Multivariate data acquisition and fusion analysis method and system for structural health monitoring

Abstract

The application provides a multi-element data acquisition and fusion analysis method and system for structural health monitoring, comprising: acquiring target building or structure monitoring data by using an integrated multi-element sensor and transmitting; processing the monitoring data; establishing an LSTM neural network modal parameter prediction model and an LSTM neural network correction inclination prediction model and training; determining a modal parameter sequence and a correction inclination sequence for prediction, inputting the modal parameter sequence and the correction inclination sequence into the LSTM neural network modal parameter prediction model and the LSTM neural network correction inclination prediction model respectively, obtaining a modal parameter prediction sequence and an inclination prediction sequence, and performing trend analysis and early warning. The application effectively eliminates the influence of environmental parameters on the measured structure parameters and the prediction results, and the LSTM neural network is used to train and fit the measured data of buildings or structures under different conditions, so that the nonlinear evolution characteristics of the measured structure time series data can be expressed with high precision.

Description

Technical Field

[0001] This invention relates to the field of building health and safety, and more specifically, to a method and system for collecting and fusing multi-data points for structural health monitoring. Background Technology

[0002] With the continuous increase in the number and service life of existing buildings and structures in cities, the probability of structural risks is constantly increasing. Real-time perception and early prediction of potential structural risks are effective means to reduce the operation and maintenance costs of existing buildings and structures. Improving the predictive accuracy of structural monitoring data in different scenarios is of great practical significance for preventing and reducing losses caused by structural risks.

[0003] Currently, many structural health and safety monitoring systems are deployed and used primarily for risk assessment based on single structural indicators. However, numerous studies have shown that environmental factors significantly impact monitoring data. In the raw structural response monitoring data, the structural response caused by environmental variables such as temperature may be on the same order of magnitude as the changes in structural parameters such as tilt angle. Therefore, the raw monitoring data may not reflect the true trends in structural parameter changes. Ignoring this environmental influence could lead to a decrease in the accuracy of structural safety assessment results. Therefore, it is necessary to propose an analytical method, system, and equipment to fuse and analyze multi-dimensional health monitoring data to achieve multi-indicator early warning for structural safety. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for multi-data acquisition and fusion analysis for structural health monitoring.

[0005] According to one aspect of the present invention, a method for multi-data acquisition and fusion analysis for structural health monitoring is provided, comprising:

[0006] The monitoring data of the target building or structure is acquired by using an integrated multi-sensor system, including acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0007] The acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data are transmitted.

[0008] Structural modal parameter identification is performed on the acceleration monitoring data to obtain modal parameter identification results, resulting in historical modal parameter monitoring data; the tilt angle monitoring data and temperature monitoring data are processed to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data;

[0009] An LSTM neural network modal parameter prediction model is established, and the historical monitoring data of the modal parameters is used as training samples for training; an LSTM neural network corrected tilt angle prediction model is established, and the historical monitoring data of the corrected tilt angle is used as training samples for training.

[0010] Obtain the modal parameter sequence to be predicted, input the modal parameter sequence into a trained LSTM neural network modal parameter prediction model, and output the modal parameter prediction sequence of the target building or structure; obtain the corrected tilt angle sequence to be predicted, input the corrected tilt angle sequence into a trained LSTM neural network tilt angle prediction model, and output the tilt angle prediction sequence of the target building or structure;

[0011] Trend determination and early warning are performed based on the predicted modal parameter sequence and the predicted tilt angle sequence.

[0012] Preferably, the method of acquiring monitoring data of the target building or structure using integrated multi-sensor sensors is for timed data collection using sensors;

[0013] When the sensor performs a single data acquisition, the acquired data includes acceleration time history, single or multiple tilt angle data, and single or multiple temperature data collected at a fixed sampling frequency within a fixed time period.

[0014] Preferably, the SSI method is used to identify modal parameters from the acceleration monitoring data of the target building or structure, and the identified modal parameters include: modal frequency, mode shape and modal damping;

[0015] During the identification process, a stability map is also generated, which includes a stability map with the modal frequency on the horizontal axis and the modal order on the vertical axis.

[0016] Based on the stability graph, unstable frequencies in the identified modal parameters are filtered out to obtain frequency-stable modal frequencies, mode shapes, and modal damping.

