Building vibration detection model training method and device, equipment, medium and product

Abstract

The invention relates to a training method and device for a vibration detection model of a building, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining a training data set, and performing optimization training on a vibration detection model based on sample vibration data and label vibration data included in the training data set to obtain target vibration detection data. The method can improve the reliability of the vibration detection model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a building vibration detection model training method and device, computer equipment, computer readable storage medium and computer program product. BACKGROUND

[0002] With the continuous development of transportation technology, various public transportation modes have emerged, among which the subway is gradually loved by the public because of its convenience, speed, and the advantage of basically no congestion and delay. The subway is built underground, and the corresponding ground surface may have various buildings. The subway may produce vibration during operation, which may affect the safety of the buildings.

[0003] In the prior art, there is a way of using a model for vibration detection, but the reliability of the trained model is poor. SUMMARY

[0004] Therefore, it is necessary to provide a building vibration detection model training method, device, computer equipment, computer readable storage medium and computer program product capable of improving the reliability of the trained vibration detection model to solve the above technical problems.

[0005] In a first aspect, the present application provides a building vibration detection model training method, comprising: obtaining a training data set, the training data set comprising sample vibration data and label vibration data, the sample vibration data being obtained by a monitoring device arranged at a fixed position of the building, and the label vibration data being obtained by a mobile monitoring device; training an initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model; and optimizing and training the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model after optimization and training.

[0006] In one embodiment, the sample vibration data acquisition process comprises: obtaining a plurality of first vibration time history data collected by a plurality of monitoring devices; the fixed position comprises a vibration source area, a transmission area and a resonance area of the building; and each first vibration time history data corresponding to each monitoring device is sorted according to time sequence to obtain the sample vibration data.

[0007] In one embodiment, the label vibration data acquisition process comprises: obtaining a plurality of second vibration time history data collected by a plurality of mobile monitoring devices according to a preset time period; and determining the label vibration data of the target building according to each second vibration time history data.

[0008] In one of the embodiments, the second vibration time series data collected by the plurality of mobile monitoring devices is acquired according to a preset time period, including: generating a collection message for the label vibration data and sending the collection message to the terminal in a case where it is detected that the model optimization period is expired; and receiving the second vibration time series data sent by the terminal.

[0009] In one of the embodiments, the candidate vibration detection model is optimized and trained based on the label vibration data to obtain a target vibration detection model of which the optimization and training are completed, including: inputting the sample vibration data into the candidate vibration detection model to obtain vibration detection data of the target building output by the candidate vibration detection model; calculating an error of the vibration detection data and the label vibration data, and adjusting parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0010] In one of the embodiments, the application process of the target vibration detection model includes: acquiring vibration sequence data collected by the monitoring device, inputting the vibration sequence data into the target vibration detection model to obtain vibration index data of the building output by the target vibration detection model; detecting whether the vibration index data is within a preset index threshold, and determining that the vibration state of the building is a safe state if the vibration index data is within the preset index threshold.

[0011] In a second aspect, the application further provides a training device of a vibration detection model of a building, including: a training data set acquisition module configured to acquire a training data set, the training data set including sample vibration data and label vibration data, the sample vibration data being obtained by a monitoring device arranged at a fixed position of the building, and the label vibration data being obtained by a mobile monitoring device; a first model training module configured to train an initial vibration detection model by using the sample vibration data to obtain a candidate vibration detection model; and a second model training module configured to optimize and train the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model of which the optimization and training are completed.

[0012] In a third aspect, the application further provides a computer device including a memory and a processor, the memory storing a computer program, and the processor implementing steps of the method of the first aspect when executing the computer program.

[0013] In a fourth aspect, the application further provides a computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement steps of the method of the first aspect.

[0014] In a fifth aspect, the application further provides a computer program product including a computer program, the computer program being executed by a processor to implement steps of the method of the first aspect.

