A cloud-computing-based building bearing capacity detection data processing method and system

Abstract

The application relates to the technical field of building engineering and discloses a building bearing capacity detection data processing method and system based on cloud computing, which realizes real-time collection of data related to bearing capacity in a building structure through an Internet of Things sensor, carries out alignment and feature extraction on the collected data by adopting a dynamic time warping algorithm to obtain sensor original data, realizes real-time transmission of the sensor original data to a cloud computing platform, constructs a real-time data stream processing engine based on Apache Flink, carries out preprocessing on the sensor original data by adopting an edge computing technology to obtain preprocessed data, inputs the preprocessed data into a model to predict a bearing capacity change trend, dynamically adjusts a threshold value in combination with a sliding window algorithm, and outputs a prediction result, generates a risk atlas based on the prediction result, marks a potential abnormal area, and records data through a blockchain technology by the cloud computing platform; and the data processing time is greatly shortened, and the work efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of building engineering technology, and specifically to a cloud computing-based method and system for processing building load-bearing capacity testing data. Background Technology

[0002] In construction engineering, accurately assessing a building's load-bearing capacity is crucial for ensuring its safety and stability. Traditional methods for processing building load-bearing capacity testing data have many limitations, such as low data processing efficiency, insufficient analytical accuracy, and inability to effectively handle large-scale data. With the continuous expansion of building scale and the increasing complexity of building structures, higher demands are placed on the accuracy, efficiency, and real-time performance of building load-bearing capacity testing data processing. At the same time, the large amount of testing data requires powerful computing and storage capabilities, which poses a significant challenge to traditional data processing methods. Summary of the Invention

[0003] The purpose of this invention is to solve the above problems by designing a cloud computing-based method and system for processing building load-bearing capacity detection data.

[0004] The first aspect of this invention provides a cloud computing-based method for processing building load-bearing capacity testing data, the method comprising the following steps:

[0005] Real-time data related to the load-bearing capacity of the building structure are collected through IoT sensors. The collected data is aligned and features are extracted using a dynamic time warping algorithm to obtain the raw sensor data.

[0006] The raw sensor data is transmitted to the cloud computing platform in real time. A real-time data stream processing engine is built based on Apache Flink. Edge computing technology is used to preprocess the raw sensor data to obtain preprocessed data.

[0007] The preprocessed data is input into the LSTM-Transformer model to predict the trend of bearing capacity change, and the threshold is dynamically adjusted by combining the sliding window algorithm to output the prediction results.

[0008] Risk maps are generated based on the prediction results, and potentially abnormal areas are marked. The cloud computing platform records the data using blockchain technology.

[0009] Optionally, in a first implementation of the first aspect of the present invention, the step of collecting load-bearing capacity-related data in the building structure in real time through IoT sensors, and aligning and extracting features from the collected data using a dynamic time warping algorithm to obtain raw sensor data includes:

[0010] IoT sensors collect data related to the load-bearing capacity of the building structure in real time according to a preset sampling frequency, forming initial time series data;

[0011] Align time series data collected by different sensors, find the optimal matching path by calculating the distance between two time series, and achieve data alignment.

[0012] Feature extraction is performed based on the aligned time series data to extract features that reflect the building's load-bearing capacity characteristics from the data, thus obtaining the original sensor data.

[0013] Optionally, in a second implementation of the first aspect of the present invention, the step of transmitting the raw sensor data to a cloud computing platform in real time, building a real-time data stream processing engine based on Apache Flink, and using edge computing technology to preprocess the raw sensor data to obtain preprocessed data includes:

[0014] The raw sensor data collected is transmitted in real time to the receiving node of the cloud computing platform via a wireless network;

[0015] Deploy the Apache Flink framework on a cloud computing platform and build a real-time data stream processing engine based on the Apache Flink framework;

[0016] Start the edge computing node, use wavelet transform algorithm for denoising in the edge computing node, and use density-based spatial clustering algorithm to identify abnormal data points to obtain preprocessed data;

[0017] Edge computing nodes transmit pre-processed data to the cloud computing platform via the network, where the pre-processed data is integrated and stored by a real-time data stream processing engine based on Apache Flink.

