Strength prediction system for prefabricated box girder transfer station
By combining the precast box girder's own parameters with the environmental parameters of the transfer station, a multi-factor fusion prediction system has been developed, which solves the problems of low prediction accuracy and poor versatility in existing technologies. It enables real-time, accurate, and multi-form early warning of the strength of precast box girders, adapts to the needs of box girders of different specifications, and ensures the safety of transfer stations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ZHENGZHOU ENG CO LTD CHINA RAILWAY SEVENTH GRP
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-31
AI Technical Summary
Existing precast box girder strength prediction systems rely on a single detection index, without considering environmental factors at transfer stations and load impacts during transportation. This results in low prediction accuracy, an inability to achieve real-time data acquisition and dynamic updates, a lack of anomaly warning functions, poor versatility, and an inability to adapt to the needs of box girders of different specifications.
It employs a sensing and acquisition module, a data transmission module, an edge computing module, a cloud processing module, and an interactive display module. Combining the parameters of the precast box girder itself, the environmental parameters of the transfer station, and the load parameters, it uses an adaptive prediction model that integrates random forest and BP neural network. Through 5G+LoRa dual-mode transmission, it achieves real-time data acquisition and processing, and has the functions of dynamic prediction and three-level early warning with multi-factor fusion.
It achieves accurate, real-time dynamic prediction of the strength of precast box girders, improving prediction accuracy by more than 30%, adapting to the needs of box girders of different specifications, and has multiple forms of early warning functions to ensure the safe management and control of transfer stations.
Smart Images
Figure CN122492400A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of precast box girder maintenance and testing technology, specifically a strength prediction system for a precast box girder transfer station. Background Technology
[0002] As a core load-bearing component in bridge engineering, the strength of precast box girders directly determines the safety and durability of the bridge structure. After the precast box girders are produced, they need to be transported to a transfer station for temporary storage and curing until their strength meets the design requirements before being transported to the construction site for installation. Therefore, monitoring and predicting the strength of box girders at the transfer station stage is a crucial step in ensuring construction safety and preventing damage or collapse of box girders due to insufficient strength.
[0003] Currently, existing precast box girder strength prediction technologies suffer from the following main drawbacks: First, existing prediction systems largely rely on single detection indicators (such as concrete age and surface temperature), failing to consider environmental factors at transfer stations (such as temperature, humidity, and wind speed) and dynamic influencing factors like load impacts during transport, resulting in low prediction accuracy and significant errors. Second, existing systems are mostly offline prediction modes, unable to achieve real-time acquisition, transmission, and dynamic updating of strength data, making it difficult to meet the efficient management and control requirements of batch box girders at transfer stations. Third, existing systems lack anomaly warning functions; when the strength of the box girder increases abnormally or falls below the safety threshold, they cannot issue timely warning signals, easily leading to safety hazards. Fourth, the prediction models of existing systems are fixed and cannot be adaptively adjusted according to different types and specifications of precast box girders, resulting in poor versatility.
[0004] Furthermore, existing technical solutions mostly focus on predicting the strength of concrete or steel structures alone, failing to achieve dynamic prediction by integrating multiple factors. They also lack integration with the actual operating conditions of transfer stations (such as the number of stacking layers and transfer routes), thus failing to fully adapt to the actual application scenarios of transfer stations. Therefore, there is an urgent need for a strength prediction system for precast box girder transfer stations that can overcome the aforementioned technical deficiencies and achieve accurate, real-time, dynamic, and highly versatile predictions to fill the gap in existing technology and ensure the operational safety of precast box girder transfer stations. Summary of the Invention
[0005] The purpose of this invention is to overcome the above-mentioned defects of the prior art and provide a strength prediction system for precast box girder transfer stations. This system enables real-time strength prediction of precast box girders during storage and transportation at transfer stations by integrating multiple factors, improving prediction accuracy, providing anomaly warning function, adapting to the prediction needs of box girders of different specifications, and providing reliable data support for the safety management of transfer stations.