[0015] The training method, device, computer equipment, computer readable storage medium and computer program product of the vibration detection model of the building, obtain a training data set, the training data set includes sample vibration data collected by a monitoring device arranged at a fixed position of the building, and label vibration data of the building collected by a mobile monitoring device, the vibration detection model is trained by the sample vibration data and the label vibration data, the initial vibration detection model is trained by the sample vibration data, the candidate vibration detection model is obtained, and the candidate vibration detection model is optimized and trained based on the label vibration data, and the target vibration detection model of the optimization training is obtained. The model is periodically optimized and trained by the mobile monitoring device, the problem of inaccurate parameters caused by training of the model based on short-term vibration data is avoided, and the reliability of the vibration detection model is improved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without creating labor.

[0017] Figure 1 The application environment diagram of the vibration detection model training method of the building in an embodiment;

[0018] Figure 2 The flowchart of the vibration detection model training method of the building in an embodiment;

[0019] Figure 3 The flowchart of the sample vibration data acquisition step in an embodiment;

[0020] Figure 4 The flowchart of the label vibration data acquisition step in an embodiment;

[0021] Figure 5 The flowchart of step 401 in an embodiment;

[0022] Figure 6 The flowchart of step 203 in an embodiment;

[0023] Figure 7 The flowchart of the application step of the target vibration detection model in an embodiment;

[0024] Figure 8 The flowchart of the vibration detection model training method of the building in another embodiment;

[0025] Figure 9 a structural block diagram of a training device of a vibration detection model of a building in an embodiment;

[0026] Figure 10 an internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0027] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not intended to limit the present application.

[0028] It should be noted that the terms "first", "second", etc. used in the present application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "include" and "have" and any variations thereof used in the present application are intended to cover non-exclusive inclusion. The term "multiple" used in the present application refers to two and more than two. The term "and / or" used in the present application refers to one of the options or any combination of multiple options.

[0029] The training method of the vibration detection model of the building provided by the embodiments of the present application can be applied in the application environment as shown in the figure. The application environment at least includes a model training device 101, a fixed monitoring device 102 and a mobile monitoring device 103. Figure 1

[0030] The model training device 101 is used to obtain sample vibration data collected by the fixed monitoring device 102 and obtain label vibration data collected by the mobile monitoring device 103, and to optimize and train the vibration detection model based on the sample vibration data and the label vibration data to obtain a target vibration detection model. The model training device 101 can be a server, which can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0031] The fixed monitoring device 102 is used to be installed at a fixed position of the building and to collect vibration data at the position, so that the model training device 101 obtains the vibration data as sample vibration data. The fixed monitoring device 102 can be a low-frequency acceleration sensor.

[0032] The mobile monitoring device 103 is used to collect overall vibration data of the building, so that the model training device 101 obtains the vibration data as label vibration data. The mobile monitoring device 103 can include a collection instrument and a vibration pickup.

[0033] ​In real-world scenarios, vibration testing of buildings above subway stations is crucial for assessing structural health and ensuring safe operation. However, existing vibration testing technologies face several challenges: Firstly, deploying sensors throughout the entire building is costly and difficult: This requires not only procuring a large number of sensors but also complex wiring and installation, significantly increasing costs. Furthermore, practical implementation is hampered by numerous technical and spatial limitations (specifically, the area of ​​the subway-adjacent building), making implementation extremely difficult. Secondly, short-term testing fails to reflect long-term structural vibration characteristics: Short-term vibration tests cannot effectively capture the gradual changes in vibration characteristics caused by factors such as subway passage, load fluctuations, and material aging during long-term use, making it difficult to accurately assess the long-term health of the structure. Thirdly, traditional methods struggle to balance accuracy and cost-effectiveness: Traditional vibration testing methods either prioritize high accuracy at the expense of cost, or compromise accuracy to meet practical needs due to cost control constraints, making it difficult to find a reasonable balance between accuracy and cost-effectiveness.

[0034] To address this, this application improves the reliability of the vibration detection model by periodically optimizing and training the pre-trained vibration detection model. The core of this approach is to build the model using long-term data from a small number of fixed points, and then use mobile detection data for full-domain calibration and verification, ultimately achieving a "point-to-surface" vibration assessment of the entire building.

[0035] In one exemplary embodiment, such as Figure 2 As shown, a training method for a vibration detection model of a building is provided, which can be applied to... Figure 1 The following steps, 201 to 203, are used as an example to illustrate the model training equipment.