[0018] Optionally, in a third implementation of the first aspect of the present invention, the step of activating the edge computing node, performing denoising processing using a wavelet transform algorithm in the edge computing node, and identifying abnormal data points using a density-based spatial clustering algorithm to obtain preprocessed data includes:

[0019] The original sensor data was decomposed using the db4 wavelet to obtain wavelet coefficients of different frequencies. The parts of the high-frequency coefficients whose absolute values ​​are less than the threshold were set to zero. The processed wavelet coefficients were then reconstructed into denoised signal data through inverse wavelet transform.

[0020] Traverse each denoised signal data point, calculate the number of other data points contained in the neighborhood radius of the current data point, and mark the current data point as the core point if the number is greater than or equal to the minimum number of points.

[0021] All data points within the neighborhood of the core point, as well as other data points reachable through the density of the core point, are grouped into the same cluster.

[0022] Data points not included in any cluster are identified as outliers, and linear interpolation is performed based on the normal data trends before and after the outliers to obtain preprocessed data.

[0023] Optionally, in the fourth implementation of the first aspect of the present invention, the step of inputting the preprocessed data into the LSTM-Transformer model to predict the trend of bearing capacity change, and dynamically adjusting the threshold in combination with the sliding window algorithm to output the prediction result includes:

[0024] The preprocessed data is input into the LSTM-Transformer model. The LSTM module filters the information that needs to be retained in the current data through the input gate and stores it in the cell state. The forget gate removes the historical information that is no longer needed in the cell state. Then, the output gate generates the output of the current time based on the updated cell state to process the temporal features in the data.

[0025] The output of the LSTM module is fed into the encoder of the Transformer module. The encoder calculates the association weights of each data position with all other positions through a multi-head self-attention mechanism. The association weights are multiplied by the features of the corresponding positions and then summed to obtain a feature representation containing global association information.

[0026] The decoder in the Transformer module receives global features from the encoder output and generates a predicted sequence of building load-bearing capacity change trends through an attention mechanism.

[0027] By dynamically adjusting the threshold using the sliding window algorithm, the predicted sequence of bearing capacity change trend output by the Transformer module is compared point by point with the dynamically adjusted threshold after the sliding window to determine whether each predicted point is within the corresponding threshold range.

[0028] Based on the comparison results, the predicted sequences are labeled. Predicted points within the threshold range are labeled as normal, and predicted points outside the threshold range are labeled as abnormal. A prediction result containing normal and abnormal labels is generated and output.

[0029] Optionally, in the fifth implementation of the first aspect of the present invention, a sliding window algorithm is initiated, and the window size and sliding step size are set according to the time interval of the data and the changing characteristics of the building's bearing capacity.

[0030] The preprocessed data is traversed using a set sliding window. Each time the window moves by one step, the data within the current window is captured. The mean, standard deviation, maximum and minimum values ​​of the window data are calculated to obtain statistical indicators. The threshold is dynamically adjusted based on the statistical indicators of each window.

[0031] Optionally, in a sixth implementation of the first aspect of the present invention, the step of generating a risk map based on the prediction results and marking potentially abnormal areas, with the cloud computing platform recording data through blockchain technology, includes:

[0032] Risk maps are generated based on the prediction results output by the LSTM-Transformer model. Different colors are used on the map to represent different levels of risk according to the predicted trend of bearing capacity change and risk level.

[0033] Based on the portion of the prediction results that exceeds the normal range, determine the location and extent of potential abnormal areas and clearly mark them on the map;

[0034] The cloud computing platform encrypts the data that needs to be recorded, generates blocks, verifies the blocks through a consensus mechanism, and adds the blocks to the blockchain after successful verification.