[0006] To achieve the above objectives, the present invention proposes the following technical solution: a strength prediction system for a precast box girder transfer station, characterized in that it includes a sensing and acquisition module, a data transmission module, an edge computing module, a cloud processing module, an early warning module, and an interactive display module, with each module electrically connected in sequence to form a closed loop of signal acquisition-transmission-processing-early warning-display; The sensing and acquisition module is used to collect parameters of the precast box girder itself, environmental parameters of the transfer station, and load parameters. The parameters of the precast box girder itself include concrete mix proportion, initial strength, curing age, and steel structure connection strength. The environmental parameters of the transfer station include ambient temperature, humidity, wind speed, and light intensity. The load parameters include stacking load, transfer impact load, and support reaction force. The data transmission module adopts a 5G+LoRa dual-mode transmission method to realize real-time data transmission between the sensing and acquisition module and the edge computing module, and between the edge computing module and the cloud processing module; The edge computing module has a built-in preprocessing unit and an adaptive prediction model. The preprocessing unit is used to perform noise reduction, anomaly removal, and normalization on the collected raw data. The adaptive prediction model is built based on the fusion of random forest algorithm and BP neural network, and can automatically adjust the model parameters according to the specifications and type of precast box girder. The cloud processing module is used to store historical strength data, environmental data and prediction results, build a precast box girder strength database, and perform secondary verification of prediction results and optimize prediction model parameters. The early warning module has a built-in safety threshold setting unit, which compares the real-time predicted strength of the box girder with the preset safety threshold and issues an early warning signal of the corresponding level according to the degree of abnormality. The interactive display module is used to display various collected parameters, predicted intensity data, and early warning information in real time, and supports parameter input, historical data query, and data export.
[0007] Furthermore, the sensing and acquisition module includes a concrete strength sensor, a steel structure stress sensor, an environmental sensor, and a load sensor; the concrete strength sensor is embedded inside the precast box girder, the steel structure stress sensor is installed at the steel structure connection of the box girder, the environmental sensor is installed in various areas of the transfer station, and the load sensor is installed on the box girder stacking supports and transfer equipment.
[0008] Furthermore, the data transmission module has a built-in data caching unit that can cache data for no less than 72 hours when the network is interrupted, and automatically synchronize the data after the network is restored.
[0009] Furthermore, the process of building an adaptive prediction model includes the following steps: S1: Collect historical strength data, intrinsic parameters, environmental parameters, and load parameters of precast box girders of different specifications and types to construct a sample dataset; S2: Preprocess the sample dataset by removing outliers and missing values, normalizing the data, and dividing it into training and test sets. S3: Selecting core features based on the random forest algorithm; S4: Input the core features into the BP neural network, build the initial prediction model and train it; S5: An adaptive adjustment mechanism is introduced to automatically optimize model parameters based on the box girder specifications and type, resulting in an adaptive prediction model; S6: Validate the model using the test set, and put it into use once the error meets the standard.
[0010] Furthermore, the cloud processing module has a built-in data encryption unit and a data backup unit. The data encryption unit uses the AES encryption algorithm, and the data backup unit automatically backs up the data once a day.
[0011] Furthermore, the warning module has three warning levels: Level 1 warning issues an audible and visual warning, Level 2 warning issues an audible and visual warning and an SMS warning, and Level 3 warning issues an audible and visual warning, an SMS warning and an emergency shutdown signal.
[0012] Furthermore, the interactive display module uses an industrial touch screen, which supports manual input of parameters and querying of historical data.
[0013] Furthermore, the sensing and acquisition module also includes a vibration sensor, which is installed on the web of the precast box girder to collect the vibration frequency and amplitude during the transport of the box girder, as a supplement to the load parameters.
[0014] Furthermore, the cloud processing module also has a built-in data statistical analysis unit, which can statistically analyze the strength growth patterns of different batches of box girders and generate statistical reports.
[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention combines the parameters of the precast box girder itself, the environmental parameters of the transfer station, and the load parameters, and adopts an adaptive prediction model that integrates random forest and BP neural network to achieve dynamic prediction of multiple factors, effectively reducing prediction errors. Compared with the existing single-index prediction system, the prediction accuracy is improved by more than 30%.
[0016] 2. The 5G+LoRa dual-mode transmission method is adopted to realize real-time data acquisition, transmission and processing. The edge computing module processes the data locally, reducing data transmission latency and enabling real-time updates of box girder strength prediction results, meeting the efficient management and control requirements of batch box girders at the transfer station.
[0017] 3. The adaptive prediction model can automatically adjust parameters according to different specifications and types of precast box girders without the need for manual modification of model settings. It can adapt to the strength prediction needs of various precast box girders and has a wide range of applications.