[0017] Preferably, the processing of the tilt angle monitoring data and temperature monitoring data to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data includes:

[0018] The tilt angle data and temperature data are normalized and a database is constructed to obtain historical tilt angle monitoring data and historical temperature monitoring data.

[0019] The XGBoost method was used to perform segmented component decomposition on the historical dip angle monitoring data, and the trend term in the component decomposition was obtained by linear fitting.

[0020] The historical temperature monitoring data is considered as an exogenous variable, and least squares fitting is used to obtain the exogenous term caused by the exogenous variable.

[0021] The trend term and the exogenous term are removed from the historical tilt monitoring data to obtain the corrected historical tilt monitoring data.

[0022] Preferably, the normalization process specifically includes: E[x] is the mean of the data sample, and Var[x] is the variance of the data sample.

[0023] Preferably, the training process of the LSTM neural network modal parameter prediction model and the LSTM neural network tilt angle prediction model is the same, both including the following process:

[0024] The hyperparameters of the neural network are determined based on the number of input variables;

[0025] The output variable of the LSTM neural network modal parameter prediction model is determined to be the modal parameter prediction sequence; the output variable of the LSTM neural network tilt angle prediction model is determined to be the tilt angle prediction sequence.

[0026] The LSTM neural network modal parameter prediction model / LSTM neural network tilt prediction model is trained by minimizing the difference between the predicted modal parameters / predicted tilt angle and the measured values ​​to obtain the parameters of each neuron node.

[0027] Preferably, the step of determining the trend and issuing an early warning based on the predicted modal parameter sequence and the predicted tilt angle sequence includes:

[0028] The tilt trend is obtained based on the modal parameters and the tilt angle prediction value;

[0029] Based on the set modal parameters and upper and lower thresholds for tilt angle trends, determine whether the modal parameters and tilt angle trends exceed the limits:

[0030] An early warning is issued if the tilt angle trend or at least one modal parameter exceeds the limit; or, an early warning is issued if the tilt angle trend and at least one modal parameter both exceed the limit simultaneously.

[0031] According to a second aspect of the present invention, a multi-data acquisition and fusion analysis system for structural health monitoring is provided, comprising:

[0032] The data acquisition module uses integrated multi-sensor to acquire monitoring data of the target building or structure, including acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0033] The data transmission module transmits the acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0034] The data processing module performs structural modal parameter identification on the acceleration monitoring data to obtain modal parameter identification results and obtain historical monitoring data of modal parameters; it also processes the tilt angle monitoring data and temperature monitoring data to obtain corrected historical monitoring data of tilt angle and historical monitoring data of temperature.

[0035] The prediction model training module establishes an LSTM neural network modal parameter prediction model, using historical monitoring data of the modal parameters as training samples; and establishes an LSTM neural network corrected tilt angle prediction model, using historical monitoring data of the corrected tilt angle as training samples.

[0036] The settlement prediction module obtains the modal parameter sequence to be predicted, inputs the modal parameter sequence into a trained LSTM neural network modal parameter prediction model, and outputs the modal parameter prediction sequence of the target building or structure; it also obtains the corrected tilt angle sequence to be predicted, inputs the corrected tilt angle sequence into a trained LSTM neural network tilt angle prediction model, and outputs the tilt angle prediction sequence of the target building or structure.

[0037] The early warning module determines trends and issues early warnings based on the predicted modal parameter sequence and the predicted tilt angle sequence.

[0038] According to a third aspect of the present invention, a terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, can be used to perform the method described therein, or to run the system described therein.

[0039] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, can be used to perform the method described thereon, or to run the system described thereon.

[0040] Compared with the prior art, the embodiments of the present invention have at least one of the following beneficial effects:

[0041] 1. The multi-data acquisition and fusion analysis method and system for structural health monitoring in this embodiment of the invention effectively isolates the influence of environmental parameters on the parameters of the measured structure and the prediction results by processing and correcting acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data. By using LSTM neural network to train and fit the measured data of buildings or structures under different conditions, the nonlinear evolution characteristics of the time series data of the measured structure can be expressed with high precision.