[0035] A second aspect of the present invention provides a cloud computing-based building load-bearing capacity testing data processing system, the system comprising:

[0036] The data acquisition module is used to collect data related to the load-bearing capacity of the building structure in real time through IoT sensors. The dynamic time warping algorithm is used to align and extract features from the collected data to obtain the raw sensor data.

[0037] The preprocessing module is used to transmit raw sensor data to the cloud computing platform in real time. It is based on Apache Flink to build a real-time data stream processing engine and uses edge computing technology to preprocess the raw sensor data to obtain preprocessed data.

[0038] The prediction module is used to input the preprocessed data into the LSTM-Transformer model to predict the trend of bearing capacity change, and dynamically adjust the threshold by combining the sliding window algorithm to output the prediction results.

[0039] The generation module is used to generate risk maps based on the prediction results and mark potential abnormal areas. The cloud computing platform records the data through blockchain technology.

[0040] A third aspect of the present invention provides a cloud-based building load-bearing capacity testing data processing device, the cloud-based building load-bearing capacity testing data processing device comprising a memory and at least one processor, the memory storing instructions; the at least one processor calling the instructions in the memory to cause the cloud-based building load-bearing capacity testing data processing device to perform the various steps of the cloud-based building load-bearing capacity testing data processing method as described in any of the preceding claims.

[0041] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the steps of the cloud computing-based building load-bearing capacity detection data processing method as described in any of the preceding claims.

[0042] The technical solution provided by this invention involves real-time collection of load-bearing capacity-related data from building structures using IoT sensors. A dynamic time warping algorithm is employed to align and extract features from the collected data, yielding raw sensor data. This raw sensor data is then transmitted in real-time to a cloud computing platform. A real-time data stream processing engine is built based on Apache Flink, and edge computing technology is used to preprocess the raw sensor data, resulting in preprocessed data. The preprocessed data is then input into an LSTM-Transformer model to predict load-bearing capacity trends. A sliding window algorithm is used to dynamically adjust thresholds, outputting the prediction results. A risk map is generated based on the prediction results, and potentially abnormal areas are marked. The cloud computing platform records the data using blockchain technology. This invention, through cloud computing processing and storage, can rapidly process large-scale building load-bearing capacity detection data. Compared to traditional methods, data processing time is significantly reduced, improving work efficiency. The LSTM-Transformer model effectively improves data accuracy and prediction precision, providing a more reliable basis for the safety assessment of building projects. Attached Figure Description

[0043] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention.

[0044] Figure 1 A flowchart illustrating a cloud-based building load-bearing capacity detection data processing method provided in an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of the structure of a cloud computing-based building load-bearing capacity detection data processing system provided in an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of the structure of a cloud computing-based building load-bearing capacity detection data processing device provided in an embodiment of the present invention. Detailed Implementation

[0047] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0048] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 A flowchart of a cloud-based building load-bearing capacity detection data processing method provided in this embodiment of the invention is shown. The method specifically includes the following steps:

[0049] Step 101: Collect data related to the load-bearing capacity of the building structure in real time through IoT sensors, and use the dynamic time warping algorithm to align and extract features from the collected data to obtain the original sensor data;

[0050] In this embodiment, based on the type and scale of the building structure and the specific requirements for load-bearing capacity detection, suitable IoT sensors are selected, such as high-precision pressure sensors, displacement sensors, and strain sensors. These sensors have the characteristics of real-time sensing, low power consumption, and anti-interference, and can work stably in complex building environments. Subsequently, the sensors are deployed at key load-bearing parts of the building structure, including but not limited to foundation caps, frame beam-column joints, shear walls, and floor load-bearing areas, ensuring that the sensors are in close contact with the surface of the building structure to accurately capture the mechanical response of the structure under different loads. The sensors collect load-bearing capacity-related data in real time according to a preset sampling frequency. These data cover various types such as pressure values, displacement amounts, strain degrees, and vibration frequencies of various parts of the structure. Due to differences in the response speed and force transmission paths of different sensors, the timestamps of the data collected by each sensor may not be completely synchronized, forming multiple sets of raw sensing data with time-series characteristics.