[0018] 4. It has a three-level early warning function, which can promptly detect abnormal intensity and load, issue multiple forms of early warning signals, and even control the shutdown of transfer equipment to effectively avoid safety accidents; at the same time, cloud data is encrypted and stored to ensure data security and prevent data leakage and tampering. Attached Figure Description
[0019] Figure 1 This is a system diagram of the present invention. Detailed Implementation
[0020] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example 1
[0021] like Figure 1 As shown: A strength prediction system for a precast box girder transfer station, characterized in that: it includes a sensing and acquisition module, a data transmission module, an edge computing module, a cloud processing module, an early warning module, and an interactive display module, and each module is electrically connected in sequence to form a closed loop of signal acquisition-transmission-processing-early warning-display; The sensing and acquisition module is used to collect parameters of the precast box girder itself, environmental parameters of the transfer station, and load parameters. The parameters of the precast box girder itself include concrete mix proportion, initial strength, curing age, and steel structure connection strength. The environmental parameters of the transfer station include ambient temperature, humidity, wind speed, and light intensity. The load parameters include stacking load, transfer impact load, and support reaction force. The sensing and acquisition module includes a concrete strength sensor, a steel structure stress sensor, an environmental sensor, and a load sensor. The concrete strength sensor is a fiber optic type, embedded in key load-bearing components inside the precast box girder at 1.5m intervals. It is used to collect real-time data on the compressive and tensile strength of the concrete, with a measurement range of 0-100MPa and an accuracy of ±0.5MPa. The steel structure stress sensor is a strain gauge type, attached to the bolt connections and welds of the box girder steel structure to collect stress changes in the steel structure, indirectly reflecting the connection strength. The environmental sensor uses… Integrated temperature, humidity, and wind speed sensors are installed in the stacking area and above the transfer channel of the transfer station, with one sensor installed every 500 square meters. These sensors are used to collect ambient temperature (measurement range -20℃~80℃), humidity (measurement range 0~100%RH), wind speed (measurement range 0~30m / s), and light intensity (measurement range 0~10000lux). The load sensors are pressure-type load sensors, installed on the top of the supports of the box girder stacking and on the bearing surface of the transfer vehicles. These sensors are used to collect stacking load (measurement range 0~500t), transfer impact load (measurement range 0~100t), and support reaction force.
[0022] The data transmission module adopts a 5G+LoRa dual-mode transmission method to realize real-time data transmission between the sensing and acquisition module and the edge computing module, and between the edge computing module and the cloud processing module; The data transmission module adopts a 5G+LoRa dual-mode transmission module. 5G transmission is used to transmit data with high real-time requirements (such as impact load and real-time intensity), with a transmission latency of ≤100ms. LoRa transmission is used to transmit batch environmental data and historical data, with a transmission distance of ≥1000m. The dual-mode transmission serves as a backup for each other, avoiding signal interruption in a single transmission mode and ensuring stable data transmission. The data transmission module also has a built-in data caching unit, which can cache data for no less than 72 hours when the network is interrupted, and automatically synchronize it to the edge computing module after the network is restored.
[0023] The edge computing module has a built-in preprocessing unit and an adaptive prediction model. The preprocessing unit is used to perform noise reduction, anomaly removal, and normalization on the collected raw data. The adaptive prediction model is built based on the fusion of random forest algorithm and BP neural network, and can automatically adjust the model parameters according to the specifications and type of precast box girder. The process of building an adaptive prediction model includes the following steps: S1: Collect historical strength data, intrinsic parameters, environmental parameters, and load parameters of precast box girders of different specifications and types to construct a sample dataset; S2: Preprocess the sample dataset by removing outliers and missing values, normalizing the data, and dividing it into training and test sets. S3: Selecting core features based on the random forest algorithm; S4: Input the core features into the BP neural network, build the initial prediction model and train it; S5: An adaptive adjustment mechanism is introduced to automatically optimize model parameters based on the box girder specifications and type, resulting in an adaptive prediction model; S6: Validate the model using the test set, and put it into use once the error meets the standard.