[0042] 2. The multi-data acquisition and fusion analysis method and system for structural health monitoring in this embodiment of the invention comprehensively considers the different characteristics of the target building or structure as well as the actual environmental conditions. It can accurately predict the structural health and safety index data under different site conditions. The interactive early warning information it provides can provide relevant operation and maintenance personnel with a low-cost, high-efficiency, and high-precision real-time structural safety management solution. Attached Figure Description

[0043] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0044] Figure 1 This is a schematic diagram of a multi-data acquisition and fusion analysis method for structural health monitoring according to an embodiment of the present invention;

[0045] Figure 2 This is a stability graph of a preferred embodiment of the present invention, with the horizontal axis representing the modal frequency and the vertical axis representing the modal order.

[0046] Figure 3 This is a schematic diagram of the internal structure of a neural network model according to a preferred embodiment of the present invention. Detailed Implementation

[0047] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0048] With the development of integrated small multi-element sensors and the increasing innovation and widespread application of artificial intelligence technology in the early detection of building risk sources, it is possible to simultaneously collect structural data and environmental parameters in different scenarios. Based on this, embodiments of the present invention provide a method for multi-element data collection and fusion analysis for structural health monitoring, applicable to the prediction and evaluation of the structural health and safety of different types of buildings or structures. Figure 1 The method includes:

[0049] Step 1: Use integrated multi-sensor systems to acquire monitoring data of the target building or structure. The monitoring data includes acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0050] Step two: Transmit the acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data;

[0051] Step 3: Perform structural modal parameter identification on the acceleration monitoring data to obtain the first modal parameter identification results of the target building or structure, and obtain historical monitoring data of modal parameters; process the tilt angle monitoring data and temperature monitoring data to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data;

[0052] Step 4: Establish an LSTM neural network modal parameter prediction model based on historical monitoring data of modal parameters, and train the LSTM neural network modal parameter prediction model using historical monitoring data of modal parameters as training samples; establish an LSTM neural network corrected tilt angle prediction model based on historical monitoring data of corrected tilt angle, and train the LSTM neural network tilt angle prediction model using historical monitoring data of corrected tilt angle as training samples.

[0053] Step 5: Determine the modal parameter sequence for prediction, input the modal parameter sequence into the trained LSTM neural network modal parameter prediction model, and output the modal parameter prediction sequence of the target building or structure; determine the corrected tilt angle sequence for prediction, input the corrected tilt angle sequence into the trained LSTM neural network tilt angle prediction model, and output the tilt angle prediction sequence of the target building or structure.

[0054] Step Six: Determine the trend and issue early warnings for the predicted modal parameter prediction sequence and tilt angle prediction sequence.

[0055] This invention comprehensively analyzes multi-dimensional sensing data such as historical modal parameter information, historical tilt angle information, and historical temperature information, and removes environmental influences. It uses an LSTM neural network to train and fit the measured structural parameters of buildings or structures under different conditions, such as modal parameters and tilt angle data. This can accurately express the nonlinear evolution characteristics of the time series data of the measured structural parameters, and improve the accuracy of structural parameter prediction under different scenarios.

[0056] In a preferred embodiment of the present invention, step one can acquire historical monitoring data through a data acquisition network. The data acquisition network comes from an integrated miniaturized multi-element sensor, which includes an acceleration acquisition module, an tilt acquisition module, and a temperature acquisition module to acquire the above data, and uses an integrated sensor transmission module to realize the unified timed transmission of multi-element data. Among them, the acceleration data acquired in a single acquisition consists of the acceleration time history acquired at a fixed sampling frequency within a fixed time period, and the tilt and temperature data acquired in a single acquisition consists of one or more acquired data.

[0057] In some other embodiments, the data acquisition network may also consist of different measuring devices, including accelerometers, inclinometers, thermometers, hygrometers, etc., arranged on the target building or structure. The data collected by the above-mentioned acquisition devices is simultaneously acquired and transmitted by a unified gateway and a timing module arranged on the gateway.

[0058] In a preferred embodiment of the invention, the data transmission method in step two includes a field distributed wired transmission network composed of a sensor data acquisition terminal, a distributed gateway, a public network data transmission protocol, a cloud or local data receiving port, etc., to realize data transmission of the sensor network using a wired data transmission method.

[0059] In some other embodiments, the data transmission method may also employ a field distributed wireless transmission network consisting of sensor wireless transmitting devices, wireless transmission protocols (IoT, WiFi, LoRa, ZigBee, etc.), distributed gateways, public network data transmission protocols, cloud or local data receiving ports, etc., to realize data transmission of sensor networks using wireless data transmission methods.