[0051] In this embodiment, IoT sensors collect data related to the load-bearing capacity of the building structure in real time according to a preset sampling frequency to form initial time series data; the time series data collected by different sensors are aligned, and the optimal matching path is found by calculating the distance between the two time series to achieve data alignment; feature extraction is performed based on the aligned time series to extract features that can reflect the load-bearing capacity characteristics of the building from the data to obtain the original sensor data.

[0052] In this embodiment, an IoT sensor network is deployed. Adapted pressure sensors, displacement sensors, and strain sensors are installed at key load-bearing components of the building structure, such as beams, columns, foundations, and floor slabs. This ensures close contact between the sensors and the building structure, enabling accurate capture of changes in physical quantities related to load-bearing capacity. The sensors collect data in real time according to a preset sampling frequency, forming initial time-series data. The time-series data collected by each sensor is then collected. Since these data come from sensors in different locations, the timestamps of the collected data may not be completely synchronized due to differences in sensor response speed and installation environment. Therefore, this data needs to be initially organized according to sensor number and collection time to form a final data set. This process involves processing multiple time series. A dynamic time warping algorithm is applied for data alignment. First, one time series is selected as the reference series, and the remaining series are used as the series to be aligned. Then, a distance matrix is ​​constructed, where each element represents the Euclidean distance between a data point in the series to be aligned and a data point in the reference series. Next, starting from the top left corner of the distance matrix, an optimal path is found from the top left corner to the bottom right corner. The two data points corresponding to the elements on the path are the alignment points. The path selection follows local constraints to ensure that the path does not deviate excessively from the diagonal. Through this optimal path, the series to be aligned and the reference series are matched in the time dimension, achieving data alignment.

[0053] After data alignment is completed, feature extraction is performed on the aligned time series. First, the statistical characteristics of the data are analyzed, and statistical measures such as the mean, variance, peak, and trough of the series are calculated. These statistical measures can reflect the overall distribution and fluctuation of the data. Then, the trend characteristics of the series are observed. The rate of change of the series in different time periods is calculated by using a sliding window to identify the rising, falling, or stable trends of the data. At the same time, the periodic characteristics of the series are extracted, and methods such as Fourier transform are used to analyze whether there are periodic fluctuations and their cycles in the data. These extracted features are integrated to form the raw sensor data that can characterize information related to building bearing capacity.

[0054] Step 102: Transmit the raw sensor data to the cloud computing platform in real time, build a real-time data stream processing engine based on Apache Flink, and use edge computing technology to preprocess the raw sensor data to obtain preprocessed data;

[0055] In this embodiment, the raw sensor data collected is transmitted in real time to the receiving node of the cloud computing platform via a wireless network. First, each sensor is configured with a suitable wireless communication module. These modules support wireless network standards such as LPWAN and 5G and can automatically select the optimal transmission path based on the signal strength of the building environment. Before data transmission, the raw sensor data is packaged. Each data packet contains a unique sensor identifier, a data acquisition timestamp, the specific measurement value, and a data checksum. The data checksum is used by the receiving node to verify whether the data has been corrupted during transmission. The wireless communication module pushes the data packets to nearby wireless base stations or gateways at set transmission intervals. The base stations or gateways then forward the packets to the receiving node of the cloud computing platform via the backbone network. Upon receiving the data packets, the receiving node first verifies the checksum to ensure the data is complete and error-free before storing it in a temporary buffer.