[0036] Step 201: Obtain the training dataset.

[0037] In this application, the training dataset includes sample vibration data and label vibration data. Sample vibration data is collected using monitoring equipment installed at fixed locations within the building, while label vibration data is collected using mobile monitoring equipment. Specifically: The fixed-location monitoring equipment refers to a vibration monitoring device fixed within the building. The fixed location may include the building's vibration source area, transmission area, and / or resonance area. The monitoring equipment can be permanently fixed within the building or installed in a portable manner, and can be installed permanently at a fixed location within the building. The mobile monitoring equipment refers to a movable vibration detection device that can be moved by hand. The mobile monitoring equipment can detect the overall vibration data of the building, and its monitoring area may overlap or partially overlap with the monitoring area of ​​the fixed-location monitoring equipment. The building may be located above a subway station.

[0038] During implementation, the model training device acquires a training dataset consisting of sample vibration data and label vibration data. This can be achieved by the model training device receiving sample vibration data from a monitoring device and label vibration data from a mobile monitoring device; alternatively, an external device can acquire vibration data collected by the monitoring device to generate sample vibration data and receive vibration data from the mobile monitoring device to generate label vibration data, while the model training device receives the training dataset from the external device.

[0039] Step 202: Train the initial vibration detection model using sample vibration data to obtain candidate vibration detection models.

[0040] In the application of the model, the vibration detection model is used to predict the vibration state / safety state of a building. Vibration data corresponding to fixed monitoring points can be input into the vibration detection model, which then makes predictions to obtain the vibration state / safety state of the building. Based on this, during the model training process, the model training equipment can first train the initial vibration detection model based on sample vibration data to obtain candidate vibration models.

[0041] The initial vibration detection model can be an untrained model, or it can be pre-trained using initial sample data and initial label data to obtain an initial vibration detection model with vibration prediction capabilities. Furthermore, the initial vibration detection model is trained using sample vibration data collected by monitoring equipment to obtain candidate vibration detection models.

[0042] At this point, the candidate vibration detection model has not yet been optimized and trained. The vibration caused by the subway may have seasonal and periodic characteristics, and the candidate vibration detection model may have poor reliability when facing long-term data. Therefore, it is necessary to optimize and train it using labeled vibration data.

[0043] Step 203: Optimize and train the candidate vibration detection model based on the label vibration data to obtain the optimized target vibration detection model.

[0044] During implementation, the parameters of the candidate vibration detection model can be periodically adjusted based on the labeled vibration data to obtain the target vibration detection model. The vibration characteristics of this period or season can be characterized by the labeled vibration data, and then the target vibration detection model adapted to the vibration characteristics of this period can be trained.

[0045] During the execution process, the sample vibration data and the label vibration data can be used together for optimization training. That is, steps 202 to 203 can be replaced by: optimizing the pre-trained vibration detection model based on the sample vibration data and label vibration data included in the training dataset to obtain the optimized target vibration detection model.

[0046] The vibration data of the sample can be input into the vibration detection model to obtain the vibration index. Based on the labeled vibration data and the vibration index, the parameters of the vibration detection model are adjusted until the training error converges, and the target vibration detection model is obtained.

[0047] In the training method of the vibration detection model of the above-mentioned building, a training dataset is obtained. The training dataset includes sample vibration data collected by monitoring equipment set at fixed locations on the building, and labeled vibration data of the building collected by mobile monitoring equipment. The vibration detection model is trained by combining the sample vibration data and the labeled vibration data. First, the initial vibration detection model is trained using the sample vibration data to obtain candidate vibration detection models. Then, the candidate vibration detection models are optimized and trained based on the labeled vibration data to obtain the target vibration detection model after optimization training. The method of periodically optimizing and training the model using mobile monitoring equipment avoids the problem of inaccurate parameters caused by training the model based on short-term vibration data, thereby improving the reliability of the vibration detection model.

[0048] Based on the above exemplary embodiment, the following provides a method for training a vibration detection model of a building in one or more exemplary embodiments, which can be applied to... Figure 1 The following explanation will be based on the model training equipment used in the example.

