of the building is displayed in the geographic information system, the location of each sensor and the data it collects are marked and updated, the changes in the water leakage risk in the target building are dynamically displayed on the map, and the risk levels of different areas are divided using color coding, with green coding indicating low risk, yellow coding indicating medium risk, and red coding indicating high risk. The pre-processed temperature and humidity, total precipitation, groundwater level, building strain value and construction record data are integrated and converted into a format suitable for transmission. The AES symmetric encryption algorithm is used to generate a random key and encrypt the data to generate an encrypted data block. The encrypted data block is split into small data packets containing data and CRC checksum to achieve data fragmentation and verification. The data packets are transmitted via the TCP / IP protocol and verified and reassembled at the receiving end. The original data is decrypted and restored using the key. A retransmission mechanism is designed to deal with data loss and damage. The RESTful API interface allows external systems to query and obtain the status data of the building.

[0022] First, the building data collection module is activated. Using temperature and humidity sensors, rain gauges, groundwater level monitors, and strain gauges, it comprehensively collects meteorological, hydrological, and structural data from prefabricated buildings. Construction records are then integrated to ensure data integrity and accuracy. The Data Transmission Assurance Center then transmits this data in real time to the building risk prediction module. This module uses an autoregressive integral moving average model to predict leakage risks and trains the model using a support vector machine classification algorithm to accurately distinguish between normal and abnormal building states, enabling early warning. Finally, the building data analysis module receives the risk prediction results and, using geographic information system technology combined with real-time data streams, creates a water leakage risk distribution map. Users can visually view the status of various building components through a visual interface, enabling them to take timely measures to address potential leakage risks. Furthermore, the system employs encryption technology and fault-tolerant mechanisms to ensure the security and reliability of data transmission.

[0023] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An abnormality identification system for monitoring water leakage in prefabricated buildings, characterized by: It includes a data transmission guarantee center, which is communicatively connected to a building data acquisition module, a building risk prediction module, and a building data analysis module, wherein the modules are connected by electrical signals; The building data acquisition module collects meteorological data, hydrological data and structural data of the prefabricated building through temperature and humidity sensors, rain gauges, groundwater level monitors, strain gauges and construction records; The building risk prediction module uses an autoregressive integral moving average model to predict the leakage risk of prefabricated buildings, and uses a support vector machine classification algorithm to train the model to distinguish between normal and abnormal states, thereby achieving early warning of potential leakage points in prefabricated buildings. The building data analysis module uses geographic information system technology and real-time data streams to draw a water leakage risk distribution map for prefabricated buildings, provides a visual interface to view the status of various parts of the building, and uses encryption technology and fault-tolerant mechanisms to transmit data.

2. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 1 is characterized by: In the building data collection module, the process of collecting meteorological data, hydrological data, and structural data of the prefabricated building through temperature and humidity sensors, rain gauges, groundwater level monitors, strain gauges, and construction records includes: Temperature and humidity sensors are deployed in the basement, corridors, and exterior walls of the building. They measure the temperature and relative humidity of the environment using changes in resistance. The internal circuits of the temperature and humidity sensors convert the physical signals into electrical signals, which are then converted into digital signals using analog-to-digital converters to collect temperature and humidity data. Rain gauges are deployed on building roofs and open areas on the ground. They use a tipping bucket design. Rainwater flows through the funnel into a water collection container. A water collection threshold is set. When the water volume reaches the collection threshold, the tipping bucket flips and records the rainfall. Deploy groundwater level monitors in the groundwater around the building, using pressure sensors to determine the groundwater level by measuring the pressure of the water column on the sensor; Strain gauges are deployed on the beams, columns, and walls of the building. Based on the resistance strain effect of metal materials, the resistance changes when the material is deformed by force. The strain gauges are attached to the surface of the building. When the building deforms, the resistance value of the strain gauge changes. The resistance change is converted into a voltage signal through a bridge circuit. Key information during the construction process is recorded manually and regularly, including the location of expansion joints and post-casting strips and the quality of construction, as well as the construction time, process, materials used and test results.

3. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 2 is characterized by: In the building data acquisition module, the process of pre-processing the collected temperature and humidity data, total precipitation, groundwater level, building strain value and construction records includes: Set the temperature threshold between -20°C and 50°C, and the relative humidity threshold between 0% and 100%, conduct a preliminary check on the collected temperature and humidity data, remove abnormal values, and perform standardization. The standardized temperature data , standardized humidity data ,in, is the original temperature data, is the relative humidity data; Remove the abnormal values of the bucket flipping time in succession in the dumping record, and accumulate the number of bucket flipping times to calculate the total precipitation P. , where N is the number of bucket flips, and X is the amount of precipitation corresponding to each flip, in millimeters; Remove the fluctuation abnormal value in the pressure sensor measurement value and convert the pressure value measured by the pressure sensor into the groundwater level height H. ,in is the pressure value measured by the pressure sensor, is the density of groundwater, g is the acceleration due to gravity; Remove the fluctuation abnormal value in the strain gauge resistance change and convert the strain gauge resistance change into strain value , , where k is the strain gauge sensitivity coefficient, is the building structure strain value, is the initial resistance value of the strain gauge.

4. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 3 is characterized by: In the building data acquisition module, the process of extracting features from the collected temperature and humidity data, total precipitation, groundwater level, building strain values, and construction records includes: Calculate the daily average temperature and relative humidity change rate. The calculation process is as follows: ； ； in, and are the initial and final relative humidity values during the acquisition period, and They are the start time and end time of the collection respectively; The daily cumulative precipitation, hourly precipitation change rate, daily groundwater level change rate, and daily strain value change rate were calculated. The text description of the construction record was converted into a standard format, and the key features in the construction record were extracted, including the spacing of expansion joints, the location of post-cast strips, and the concrete strength grade.

5. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 4 is characterized by: In the building risk prediction module, the process of training the autoregressive integrated moving average model includes: Constructing an autoregressive integrated moving average model, the autoregressive integrated moving average model includes an autoregressive term coefficient, a difference order, and a moving average term number. The number of autoregressive terms p is determined by a partial autocorrelation function graph, and the difference order d is determined by observing the stationarity of the time series data. If the original data is non-stationary, difference processing is performed until the data is stationary, and the number of moving average terms q is determined by an autocorrelation function graph. Using the collected historical data, the autoregressive integrated moving average model is trained. The calculation process is as follows: ； in, represents the penetration probability of the target building at time t, c is a constant term, is the white noise error term, and are the autoregressive and moving average coefficients, respectively.

6. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 5 is characterized by: In the building risk prediction module, the process of predicting the leakage risk of prefabricated buildings using the trained autoregressive integrated moving average model includes: The preprocessed temperature and humidity data, total precipitation, groundwater level and building strain value are used as input data of the autoregressive integral moving average model. According to the output results of the autoregressive integral moving average model, the leakage risk prediction value of the target building is obtained.

7. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 6 is characterized by: In the building risk prediction module, the process of training the support vector machine classification model includes: From the processed data, standard values of temperature and relative humidity, total precipitation, groundwater level, building strain, expansion joint spacing, post-casting strip location, and concrete strength grade were selected as input features for the support vector machine classification model. 70% of the data was used as the training set, and 30% of the data was used as the test set. The support vector machine classification model is trained using the training set data, the radial basis function kernel is selected, and the optimal hyperparameter combination is determined by the cross-validation method. The hyperparameters include the regularization parameter C and the kernel function parameter , the training process is as follows: ； in, is the output of the support vector machine classification model, is the Lagrange multiplier, is the sample label, 1 indicates normal state, -1 indicates abnormal state, is the radial basis function kernel, b is the bias term; Use the test set to evaluate the performance of the support vector machine classification model. By comparing the prediction results of the support vector machine classification model with the actual labels, the accuracy of the support vector machine classification model is calculated. Generate a confusion matrix and analyze the classification effect of the support vector machine classification model. The confusion matrix consists of four parts: true positives, false positives, true negatives, and false negatives. Among them, true positives are the number of samples that are actually abnormal and correctly predicted to be abnormal, false positives are the number of samples that are actually normal but incorrectly predicted to be abnormal, true negatives are the number of samples that are actually normal and correctly predicted to be normal, and false negatives are the number of samples that are actually abnormal but incorrectly predicted to be normal. Use precision, recall, and F1 score as evaluation indicators to comprehensively evaluate the performance of the support vector machine classification model.

8. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 7 is characterized by: In the building risk prediction module, the process of distinguishing normal and abnormal states based on the support vector machine classification model to achieve early warning of potential leakage points in prefabricated buildings includes: The preprocessed temperature and humidity data, total precipitation, groundwater level and building strain value are used as input data of the support vector machine classification model. According to the output results of the support vector machine classification model, the target building state is obtained. The target building state includes normal state and abnormal state. If the prediction result is an abnormal state, the corresponding early warning mechanism is triggered. If the prediction result is a normal state, no early warning prompt is issued.

9. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 8 is characterized by: In the building data analysis module, the process of drawing a water leakage risk distribution map for prefabricated buildings by using geographic information system technology in combination with real-time data streams and providing a visual interface to view the status of various parts of the building includes: Integrate and convert temperature and humidity data, total precipitation data, groundwater level data, building strain data, and construction record data into a format suitable for geographic information systems. Associate the spatial coordinate data of deployed sensors with the corresponding sensor data, and arrange the data at each time point in chronological order. Using the QGIS platform, the pre-processed data was imported into the geographic information system in CSV format. Utilizing the spatial analysis capabilities of GIS, combined with sensor data and the spatial structure of the building, a water leakage risk distribution map was generated. Thresholds for temperature, humidity, total precipitation, groundwater level, and strain were set. If any detected area exceeded the set threshold, it was marked as a high-risk area. If no detected area exceeded the set threshold, it was marked as a normal area. The three-dimensional model of the building is displayed in the geographic information system, the location of each sensor and the data it collects are marked and updated, the changes in the water leakage risk in the target building are dynamically displayed on the map, and color coding is used to divide the risk levels of different areas, with green coding indicating low risk, yellow coding indicating medium risk, and red coding indicating high risk.

10. The abnormality identification system for monitoring water leakage in prefabricated buildings according to claim 9, characterized in that: In the building data analysis module, the process of transmitting data using encryption technology and fault tolerance mechanism includes: The pre-processed temperature and humidity, total precipitation, groundwater level, building strain value and construction record data are integrated and converted into a format suitable for transmission. The AES symmetric encryption algorithm is used to generate a random key and encrypt the data to generate an encrypted data block. The encrypted data block is split into small data packets containing data and CRC checksum to achieve data fragmentation and verification. The data packets are transmitted via the TCP / IP protocol and checked and reassembled at the receiving end. The original data is decrypted using the key and a retransmission mechanism is designed to deal with data loss and corruption. The RESTful API interface allows external systems to query and obtain the building status data.

Images