[0024] The edge computing module adopts an industrial-grade edge gateway, with a built-in preprocessing unit and adaptive prediction model. The preprocessing unit is written in Python and uses an outlier detection algorithm (3σ criterion) to remove outliers from the collected data, linear interpolation to supplement missing values, and Min-Max normalization to normalize the data to the [0,1] interval, improving data quality. The adaptive prediction model is based on the fusion of random forest algorithm and BP neural network. The random forest algorithm is used for feature selection, selecting concrete mix ratio, ambient temperature, stacking load, and curing age as core features (cumulative contribution rate ≥85%). The BP neural network adopts a 3-layer structure (input layer, hidden layer, output layer). The input layer has 4 core features, the hidden layer has 12 nodes, and the output layer is the real-time strength of the box girder. The model automatically adjusts the number of hidden layer nodes and network weights according to the input box girder specifications (such as box girder length and cross-sectional dimensions) and type (such as prestressed box girder and ordinary box girder) through an adaptive adjustment mechanism to achieve accurate prediction of different box girders.
[0025] The cloud processing module is used to store historical strength data, environmental data and prediction results, build a precast box girder strength database, and perform secondary verification of prediction results and optimize prediction model parameters. The cloud processing module uses a cloud server with a built-in MySQL database to store raw data collected by the sensing and acquisition module, prediction results transmitted by the edge computing module, and historical intensity data. The database supports millions of data storages and has a data backup function, automatically backing up once a day. The cloud processing module also has a built-in model optimization unit, which optimizes the parameters of the adaptive prediction model by comparing the prediction results with the measured intensity data, using the gradient descent method, and optimizes once every 7 days to continuously improve prediction accuracy. At the same time, it has a built-in AES encryption unit to encrypt the transmitted and stored data to prevent data leakage and tampering.
[0026] The early warning module has a built-in safety threshold setting unit, which compares the real-time predicted strength of the box girder with the preset safety threshold and issues an early warning signal of the corresponding level according to the degree of abnormality. The early warning module includes an audible and visual alarm and an SMS warning module, with a built-in safety threshold setting unit. Staff can set the strength safety threshold for different specifications of box girders (e.g., the safety strength threshold for prestressed box girders is ≥35MPa) via an interactive display module. The early warning levels are divided into three levels: Level 1 (minor anomaly): When the predicted strength is less than 5% below the safety threshold or the strength growth rate is less than 0.5MPa / day, the audible and visual alarm emits a yellow light and a warning sound (1 time / second); Level 2 (moderate anomaly): When the predicted strength is 5%-10% below the safety threshold or the strength growth rate is less than 0.3MPa / day, the audible and visual alarm emits an orange light and a warning sound (2 times / second), and simultaneously sends an SMS warning to the staff's mobile phone; Level 3 (serious anomaly): When the predicted strength is more than 10% below the safety threshold or the load exceeds the bearing capacity by 20%, the audible and visual alarm emits a red light and a rapid warning sound (3 times / second), sends an SMS warning to the staff's mobile phone, and outputs an emergency stop signal to control the transfer equipment to stop operating.
[0027] The interactive display module is used to display various collected parameters, predicted intensity data, and early warning information in real time, and supports parameter input, historical data query, and data export.
[0028] The interactive display module uses an industrial touch screen (15 inches) installed in the monitoring room of the transfer station. It displays various collected parameters (temperature, humidity, load, strength), predicted strength data and early warning information in real time. It supports staff to manually input parameters such as box girder specifications and safety thresholds, query historical data and strength prediction curves (trend graphs), and also supports data export function for subsequent analysis. Example 2
[0029] The difference between this embodiment and Embodiment 1 is that the sensing and acquisition module also includes a vibration sensor, which is installed on the web of the precast box girder to collect the vibration frequency and amplitude during the transport of the box girder, as a supplement to the load parameters.
[0030] The sensing and acquisition module has added vibration sensors, which are installed on the web of the precast box girder to collect the vibration frequency and amplitude of the box girder during transportation. The vibration parameters are used as a supplement to the load parameters and input into the adaptive prediction model to further improve the prediction accuracy. At the same time, the cloud processing module has added a data statistical analysis unit, which can automatically count the strength growth patterns of different batches of box girders and generate statistical reports to provide data support for the optimization of maintenance plans at the transfer station.