[0049] In real-world scenarios, multiple monitoring devices can be installed on a building, distributed across different areas. These devices can acquire vibration time-history data from multiple areas, and then be sorted sequentially to obtain sample vibration data. One optional implementation provided in this application is as follows: Figure 3 As shown, the process of acquiring sample vibration data includes steps 301 to 302:

[0050] Step 301: Acquire multiple first vibration time history data collected by multiple monitoring devices.

[0051] During implementation, monitoring devices at multiple locations collect multiple first vibration time history data; the model training device acquires these multiple first vibration time history data.

[0052] In this application, the monitoring equipment can be installed in the vibration source area, transmission area, and resonance area of ​​the building; optionally, the fixed location includes the vibration source area, transmission area, and resonance area of ​​the building. The vibration source area can be the basement floor slab and side walls; the transmission area can be the bottom of the core tube, corner columns, transfer floors, and equipment floors; the resonance area can be the high-rise area or high-floor area of ​​the superstructure.

[0053] During execution, the vibration data collected by the monitoring equipment can be presented in the form of low-frequency acceleration data, which can be called vibration time history data. The monitoring equipment can collect vibration time history data of different areas within a fixed time period. The model training equipment can obtain the vibration time history data of different areas within a fixed time period collected by the monitoring equipment as the first time history data.

[0054] Step 302: Sort the first vibration time history data corresponding to each monitoring device according to the time sequence to obtain sample vibration data.

[0055] During implementation, the model training device can sort the first vibration time history data corresponding to each monitoring device in chronological order to obtain each first vibration time series data. After region labeling, the first vibration time series data are integrated to obtain sample vibration data. Optionally, the sample vibration data may include sample vibration time series data.

[0056] One optional implementation provided in this application involves collecting vibration time history data through monitoring equipment and integrating the vibration time history data into time series data. This method can better reflect the building's response when the subway passes by, thereby more accurately determining the vibration state of the entire building and improving the reliability and accuracy of the vibration detection model obtained through training.

[0057] In real-world scenarios, mobile monitoring devices can also collect vibration time-history data. These devices can collect vibration data from buildings in more dimensions, thus reflecting the building's vibration state and using this data as tag data to optimize the parameters of the vibration detection model. In one optional embodiment provided in this application, the process of acquiring tag vibration data includes steps 401 to 402:

[0058] Step 401: Acquire multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period.

[0059] During implementation, the mobile monitoring equipment can be carried by members and used to periodically detect vibration time history data of the building according to a preset route. It can periodically (every quarter or every six months) collect vibration data of the entire building. For example, it can collect vibration time history data of the floor slabs, beams, columns and other areas of the building. The model training equipment obtains each vibration time history data as the second vibration time history data.

[0060] Step 402: Determine the label vibration data of the target building based on each second vibration time history data.

[0061] During implementation, the model training equipment can determine the vibration index data of the building based on each second vibration time history data, and use the vibration index data as vibration label data. Among them, the vibration index data can be used to characterize the vibration state / safety state of the building, and the vibration index data may include vibration acceleration level, dominant frequency, 1 / 3 octave band spectrum, and RMS (Root Mean Square) value.

[0062] One optional implementation provided in this application uses mobile monitoring equipment to collect the overall vibration time history data of the building as tagged vibration data, which greatly expands the monitoring spatial range, eliminates blind spots of fixed monitoring points, and can capture abnormal vibrations of local components. This provides a wealth of calibration and validation data for the subsequent model development, ensuring that the model can accurately reflect the current state of the structure and improving the reliability of the trained model.

[0063] Furthermore, the vibrations caused by the subway can exhibit periodic and seasonal characteristics, allowing for the collection of second vibration time-history data at the beginning of each cycle to save data acquisition costs; in one optional implementation provided by this application, such as Figure 5 As shown, step 401 includes steps 501 to 502:

[0064] Step 501: When the model optimization period expires, generate a collection message for the label vibration data and send the collection message to the terminal.

[0065] During implementation, upon detecting the expiration of the model optimization period, the model training equipment generates a collection message for the label vibration data and sends the collection message to the terminals of the organization members. The model optimization period can be a pre-set period for optimizing and training the vibration detection model, which can be set to quarterly or semi-annual; the terminal information of the organization members can also be pre-set so that the collection message is sent to the terminals of the organization members when the model optimization period is detected to have expired.