[0060] To achieve high-precision representation of the nonlinear evolution characteristics of the measured structure's time-series data, effective processing of the acquired data is necessary. In a preferred embodiment of this invention, the Stochastic Subspace Identification (SSI) method is used to identify structural modal parameters from the acceleration monitoring data. The identified modal parameters include: modal frequencies, mode shapes, and modal damping. Furthermore, during the SSI algorithm's modal parameter identification of the acceleration time history, a stability diagram is also generated, such as... Figure 2 As shown, the stability diagram includes a horizontal axis representing modal frequency and a vertical axis representing modal order. Based on the stability diagram, stable frequencies are identified, and unstable frequencies caused by noise interference in the original data are filtered out. This allows for the identification of the actual modes of the target building or structure, and the identification results of the modal frequencies, mode shapes, and damping of the stable frequencies are obtained. These results serve as the input data for the structural modal parameter prediction model in step four.

[0061] The acceleration monitoring data in the above embodiments were processed. Similarly, in another embodiment of the present invention, the tilt angle monitoring data and temperature monitoring data were also processed to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data, which mainly include two parts:

[0062] First, considering the differences in units among diverse data sources, it is necessary to normalize the collected multi-source data to achieve better neural network training results. Normalized data can be represented as follows: E[x ] Here, Var[x] represents the average value of the data sample, and Var[x] represents the variance of the data sample. After normalization, variables with different dimensions can be processed uniformly.

[0063] Second, the XGBoost algorithm is used to perform segmented component decomposition on the historical tilt angle monitoring data. The historical temperature monitoring data is regarded as an exogenous environmental variable, and the possible linear data drift caused by sensor hardware is also taken into account. The purpose is to remove the components related to environmental variables and linear data drift from the original monitoring data to obtain the corrected true structural tilt angle. In the component decomposition, the trend term caused by the linear data drift is fitted with linear fitting, and the exogenous term caused by the exogenous environmental variable is fitted with least squares fitting. The corrected historical tilt angle monitoring data is obtained by removing the trend term and exogenous term from the historical tilt angle monitoring data, thereby improving the model prediction accuracy.

[0064] After obtaining the processed data, these data can be used as input for model training. In a preferred embodiment of the present invention, step four involves using historical monitoring data of modal parameters as training samples to train an LSTM neural network modal parameter prediction model; and using historical monitoring data of corrected tilt angles as training samples to train an LSTM neural network tilt angle prediction model.

[0065] In a preferred embodiment, the training process of the LSTM neural network modal parameter prediction model includes the following steps:

[0066] The hyperparameters of the neural network are determined based on the number of input variables;

[0067] The output variable of the LSTM neural network modal parameter prediction model is the modal parameter prediction sequence;

[0068] The LSTM neural network modal parameter prediction model is trained by minimizing the difference between the predicted and measured modal parameters to obtain the parameters of each neuron node.

[0069] In a preferred implementation, the training process of the LSTM neural network tilt prediction model includes the following steps:

[0070] The hyperparameters of the neural network are determined based on the number of input variables;

[0071] The output variable of the LSTM neural network tilt prediction model is the tilt prediction sequence;

[0072] The LSTM neural network tilt prediction model is trained by minimizing the difference between the predicted and measured tilt angle values ​​to obtain the parameters of each neuron node.

[0073] Through the training process described in the above embodiments, a trained LSTM neural network tilt prediction model and an LSTM neural network tilt prediction model are obtained. The LSTM neural network divides the input time series data into sub-units of the same time slice for processing. Through the learning iteration of input gate vector, forget gate vector, and output gate vector, it realizes the extraction and representation of the information of the previous sub-unit.

[0074] Furthermore, such as Figure 3 The diagram shows the internal structure of each LSTM neural network prediction model. LSTM neural networks utilize their chain-like structure to transmit information sequentially over time, and are trained and learned through updates to gate vectors and hidden node states, making them suitable for processing and predicting time series data. This neural network divides the time series data in the historical data sequence into equal time slices for processing; the t-th time slice corresponds to processing the t-th input data. And consider the cell state C transmitted from the (t-1)th time slice cell. t-1 With hidden node state h t-1 And the element state C calculated for this sub-unit t With hidden node state h t Pass it to the next subunit.