[0056] Deploying the Apache Flink framework on a cloud computing platform and building a real-time data stream processing engine based on it requires, first, planning the size of the Flink cluster according to the hardware resources of the cloud computing platform (such as the number of servers, CPU cores, and memory capacity), and determining the number and configuration of JobManagers and TaskManagers in the cluster. Next, the necessary dependencies for Flink operation are installed on each node of the cluster, including the Java development environment and the Hadoop Distributed File System (HDFS), and the cluster's communication ports, memory allocation strategies, and fault tolerance parameters are set by modifying the Flink configuration files. Then, the Flink cluster is started. The JobManager, as the master node, is responsible for receiving and scheduling data processing tasks, while the TaskManagers, as worker nodes, are responsible for executing the specific computation tasks. When building a real-time data stream processing engine based on this cluster, the data input source is first defined, and the sensor data in the receiving node buffer is input into the engine in the form of a data stream. Then, the processing operator chain is designed according to the data processing requirements, including data parsing, format conversion, and stream filtering operators. Finally, the data output sink is configured to guide the processed data stream to the subsequent storage or computing modules. At the same time, the fault tolerance guarantee of data stream processing is realized through Flink's checkpoint mechanism to ensure that the engine can quickly recover the processing state when a node failure occurs.

[0057] The edge computing node is started, and wavelet transform algorithm is used for denoising in the edge computing node. Density-based spatial clustering algorithm is used to identify abnormal data points to obtain preprocessed data. The edge computing node transmits the preprocessed data to the cloud computing platform through the network, where the real-time data stream processing engine based on Apache Flink integrates and stores the preprocessed data.

[0058] The raw sensor data is decomposed using the db4 wavelet to obtain wavelet coefficients of different frequencies. For high-frequency coefficients, those with absolute values ​​less than a threshold are set to zero. The processed wavelet coefficients are then reconstructed into denoised signal data through inverse wavelet transform. For each denoised signal data point, the number of other data points within its neighborhood radius is calculated. If this number is greater than or equal to the minimum number of points, the current data point is marked as a core point. All data points within the core point's neighborhood, as well as other data points reachable through the core point's density, are grouped into the same cluster. Data points not included in any cluster are identified as outliers, and linear interpolation is performed based on the normal data trend before and after the outlier data points to obtain the preprocessed data.

[0059] Step 103: Input the preprocessed data into the LSTM-Transformer model to predict the trend of bearing capacity change, and dynamically adjust the threshold using the sliding window algorithm to output the prediction results;

[0060] In this embodiment, the preprocessed data is input into the LSTM-Transformer model. The LSTM module filters the information that needs to be retained from the current data through the input gate and stores it in the cell state. The forget gate removes the historical information that is no longer needed from the cell state. Then, the output gate generates the output of the current time based on the updated cell state to process the temporal features in the data. The output of the LSTM module enters the encoder of the Transformer module. The encoder uses a multi-head self-attention mechanism to calculate the association weight of each data position with all other positions. The association weights are multiplied by the features of the corresponding positions and then summed to obtain a feature representation containing global association information. The decoder of the Transformer module receives the global features output by the encoder and generates a predicted sequence of building bearing capacity change trends through the attention mechanism.

[0061] The threshold is dynamically adjusted by combining the sliding window algorithm. The sliding window algorithm is started, and the window size and sliding step size are set according to the time interval of the data and the changing characteristics of the building's bearing capacity. The preprocessed data is traversed by sliding the set sliding window. At each step, the data in the current window is captured, and the mean, standard deviation, maximum and minimum values ​​of the window data are calculated to obtain statistical indicators. The threshold is dynamically adjusted according to the statistical indicators of each window.

[0062] The load-bearing capacity change trend prediction sequence output by the Transformer module is compared point by point with the dynamic threshold adjusted by the sliding window to determine whether each prediction point is within the corresponding threshold range. Based on the comparison results, the prediction sequence is marked. Prediction points within the threshold range are marked as normal, and prediction points outside the threshold range are marked as abnormal. A prediction result containing normal and abnormal labels is generated and output.

[0063] Step 104: Generate a risk map based on the prediction results and mark potential abnormal areas. The cloud computing platform records the data through blockchain technology.