[0066] After receiving the collection message, the members of the organization can respond to the collection message and retrieve the second vibration time history data collected by the mobile monitoring equipment.

[0067] Step 502: Receive the second vibration time history data sent by the terminal.

[0068] During implementation, members of the organization can transmit the second vibration time history data collected by the mobile monitoring device to the terminal, and then send the second time history data to the model training device through the terminal; the model training device receives the second vibration time history data sent by the terminal.

[0069] One optional implementation provided in this application enables the model monitoring device to obtain label vibration data in a timely manner by automatically sending messages after a set time, thereby improving the timeliness of model optimization training and thus improving the accuracy and reliability of the trained model.

[0070] During the optimization training process, the vibration detection model can be further optimized based on sample vibration data and label vibration data; in one optional implementation provided in this application, such as Figure 6 As shown, step 203 includes steps 601 to 602:

[0071] Step 601: Input the sample vibration data into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model.

[0072] During implementation, the model training device inputs the sample vibration model into the candidate vibration detection model, and the candidate vibration detection model performs vibration prediction processing to obtain vibration detection data; among which, the vibration detection data can be vibration index data.

[0073] During execution, the vibration detection model can be divided into multiple layers of network. The vibration detection data can be obtained by the prediction layer of the vibration detection model. The prediction layer of the vibration detection model can be trained based on sample vibration data and labeled vibration data. Step 601 can be replaced by: inputting the sample vibration data into the prediction layer of the candidate vibration detection model to obtain the vibration detection data of the target building output by the prediction layer of the candidate vibration detection model.

[0074] Step 602: Calculate the error between the vibration detection data and the tag vibration data, and adjust the parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0075] During implementation, the model training equipment calculates the error between the vibration detection data and the label vibration data, adjusts the parameters of the vibration detection model according to the error, and performs this operation for each pair of sample vibration data and label vibration data until the error of the candidate vibration detection model converges, thus obtaining the target vibration detection model.

[0076] During execution, the model training device can also calculate the error between the vibration detection data and the label vibration data, and adjust the parameters of the prediction layer of the vibration detection model according to the error; step 602 can be replaced by: calculating the error between the vibration detection data and the label vibration data, and adjusting the parameters of the prediction layer of the candidate vibration detection model according to the error until the error converges, and obtaining the target vibration detection model.

[0077] For example, the prediction layer of the model can be established using the Transmission Path Analysis (TPA) method, and the prediction layer of the model can be established based on formula (1):

[0078] Formula (1);

[0079] Where Pt is the total vibration response at the target point, Pi is the contribution of the i-th transmission path to the total response Pt at the target point, (P / F) is the transfer function or frequency response function (FRF) of the i-th path, which describes the dynamic transmission characteristics from the input point of the path (usually the point of force application) to the target point, and Fi is the excitation force (or vibration intensity) acting on the input point of the i-th path.

[0080] One optional implementation method provided in this application improves the reliability and accuracy of the vibration detection model by optimizing and training the candidate vibration detection model.

[0081] In real-world scenarios, target vibration detection models can be used to perform overall vibration detection of buildings to determine whether the building's vibration state is safe; in one optional implementation method provided in this application, such as... Figure 7 As shown, the application process of the target vibration detection model includes steps 701 to 702:

[0082] Step 701: Obtain vibration sequence data collected by the monitoring equipment, input the vibration sequence data into the target vibration detection model, and obtain the vibration index data of the building output by the target vibration detection model.

[0083] During implementation, the model application equipment acquires vibration sequence data / vibration time series data collected by monitoring equipment at fixed locations, and inputs the vibration sequence data into the target vibration detection model; the target vibration detection model performs prediction processing based on the vibration sequence data / vibration time series data to obtain the vibration index data of the building; the model application equipment obtains the vibration index data output from the target vibration detection data.

[0084] Step 702: Detect whether the vibration index data is within the preset index threshold. If the vibration index data is within the preset index threshold, determine that the vibration state of the building is safe.