[0031] The workflow of this invention is as follows: 1. System initialization: Staff input parameters such as the specifications, type, concrete mix ratio, and initial strength of the precast box girder through the interactive display module, and set the strength safety threshold. 2. Data Acquisition: Various sensors in the sensing and acquisition module collect the box girder's own parameters, environmental parameters, and load parameters in real time, and transmit the collected raw data to the data transmission module; 3. Data transmission and preprocessing: The data transmission module adopts a 5G+LoRa dual-mode approach to transmit the raw data to the edge computing module in real time. The preprocessing unit performs noise reduction, anomaly removal, and normalization on the raw data. 4. Strength Prediction: The adaptive prediction model of the edge computing module analyzes the preprocessed data, extracts core features, predicts the real-time strength of the box girder, and transmits the prediction results to the cloud processing module and interactive display module; 5. Data Validation and Optimization: The cloud processing module performs secondary validation on the prediction results, optimizes the parameters of the adaptive prediction model by combining historical data, and stores relevant data. 6. Early Warning and Display: The early warning module compares the predicted intensity with the safety threshold and issues corresponding warnings based on the degree of anomaly; the interactive display module displays various data and early warning information in real time for staff to view and operate.
[0032] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0033] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.
Claims
1. A strength prediction system for a precast box girder transfer station, characterized in that: It includes a sensing and acquisition module, a data transmission module, an edge computing module, a cloud processing module, an early warning module, and an interactive display module. These modules are electrically connected in sequence to form a closed loop of signal acquisition, transmission, processing, early warning, and display. The sensing and acquisition module is used to collect parameters of the precast box girder itself, environmental parameters of the transfer station, and load parameters. The parameters of the precast box girder itself include concrete mix proportion, initial strength, curing age, and steel structure connection strength. The environmental parameters of the transfer station include ambient temperature, humidity, wind speed, and light intensity. The load parameters include stacking load, transfer impact load, and support reaction force. The data transmission module adopts a 5G+LoRa dual-mode transmission method to realize real-time data transmission between the sensing and acquisition module and the edge computing module, and between the edge computing module and the cloud processing module; The edge computing module has a built-in preprocessing unit and an adaptive prediction model. The preprocessing unit is used to perform noise reduction, anomaly removal, and normalization on the collected raw data. The adaptive prediction model is built based on the fusion of random forest algorithm and BP neural network, and can automatically adjust the model parameters according to the specifications and type of precast box girder. The cloud processing module is used to store historical strength data, environmental data and prediction results, build a precast box girder strength database, and perform secondary verification of prediction results and optimize prediction model parameters. The early warning module has a built-in safety threshold setting unit, which is used to compare the real-time predicted box girder strength with the preset safety threshold and issue an early warning signal of the corresponding level according to the degree of abnormality. The interactive display module is used to display various collected parameters, predicted intensity data, and early warning information in real time, and supports parameter input, historical data query, and data export.
2. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The sensing and acquisition module includes a concrete strength sensor, a steel structure stress sensor, an environmental sensor, and a load sensor. The concrete strength sensor is embedded inside the precast box girder, the steel structure stress sensor is installed at the steel structure connection of the box girder, the environmental sensor is installed in various areas of the transfer station, and the load sensor is installed on the box girder stacking supports and transfer equipment.
3. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The data transmission module has a built-in data caching unit that can cache data for no less than 72 hours when the network is interrupted, and automatically synchronize the data after the network is restored.
4. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The process of constructing the adaptive prediction model includes the following steps: S1: Collect historical strength data, intrinsic parameters, environmental parameters, and load parameters of precast box girders of different specifications and types to construct a sample dataset; S2: Preprocess the sample dataset by removing outliers and missing values, normalizing the data, and dividing it into training and test sets. S3: Core feature selection based on random forest algorithm; S4: Input the core features into the BP neural network, build the initial prediction model and train it; S5: An adaptive adjustment mechanism is introduced to automatically optimize model parameters based on the box girder specifications and type, resulting in an adaptive prediction model; S6: Validate the model using the test set, and put it into use once the error meets the standard.
5. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The cloud processing module has a built-in data encryption unit and a data backup unit. The data encryption unit uses the AES encryption algorithm, and the data backup unit automatically backs up the data once a day.
6. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The warning module has three warning levels: Level 1 warning issues an audible and visual warning, Level 2 warning issues an audible and visual warning and an SMS warning, and Level 3 warning issues an audible and visual warning, an SMS warning and an emergency shutdown signal.
7. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The interactive display module uses an industrial touch screen, which supports manual input of parameters and querying of historical data.
8. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The sensing and acquisition module also includes a vibration sensor, which is installed on the web of the precast box girder to collect the vibration frequency and amplitude during the transport of the box girder, as a supplement to the load parameters.
9. The strength prediction system for a precast box girder transfer station according to claim 1, characterized in that: The cloud processing module also has a built-in data statistical analysis unit, which can statistically analyze the strength growth pattern of different batches of box girders and generate statistical reports.