[0075] For the t-th input data The hidden node state h transmitted from the (t-1)th time slice unit t-1 In this embodiment, the neural network model obtains the forgetting gate vector f through equation (1). t f represents the degree to which the information learned in the previous subunit is retained in the current subunit, with each element located in the range [0,1]. t A value of 0 represents "complete forgetting", f t A value of 1 represents "complete retention". In equation (1), W... f and b f These represent the corresponding weight coefficient matrix and bias term, respectively, with σ being the sigmoid activation function. Meanwhile, in this embodiment, the neural network model obtains the input gate vector i through equations (2) and (3). t and cell status update value Where the input gate vector i t Each element is within the range [0,1] and is used to control the control system state update value. Update the state of this sub-unit C t Retention rate at time, i t A value of 0 represents "no update at all", i t 1 represents "full update". In equation (2), W... i and b i These represent the corresponding weight coefficient matrix and bias term, respectively, with σ being the sigmoid activation function. In equation (3), W...C and b C These are the corresponding weight coefficient matrix and bias term, respectively.

[0076]

[0077] Input data passes through the forgetting gate vector f t With input gate vector i t Afterwards, valuable historical information will be retained, and worthless information will be removed. The current sub-unit state is then combined with the updated state using equation (4) to complete the unit state C. t Update.

[0078]

[0079] To obtain the predicted value And complete the hidden node state h t The update is performed by obtaining the output gate vector o through equations (5) and (6). t and the hidden node state h of this sub-unit t Output gate vector o t Each element is located within the range [0,1] and is used to control the hidden node state h. t Update level, o t A value of 0 represents "no update at all". t 1 represents "full update". In equation (5), W o and b o These represent the corresponding weight coefficient matrix and bias term, respectively, with σ being the sigmoid activation function.

[0080]

[0081] h t =o t *tanh(C t (6)

[0082] Using the LSTM structure and data flow described above, the modal parameters and tilt angle prediction of the LSTM neural network prediction model are realized.

[0083] In a preferred embodiment of the present invention, in step five, when performing modal parameter prediction, a modal parameter sequence for prediction is determined, and the aforementioned data sequence is input as input data into a trained LSTM neural network model, outputting predicted modal parameter values ​​for the monitored target building or structure. When performing tilt angle prediction, a corrected tilt angle sequence for prediction is determined, and the aforementioned data sequence is input as input data into a trained LSTM neural network model, outputting predicted tilt angle values ​​for the monitored target building or structure.

[0084] After obtaining the predicted values ​​of modal parameters and tilt angle, step six is ​​further implemented in this embodiment to determine the trend and generate early warning information. Specifically, the modal parameter and tilt angle early warning processing is integrated into the cloud or local server where the data analysis is located. Based on the predicted modal parameters and tilt angle, a slope is obtained through linear fitting, and this slope is taken as the tilt angle trend. When the predicted values ​​of the target building or structure obtained by the above prediction methods exceed the settlement threshold set by the system, or when the slope obtained by the linear fitting of modal parameters and tilt angle increases significantly, it is considered that the monitored settlement has deteriorated, and early warning information is sent to relevant operation and maintenance management personnel via SMS, email, etc. The above early warning information can be managed and warned in a hierarchical manner according to different thresholds, thereby realizing fully automatic monitoring and early warning of settlement. For example, an alarm is triggered when the tilt angle trend or at least one modal parameter exceeds the threshold, or when the tilt angle trend and at least one modal parameter exceed the threshold simultaneously, realizing data fusion analysis and reducing the probability of false alarms.

[0085] The above embodiments of the present invention comprehensively consider the different characteristics of the target building or structure and the actual environmental conditions, and can accurately predict the structural health and safety index data under different site conditions. The interactive information it provides can provide relevant operation and maintenance personnel with a low-cost, high-efficiency, and high-precision real-time structural safety management solution.

[0086] Based on the same inventive concept, other embodiments of the present invention also provide a multivariate data fusion analysis system for structural health monitoring, comprising:

[0087] The data acquisition module uses integrated multi-sensor to acquire monitoring data of the target building or structure, including acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0088] The data transmission module transmits the acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data.

[0089] The data processing module performs structural modal parameter identification on the acceleration monitoring data, obtains the first modal parameter identification result of the target building or structure, and obtains historical monitoring data of modal parameters; it also processes the tilt angle monitoring data and temperature monitoring data to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data.