[0085] During implementation, vibration index data can be of various types. The model application equipment can detect multiple vibration index data in sequence, and check whether each vibration index data is within a preset index threshold. If it is, the vibration state of the building is determined to be safe; if not, the vibration state of the building is determined to be abnormal. Alternatively, normalized vibration index data can be calculated based on multiple vibration index data, and the normalized vibration index data can be checked whether it is within a preset index threshold. If it is, the vibration state of the building is determined to be safe; if not, the vibration state of the building is determined to be abnormal.

[0086] During execution, the target vibration detection model can also directly output the building's vibration state based on the vibration sequence data. Specifically, vibration index data can be output through the prediction layer of the target vibration detection model. That is, steps 701 to 702 can be replaced by: acquiring vibration sequence data collected by the monitoring equipment, inputting the vibration sequence data into the target vibration detection model, and obtaining the building's vibration state output by the target vibration detection model. The prediction layer of the target vibration model outputs vibration index data based on the vibration sequence data, and the output layer of the target vibration model outputs the building's vibration state based on the vibration index data.

[0087] One optional implementation method provided in this application uses a target vibration detection model to detect the vibration of a building, achieving a comprehensive, real-time, and high-precision vibration status assessment. Managers can monitor the vibration response of the entire building during subway operation in real time, accurately locate areas with excessive vibration, assess the impact of subway operation on sensitive areas within the building, and, based on long-term data, analyze trends in structural stiffness degradation and other health conditions, thereby improving the accuracy of vibration status detection for buildings.

[0088] In one embodiment, see Figure 8 The document illustrates a flowchart of a method for training a vibration detection model of a building, as provided in an embodiment of this application. This method can be applied to... Figure 1 The model training device shown. (As indicated) Figure 8 As shown, the training method for the vibration detection model of this building may include the following steps:

[0089] Step 801: Obtain sample vibration data collected by fixed monitoring equipment for the vibration source area, transmission area and resonance area of ​​the building above the subway.

[0090] Step 802: Obtain the tag vibration data collected by the mobile monitoring device for the building above the subway.

[0091] Step 803: Input the sample vibration data into the pre-trained vibration detection model to obtain the vibration detection data output by the prediction layer of the pre-trained vibration detection model.

[0092] Step 804: Calculate the training error based on the vibration detection data and the label vibration data, and adjust the parameters of the prediction layer of the vibration detection model according to the training error until the training error converges to obtain the target vibration detection model.

[0093] It should be noted that any one or more of steps 801 to 804 can be combined to form a new implementation method according to the needs of implementation and deployment. Furthermore, any one or more technical features in the technical solution composed of steps 801 to 804 can also be combined to form a new implementation method according to the actual deployment needs, or technical features in one or more optional implementations provided by one or more of the above embodiments can be combined to form a new implementation method. These will not be elaborated on here.

[0094] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0095] Based on the same inventive concept, this application also provides a training device for a building vibration detection model to implement the training method for the building vibration detection model described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more building vibration detection model training device embodiments provided below can be found in the limitations of the building vibration detection model training method described above, and will not be repeated here.

[0096] In one exemplary embodiment, such as Figure 9As shown, a training device for a vibration detection model of a building is provided, including: a training dataset acquisition module 901, a first model training module 902, and a second model training module 903. The training dataset acquisition module 901 is used to acquire a training dataset, which includes sample vibration data and label vibration data. The sample vibration data is collected by a monitoring device installed at a fixed location on the building, and the label vibration data is collected by a mobile monitoring device. The first model training module 902 is used to train an initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model. The second model training module 903 is used to optimize the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model after optimization training.

[0097] In one embodiment, the device further includes a vibration data acquisition module and a sample generation module, wherein: the vibration data acquisition module is used to acquire multiple first vibration time history data collected by multiple monitoring devices; the sample generation module is used to sort the first vibration time history data corresponding to each monitoring device according to the time order to obtain sample vibration data.

[0098] In one embodiment, the device further includes a vibration time history data acquisition module and a tag generation module, wherein: the vibration time history data acquisition module is used to acquire multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period; the tag generation module is used to determine the tag vibration data of the target building based on each second vibration time history data.