[0064] In this embodiment, a risk map is generated based on the prediction results output by the LSTM-Transformer model. First, the three-dimensional model of the building structure is associated and mapped with the predicted bearing capacity change trend data to determine the correspondence between the spatial coordinate system of the map and the actual building structure. According to the numerical range and rate of change of the bearing capacity in the prediction results, multiple risk levels are divided. For example, areas where the bearing capacity is above 80% of the safety threshold and changes steadily are defined as low risk and represented by green; areas where the bearing capacity is between 50% and 80% of the safety threshold and shows a slow downward trend are defined as medium risk and represented by yellow; and areas where the bearing capacity is below 50% of the safety threshold or has a rapid rate of decline are defined as high risk and represented by red. When generating the map, according to the above color division rules, areas of different risk levels are filled and rendered on the three-dimensional model. At the same time, risk level descriptions and numerical scales corresponding to the colors are added to the edges of the map to make the risk distribution intuitively visible.

[0065] Based on the portion of the prediction results that exceeds the normal range, the location and extent of potential abnormal areas are determined. First, data points whose bearing capacity values ​​exceed the upper or lower limit of the safety threshold and whose trends do not conform to normal patterns are screened from the prediction results. Combining the sensor deployment locations corresponding to these abnormal data points, the corresponding physical areas are located in the 3D model of the building structure. By calculating the spatial distance and distribution density between adjacent abnormal data points, the boundary range of the abnormal area is determined. For discretely distributed abnormal points, a buffer analysis method is used to include the area within a certain range around them in the abnormal area. When marking on the map, a conspicuous black border is used to outline the contour of the abnormal area, and a special symbol such as an exclamation mark icon is added to the center of the area. At the same time, the legend of the map explains the meaning of the mark to ensure that the abnormal area can be quickly identified.

[0066] The cloud computing platform encrypts the data that needs to be recorded, generates blocks, and verifies the blocks through a consensus mechanism. Once verified, the block is added to the blockchain. First, the data content to be recorded is defined, including raw sensor data, preprocessing results, prediction results, and risk map generation parameters. This data is then encrypted using an asymmetric encryption algorithm, i.e., using a private key to encrypt the data, generating encrypted ciphertext. Simultaneously, the corresponding public key is stored in the blockchain nodes for subsequent decryption. When generating a block, the encrypted ciphertext, the timestamp of data generation, the hash value of the previous block, and the metadata of this block, such as the block number and data size, are integrated to calculate the hash value of this block and add it to the block header. Through a consensus mechanism in the blockchain network, such as proof-of-work or proof-of-stake, multiple nodes in the network verify the legality of the block. Verification includes the correctness of the hash value, the integrity of the data, and whether it conforms to preset rules. When more than half of the nodes have verified it, the block is officially added to the end of the blockchain, becoming an immutable and permanent record.

[0067] Please see Figure 2 A schematic diagram of the structure of a cloud-based building load-bearing capacity detection data processing system provided in this embodiment of the invention. The system includes:

[0068] The data acquisition module is used to collect data related to the load-bearing capacity of the building structure in real time through IoT sensors. The dynamic time warping algorithm is used to align and extract features from the collected data to obtain the raw sensor data.

[0069] The preprocessing module is used to transmit raw sensor data to the cloud computing platform in real time. It is based on Apache Flink to build a real-time data stream processing engine and uses edge computing technology to preprocess the raw sensor data to obtain preprocessed data.

[0070] The prediction module is used to input the preprocessed data into the LSTM-Transformer model to predict the trend of bearing capacity change, and dynamically adjust the threshold by combining the sliding window algorithm to output the prediction results.

[0071] The generation module is used to generate risk maps based on the prediction results and mark potential abnormal areas. The cloud computing platform records the data through blockchain technology.