[0090] The prediction model training module is used to establish an LSTM neural network modal parameter prediction model, using the historical monitoring data of the modal parameters as training samples to train the LSTM neural network modal parameter prediction model; and to establish an LSTM neural network tilt angle prediction model, using the historical monitoring data of the corrected tilt angle as training samples to train the LSTM neural network tilt angle prediction model.

[0091] The settlement prediction module determines the modal parameter sequence for prediction, inputs the modal parameter sequence into a trained LSTM neural network modal parameter prediction model, and outputs the modal parameter prediction sequence of the target building or structure; it also determines the corrected tilt angle sequence for prediction, inputs the corrected tilt angle sequence into a trained LSTM neural network tilt angle prediction model, and outputs the tilt angle prediction sequence of the target building or structure.

[0092] The early warning module performs trend determination and early warning on the predicted modal parameter prediction sequence and tilt angle prediction sequence.

[0093] The specific implementation techniques of each module / unit in the above examples of the present invention can be referred to the steps of the multi-data acquisition and fusion analysis method for structural health monitoring in the above embodiments, and will not be repeated here.

[0094] Based on the same inventive concept, in other embodiments of the present invention, a terminal is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it can be used to perform the above-described method or to run the above-described system.

[0095] Optionally, the memory is used to store programs; the memory may include volatile memory, such as random-access memory (RAM), such as static random-access memory (SRAM), double data rate synchronous dynamic random-access memory (DDR SDRAM), etc.; the memory may also include non-volatile memory, such as flash memory. The memory is used to store computer programs (such as application programs, functional modules, etc. that implement the above methods), computer instructions, etc., and the above-mentioned computer programs, computer instructions, etc., may be partitioned and stored in one or more memories.

[0096] A processor is used to execute a computer program stored in memory to implement the various steps of the methods involved in the above embodiments. For details, please refer to the relevant descriptions in the preceding method embodiments.

[0097] The processor and memory can be separate structures or integrated structures. When the processor and memory are separate structures, they can be coupled together via a bus.

[0098] Based on the same inventive concept, in other embodiments of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, can be used to perform the above-described method or to run the above-described system.

[0099] Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of computer programs from one place to another. Storage media can be any available medium accessible to a general-purpose or special-purpose computer. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in a user device. Of course, the processor and storage medium can also exist as discrete components in a communication device.

[0100] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0103] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0104] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for multi-data acquisition and fusion analysis for structural health monitoring, characterized in that, include: The monitoring data of the target building or structure is acquired by using an integrated multi-sensor system, including acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data. The acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data are transmitted. Structural modal parameter identification is performed on the acceleration monitoring data to obtain modal parameter identification results, resulting in historical modal parameter monitoring data; the tilt angle monitoring data and temperature monitoring data are processed to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data; An LSTM neural network modal parameter prediction model is established, and the historical monitoring data of the modal parameters is used as training samples for training; an LSTM neural network corrected tilt angle prediction model is established, and the historical monitoring data of the corrected tilt angle is used as training samples for training. Obtain the modal parameter sequence to be predicted, input the modal parameter sequence into a trained LSTM neural network modal parameter prediction model, and output the modal parameter prediction sequence of the target building or structure; obtain the corrected tilt angle sequence to be predicted, input the corrected tilt angle sequence into a trained LSTM neural network tilt angle prediction model, and output the tilt angle prediction sequence of the target building or structure; Trend determination and early warning are performed based on the predicted modal parameter sequence and the predicted tilt angle sequence. The method of using integrated multi-sensor to acquire monitoring data of target buildings or structures is for timed data collection using sensors. When the sensor performs a single data acquisition, the acquired data includes acceleration time history, single or multiple tilt angle data, and single or multiple temperature data acquired at a fixed sampling frequency within a fixed time period; The SSI method is used to identify modal parameters from the acceleration monitoring data of the target building or structure, including modal frequency, mode shape, and modal damping. During the identification process, a stability map is also generated, which includes a stability map with the modal frequency on the horizontal axis and the modal order on the vertical axis. Based on the stability graph, unstable frequencies in the identified modal parameters are filtered out to obtain frequency-stable modal frequencies, mode shapes, and modal damping.