[0099] In one embodiment, the vibration time history data acquisition module includes a data acquisition message generation unit and a data receiving unit, wherein: the data acquisition message generation unit is used to generate a data acquisition message for the label vibration data when the model optimization period is detected to be expired, and send the data acquisition message to the terminal; the data receiving unit is used to receive each second vibration time history data sent by the terminal.

[0100] In one embodiment, the second model training module 903 further includes a data input unit and a parameter adjustment unit, wherein: the data input unit is used to input sample vibration data into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model; the parameter adjustment unit is used to calculate the error between the vibration detection data and the label vibration data, and adjust the parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0101] In one embodiment, the device further includes a vibration index acquisition module and a state determination module, wherein: the vibration index acquisition module is used to acquire vibration sequence data collected by the monitoring equipment, input the vibration sequence data into the target vibration detection model, and obtain the vibration index data of the building output by the target vibration detection model; the state determination module is used to detect whether the vibration index data is within a preset index threshold, and if the vibration index data is within the preset index threshold, the vibration state of the building is determined to be a safe state.

[0102] The modules in the training device for the vibration detection model of the aforementioned building can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0103] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores model training data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a training method for a vibration detection model of a building.

[0104] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0105] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring a training dataset, the training dataset including sample vibration data and label vibration data, the sample vibration data being collected by a monitoring device installed at a fixed location on a building, and the label vibration data being collected by a mobile monitoring device; training an initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model; and optimizing the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model after optimization training.

[0106] In one embodiment, when the processor executes the computer program, it further performs the following steps: acquiring multiple first vibration time history data collected by multiple monitoring devices; fixing the location including the vibration source region, transmission region and resonance region of the building; sorting the first vibration time history data corresponding to each monitoring device according to the time sequence to obtain sample vibration data.

[0107] In one embodiment, when the processor executes the computer program, it further performs the following steps: acquiring multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period; and determining the tag vibration data of the target building based on each second vibration time history data.

[0108] In one embodiment, when the processor executes the computer program, it further performs the following steps: upon detecting that the model optimization period has expired, generating a collection message for the label vibration data and sending the collection message to the terminal; and receiving each second vibration time history data sent by the terminal.

[0109] In one embodiment, when the processor executes the computer program, it further performs the following steps: inputting sample vibration data into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model; calculating the error between the vibration detection data and the label vibration data, and adjusting the parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0110] In one embodiment, when the processor executes the computer program, it further implements the following steps: acquiring vibration sequence data collected by the monitoring device, inputting the vibration sequence data into the target vibration detection model, and obtaining the vibration index data of the building output by the target vibration detection model; detecting whether the vibration index data is within a preset index threshold, and if the vibration index data is within the preset index threshold, determining that the vibration state of the building is a safe state.

[0111] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, performs the following steps: acquiring a training dataset, the training dataset including sample vibration data and label vibration data, wherein the sample vibration data is acquired by a monitoring device installed at a fixed location on a building, and the label vibration data is acquired by a mobile monitoring device; training an initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model; and optimizing the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model that has been optimized and trained.

[0112] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring multiple first vibration time history data collected by multiple monitoring devices; fixing the location including the vibration source region, transmission region and resonance region of the building; sorting the first vibration time history data corresponding to each monitoring device according to the time sequence to obtain sample vibration data.

[0113] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period; and determining the tag vibration data of the target building based on each second vibration time history data.

[0114] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: upon detecting that the model optimization period has expired, generating a collection message for the label vibration data and sending the collection message to the terminal; and receiving each second vibration time history data sent by the terminal.

[0115] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: inputting sample vibration data into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model; calculating the error between the vibration detection data and the label vibration data, and adjusting the parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0116] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring vibration sequence data collected by the monitoring device, inputting the vibration sequence data into the target vibration detection model, and obtaining the vibration index data of the building output by the target vibration detection model; detecting whether the vibration index data is within a preset index threshold, and if the vibration index data is within the preset index threshold, determining that the vibration state of the building is a safe state.