[0072] Figure 3This is a schematic diagram of a cloud-based building load-bearing capacity testing data processing device 300 provided in an embodiment of the present invention. The cloud-based building load-bearing capacity testing data processing device 300 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 310 (e.g., one or more processors) and a memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) for storing application programs 333 or data 332. The memory 320 and storage media 330 can be temporary or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the cloud-based building load-bearing capacity testing data processing device 300. Furthermore, the processor 310 may be configured to communicate with the storage media 330 and execute the series of instruction operations in the storage media 330 on the cloud-based building load-bearing capacity testing data processing device 300 to implement the method provided in the above embodiment.

[0073] The cloud-based building load-bearing capacity testing data processing device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The structure of the cloud-based building load-bearing capacity detection data processing device shown does not constitute a limitation on the computer device provided by the present invention. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0074] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the various steps of the cloud computing-based building load-bearing capacity detection data processing method provided in the above embodiments.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described equipment or apparatus / unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0076] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0077] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A cloud computing-based method for processing building load-bearing capacity testing data, characterized in that, The method includes the following steps: Real-time data related to the load-bearing capacity of the building structure are collected through IoT sensors. The collected data is aligned and features are extracted using a dynamic time warping algorithm to obtain the raw sensor data. The raw sensor data is transmitted to the cloud computing platform in real time. A real-time data stream processing engine is built based on Apache Flink. Edge computing technology is used to preprocess the raw sensor data to obtain preprocessed data. The preprocessed data is input into the LSTM-Transformer model to predict the trend of bearing capacity change, and the threshold is dynamically adjusted by combining the sliding window algorithm to output the prediction results. Risk maps are generated based on the prediction results, and potential abnormal areas are marked. The cloud computing platform records the data through blockchain technology. The process of inputting preprocessed data into an LSTM-Transformer model to predict the trend of bearing capacity changes, and dynamically adjusting the threshold using a sliding window algorithm to output the prediction results includes: The preprocessed data is input into the LSTM-Transformer model. The LSTM module filters the information that needs to be retained in the current data through the input gate and stores it in the cell state. The forget gate removes the historical information that is no longer needed in the cell state. Then, the output gate generates the output of the current time based on the updated cell state to process the temporal features in the data. The output of the LSTM module is fed into the encoder of the Transformer module. The encoder calculates the association weights of each data position with all other positions through a multi-head self-attention mechanism. The association weights are multiplied by the features of the corresponding positions and then summed to obtain a feature representation containing global association information. The decoder in the Transformer module receives global features from the encoder output and generates a predicted sequence of building load-bearing capacity change trends through an attention mechanism. By dynamically adjusting the threshold using the sliding window algorithm, the predicted sequence of bearing capacity change trend output by the Transformer module is compared point by point with the dynamically adjusted threshold after the sliding window to determine whether each predicted point is within the corresponding threshold range. Based on the comparison results, the predicted sequence is marked. Predicted points within the threshold range are marked as normal, and predicted points outside the threshold range are marked as abnormal. A prediction result containing normal and abnormal labels is generated and output. The sliding window algorithm is activated, and the window size and sliding step size are set according to the time interval of the data and the changing characteristics of the building's load-bearing capacity. The preprocessed data is traversed using a set sliding window. Each time the window moves by one step, the data within the current window is captured. The mean, standard deviation, maximum and minimum values ​​of the window data are calculated to obtain statistical indicators. The threshold is dynamically adjusted based on the statistical indicators of each window.

2. The cloud computing-based building load-bearing capacity detection data processing method as described in claim 1, characterized in that, The process involves real-time collection of load-bearing capacity-related data from the building structure via IoT sensors. A dynamic time warping algorithm is then used to align and extract features from the collected data to obtain raw sensor data, including: IoT sensors collect data related to the load-bearing capacity of the building structure in real time according to a preset sampling frequency, forming initial time series data; Align time series data collected by different sensors, find the optimal matching path by calculating the distance between two time series, and achieve data alignment. Feature extraction is performed based on the aligned time series data to extract features that reflect the building's load-bearing capacity characteristics from the data, thus obtaining the original sensor data.