2. The method for multi-data acquisition and fusion analysis for structural health monitoring according to claim 1, characterized in that, The process of processing the tilt angle monitoring data and temperature monitoring data to obtain corrected tilt angle historical monitoring data and temperature historical monitoring data includes: The tilt angle data and temperature data are normalized and a database is constructed to obtain historical tilt angle monitoring data and historical temperature monitoring data. The XGBoost method was used to perform segmented component decomposition on the historical dip angle monitoring data, and the trend term in the component decomposition was obtained by linear fitting. The historical temperature monitoring data is considered as an exogenous variable, and least squares fitting is used to obtain the exogenous term caused by the exogenous variable. The trend term and the exogenous term are removed from the historical tilt monitoring data to obtain the corrected historical tilt monitoring data.

3. The method for multi-data acquisition and fusion analysis for structural health monitoring according to claim 2, characterized in that, The normalization process specifically includes: , The average value of the data sample. This represents the variance of the data samples.

4. The method for multi-data acquisition and fusion analysis for structural health monitoring according to claim 1, characterized in that, The training process for the LSTM neural network modal parameter prediction model and the LSTM neural network tilt angle prediction model is the same, both including the following process: The hyperparameters of the neural network are determined based on the number of input variables; The output variable of the LSTM neural network modal parameter prediction model is determined to be the modal parameter prediction sequence; the output variable of the LSTM neural network tilt angle prediction model is determined to be the tilt angle prediction sequence. The LSTM neural network modal parameter prediction model / LSTM neural network tilt prediction model is trained by minimizing the difference between the predicted modal parameters / predicted tilt angle and the measured values ​​to obtain the parameters of each neuron node.

5. The method for multi-data acquisition and fusion analysis for structural health monitoring according to claim 1, characterized in that, The trend determination and early warning based on the predicted modal parameter sequence and the predicted tilt angle sequence include: The tilt trend is obtained based on the modal parameters and the tilt angle prediction value; Based on the set modal parameters and upper and lower thresholds for tilt angle trends, determine whether the modal parameters and tilt angle trends exceed the limits: An early warning is issued if the tilt angle trend or at least one modal parameter exceeds the limit; or, an early warning is issued if both the tilt angle trend and at least one modal parameter exceed the limit simultaneously.

6. A multi-data acquisition and fusion analysis system for structural health monitoring, characterized in that, include: The data acquisition module uses integrated multi-sensor to acquire monitoring data of the target building or structure, including acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data. The data transmission module transmits the acceleration monitoring data, tilt angle monitoring data, and temperature monitoring data. The data processing module performs structural modal parameter identification on the acceleration monitoring data to obtain modal parameter identification results and obtain historical monitoring data of modal parameters; it also processes the tilt angle monitoring data and temperature monitoring data to obtain corrected historical monitoring data of tilt angle and historical monitoring data of temperature. The prediction model training module establishes an LSTM neural network modal parameter prediction model, using historical monitoring data of the modal parameters as training samples; and establishes an LSTM neural network corrected tilt angle prediction model, using historical monitoring data of the corrected tilt angle as training samples. The settlement prediction module obtains the modal parameter sequence to be predicted, inputs the modal parameter sequence into a trained LSTM neural network modal parameter prediction model, and outputs the modal parameter prediction sequence of the target building or structure; it also obtains the corrected tilt angle sequence to be predicted, inputs the corrected tilt angle sequence into a trained LSTM neural network tilt angle prediction model, and outputs the tilt angle prediction sequence of the target building or structure. The early warning module determines trends and issues early warnings based on the predicted modal parameter sequence and the predicted tilt angle sequence. The method of using integrated multi-sensor to acquire monitoring data of target buildings or structures is for timed data collection using sensors. When the sensor performs a single data acquisition, the acquired data includes acceleration time history, single or multiple tilt angle data, and single or multiple temperature data acquired at a fixed sampling frequency within a fixed time period; The SSI method is used to identify modal parameters from the acceleration monitoring data of the target building or structure, including modal frequency, mode shape, and modal damping. During the identification process, a stability map is also generated, which includes a stability map with the modal frequency on the horizontal axis and the modal order on the vertical axis. Based on the stability graph, unstable frequencies in the identified modal parameters are filtered out to obtain frequency-stable modal frequencies, mode shapes, and modal damping.

7. A terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it can be used to perform the method of any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, this program can be used to perform the method of any one of claims 1-5.

Images