[0117] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps: acquiring a training dataset, the training dataset including sample vibration data and label vibration data, wherein the sample vibration data is acquired by a monitoring device installed at a fixed location on a building, and the label vibration data is acquired by a mobile monitoring device; training an initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model; and optimizing the candidate vibration detection model based on the label vibration data to obtain a target vibration detection model that has been optimized and trained.

[0118] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring multiple first vibration time history data collected by multiple monitoring devices; fixing the location including the vibration source region, transmission region and resonance region of the building; sorting the first vibration time history data corresponding to each monitoring device according to the time sequence to obtain sample vibration data.

[0119] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period; and determining the tag vibration data of the target building based on each second vibration time history data.

[0120] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: upon detecting that the model optimization period has expired, generating a collection message for the label vibration data and sending the collection message to the terminal; and receiving each second vibration time history data sent by the terminal.

[0121] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: inputting sample vibration data into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model; calculating the error between the vibration detection data and the label vibration data, and adjusting the parameters of the candidate vibration detection model according to the error until the error of the candidate vibration detection model converges to obtain the target vibration detection model.

[0122] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: acquiring vibration sequence data collected by the monitoring device, inputting the vibration sequence data into the target vibration detection model, and obtaining the vibration index data of the building output by the target vibration detection model; detecting whether the vibration index data is within a preset index threshold, and if the vibration index data is within the preset index threshold, determining that the vibration state of the building is a safe state.

[0123] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0124] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0125] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0126] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A training method for a vibration detection model of a building, characterized in that, The method includes: A training dataset is obtained, which includes sample vibration data and label vibration data. The sample vibration data is collected by monitoring devices installed at fixed locations on the building, and the label vibration data is collected by mobile monitoring devices. The initial vibration detection model is trained using the sample vibration data to obtain candidate vibration detection models; The candidate vibration detection model is optimized and trained based on the labeled vibration data to obtain the target vibration detection model after optimization and training.

2. The method according to claim 1, characterized in that, The process of acquiring the sample vibration data includes: Multiple first vibration time history data collected by multiple monitoring devices are acquired; the fixed location includes the vibration source region, transmission region and resonance region of the building; The first vibration time history data corresponding to each of the monitoring devices are sorted according to time sequence to obtain the sample vibration data.

3. The method according to claim 1 or 2, characterized in that, The process of acquiring the tag vibration data includes: Multiple second vibration time history data collected by multiple mobile monitoring devices are acquired according to a preset time period; The label vibration data of the target building is determined based on each of the second vibration time history data.

4. The method according to claim 3, characterized in that, The step of acquiring multiple second vibration time history data collected by multiple mobile monitoring devices according to a preset time period includes: If the model optimization period expires, a data acquisition message for the label vibration data is generated and sent to the terminal. Receive each of the second vibration time history data sent by the terminal.

5. The method according to claim 1 or 2, characterized in that, The step of optimizing and training the candidate vibration detection model based on the labeled vibration data to obtain the optimized target vibration detection model includes: The sample vibration data is input into the candidate vibration detection model to obtain the vibration detection data of the target building output by the candidate vibration detection model. The error between the vibration detection data and the tag vibration data is calculated, and the parameters of the candidate vibration detection model are adjusted according to the error until the error of the candidate vibration detection model converges, thereby obtaining the target vibration detection model.

6. The method according to claim 1 or 2, characterized in that, The application process of the target vibration detection model includes: The vibration sequence data collected by the monitoring device is obtained, and the vibration sequence data is input into the target vibration detection model to obtain the vibration index data of the building output by the target vibration detection model. The vibration index data is detected to be within a preset index threshold. If the vibration index data is within the preset index threshold, the vibration state of the building is determined to be a safe state.

7. A training device for a vibration detection model of a building, characterized in that, The device includes: The training dataset acquisition module is used to acquire the training dataset, which includes sample vibration data and label vibration data. The sample vibration data is collected by monitoring equipment set at a fixed location on the building, and the label vibration data is collected by mobile monitoring equipment. The first model training module is used to train the initial vibration detection model using the sample vibration data to obtain a candidate vibration detection model. The second model training module is used to optimize and train the candidate vibration detection model based on the label vibration data to obtain the optimized target vibration detection model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images