3. The cloud computing-based building load-bearing capacity detection data processing method as described in claim 1, characterized in that, The process involves transmitting raw sensor data to a cloud computing platform in real time, building a real-time data stream processing engine based on Apache Flink, and using edge computing technology to preprocess the raw sensor data to obtain preprocessed data, including: The raw sensor data collected is transmitted in real time to the receiving node of the cloud computing platform via a wireless network; Deploy the Apache Flink framework on a cloud computing platform and build a real-time data stream processing engine based on the Apache Flink framework; Start the edge computing node, use wavelet transform algorithm for denoising in the edge computing node, and use density-based spatial clustering algorithm to identify abnormal data points to obtain preprocessed data; Edge computing nodes transmit pre-processed data to the cloud computing platform via the network, where the pre-processed data is integrated and stored by a real-time data stream processing engine based on Apache Flink.

4. The cloud computing-based building load-bearing capacity testing data processing method as described in claim 3, characterized in that, The step involves activating an edge computing node, where wavelet transform is used for denoising, and density-based spatial clustering is employed to identify outlier data points, resulting in preprocessed data, including: The original sensor data was decomposed using the db4 wavelet to obtain wavelet coefficients of different frequencies. The parts of the high-frequency coefficients whose absolute values ​​are less than the threshold were set to zero. The processed wavelet coefficients were then reconstructed into denoised signal data through inverse wavelet transform. Traverse each denoised signal data point, calculate the number of other data points contained in the neighborhood radius of the current data point, and mark the current data point as the core point if the number is greater than or equal to the minimum number of points. All data points within the neighborhood of the core point, as well as other data points reachable through the density of the core point, are grouped into the same cluster. Data points not included in any cluster are identified as outliers, and linear interpolation is performed based on the normal data trends before and after the outliers to obtain preprocessed data.

5. The cloud computing-based building load-bearing capacity testing data processing method as described in claim 1, characterized in that, The process involves generating a risk map based on the prediction results and marking potentially anomalous areas. The cloud computing platform records the data using blockchain technology, including: Risk maps are generated based on the prediction results output by the LSTM-Transformer model. Different colors are used on the map to represent different levels of risk according to the predicted trend of bearing capacity change and risk level. Based on the portion of the prediction results that exceeds the normal range, determine the location and extent of potential abnormal areas and clearly mark them on the map; The cloud computing platform encrypts the data that needs to be recorded, generates blocks, verifies the blocks through a consensus mechanism, and adds the blocks to the blockchain after successful verification.

6. A cloud-based building load-bearing capacity testing data processing system, used to implement the cloud-based building load-bearing capacity testing data processing method as described in claim 1, characterized in that, The system includes: The data acquisition module is used to collect data related to the load-bearing capacity of the building structure in real time through IoT sensors. The dynamic time warping algorithm is used to align and extract features from the collected data to obtain the raw sensor data. The preprocessing module is used to transmit raw sensor data to the cloud computing platform in real time. It is based on Apache Flink to build a real-time data stream processing engine and uses edge computing technology to preprocess the raw sensor data to obtain preprocessed data. The prediction module is used to input the preprocessed data into the LSTM-Transformer model to predict the trend of bearing capacity change, and dynamically adjust the threshold by combining the sliding window algorithm to output the prediction results. The generation module is used to generate risk maps based on the prediction results and mark potential abnormal areas. The cloud computing platform records the data through blockchain technology.

7. A cloud-based building load-bearing capacity testing data processing device, characterized in that, The cloud-based building load-bearing capacity testing data processing device includes a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the cloud-based building load-bearing capacity testing data processing device to perform each step of the cloud-based building load-bearing capacity testing data processing method as described in any one of claims 1-5.

8. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement each step of the cloud computing-based building load-bearing capacity detection data processing method as described in any one of claims 1-5.

Images