Cultural relic showcase quakeproof prediction method and system based on big data analysis

Abstract

The invention discloses a cultural relic showcase quakeproof prediction method and system based on big data analysis, and the method comprises the steps: obtaining physical parameters and historical vibration data of a cultural relic showcase, carrying out the preprocessing of the physical parameters and historical vibration data of the cultural relic showcase, and generating a weighted multi-dimensional dynamic time sequence matrix; initializing a deep learning model, sending the weighted multi-dimensional dynamic time sequence matrix into the deep learning model, optimizing the deep learning model, and generating a vibration risk factor of the cultural relic showcase; carrying out anti-seismic prediction on the cultural relic showcase according to the vibration risk factor, dividing a prediction result into short-term prediction and medium-and-long-term prediction, and carrying out reliability evaluation on the two prediction results to ensure the accuracy of anti-seismic prediction on the cultural relic showcase; compared with a traditional rule-based vibration assessment method, the method has the advantages of high prediction precision and wide adaptability, and can be widely applied to vibration protection and risk management scenes of cultural relic showcases.

Description

Technical Field

[0001] The present invention relates to the technical field of earthquake-proof prediction for cultural relic display cabinets, and particularly to a method and system for earthquake-proof prediction of cultural relic display cabinets based on big data analysis. Background Art

[0002] With the development of cultural relic protection and display technologies, cultural relic display cabinets, as important carriers for protecting and displaying cultural relics, have gradually become the focus of research in the field of cultural relic safety management. However, since cultural relic display cabinets are usually in complex environments, such as areas with frequent seismic activities or crowded museums, the vibration effects they bear cannot be ignored. Traditional anti-seismic designs for display cabinets mostly rely on static parameters and simple structural optimizations. Although this method can improve the seismic resistance of display cabinets to a certain extent, in the face of diverse and dynamic vibration environments, the adaptability and accuracy of traditional methods are significantly insufficient. In addition, the multi-dimensional dynamic characteristics of vibration data cannot be fully captured by simple models, which leads to great limitations in the reliability and timeliness of existing display cabinet anti-seismic measures for vibration prediction.

[0003] In recent years, with the rapid development of big data technologies and deep learning algorithms, new solutions have been provided for the earthquake-proof prediction of cultural relic display cabinets. However, existing data-driven methods still have many deficiencies. For example, most models fail to comprehensively consider the physical characteristics of cultural relic display cabinets and the dynamic changes of vibration signals, resulting in relatively high accuracy of prediction results in the short term, but lack of stability in medium- and long-term predictions. Moreover, most existing models are based on rules or hierarchical thresholds, and the effects are not ideal when facing complex periodic changes and local features. In addition, the risk assessment means of existing methods are relatively single, and the reliability of prediction results cannot be fully evaluated, making it difficult to meet the earthquake-proof requirements of cultural relic display cabinets in different scenarios. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title, and such simplifications or omissions shall not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for earthquake-proof prediction of cultural relic display cabinets based on big data analysis to solve the problems mentioned in the background art.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for earthquake-proof prediction of cultural relic display cabinets based on big data analysis, including:

[0008] Obtain the physical parameters and historical vibration data of the cultural relic display cabinet, preprocess the physical parameters and historical vibration data of the cultural relic display cabinet, and generate a weighted multi-dimensional dynamic time series matrix;

[0009] Initialize the deep learning model, send the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimize the deep learning model, and generate the vibration risk factor of the cultural relic display cabinet;

[0010] Perform earthquake prevention prediction on the cultural relic display cabinet with the vibration risk factor, divide the prediction results into short-term prediction and medium- and long-term prediction, and evaluate the reliability of the above two prediction results to ensure the accuracy of the earthquake prevention prediction of the cultural relic display cabinet.

[0011] As a preferred solution of the earthquake prevention prediction method for cultural relic display cabinets based on big data analysis according to the present invention, wherein: obtaining the physical parameters and historical vibration data of the cultural relic display cabinet, preprocessing the physical parameters and historical vibration data of the cultural relic display cabinet, and generating a weighted multi-dimensional dynamic time series matrix, including:

[0012] Take the physical parameters of the cultural relic display cabinet as static data, and take the historical vibration data of the cultural relic display cabinet as dynamic data;

[0013] Process the static data in a normalized manner to obtain a static feature vector, and use the static feature vector as the global information of the time series matrix, running through all time steps;

[0014] The historical vibration data of the cultural relic display cabinet is indexed by the time step to construct an original time series matrix. Taking the columns of the original time series matrix as units, perform standardized processing of sliding windows on the data of each column to generate a standardized dynamic time series matrix;

[0015] Based on each time step, merge the static feature vector and the dynamic feature vectors in the dynamic time series matrix to obtain multiple multi-dimensional feature vectors, calculate the contribution degree of each multi-dimensional feature vector to obtain a weight list of the multi-dimensional feature vectors, and multiply the weight list of the multi-dimensional feature vectors by the standardized dynamic time series matrix to obtain a weighted multi-dimensional dynamic time series matrix.

[0016] As a preferred solution of the earthquake prevention prediction method for cultural relic display cabinets based on big data analysis according to the present invention, wherein: further including:

[0017] Add periodic labels to the weighted multi-dimensional dynamic time series matrix to capture the periodic changes of the cultural relic display cabinet during vibration, and at the same time consider the first derivative of the acceleration of the cultural relic display cabinet during vibration to capture the speed and displacement changes of the cultural relic display cabinet during vibration.

[0018] As a preferred solution of the anti-seismic prediction method for cultural relic display cabinets based on big data analysis according to the present invention, it includes:

[0019] Initialize the dimension of the input layer;

[0020] Initialize the convolutional kernels in the convolutional layer, where the size of the convolutional kernels is composed of the number of time windows and the number of features;

[0021] Initialize the size of the pooling window in the pooling layer;

[0022] Initialize the weight matrix and bias vector in the fully connected layer;

[0023] Initialize the output layer as a single neuron, and the activation function is the Tanh function.

[0024] As a preferred solution of the anti-seismic prediction method for cultural relic display cabinets based on big data analysis according to the present invention, it includes: sending the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimizing the deep learning model, and generating the vibration risk factor of the cultural relic display cabinet, including:

[0025] The input layer sets its own input dimension to the same value as the dimension of the weighted multi-dimensional dynamic time series matrix, and sends the multi-dimensional dynamic time series matrix into the convolutional layer. The convolutional layer determines the number of features and the number of time windows of the convolutional kernels respectively according to the number of feature vectors and the number of time steps in the weighted multi-dimensional dynamic time series matrix;

[0026] Obtain the corresponding feature map through the convolutional kernels, send the feature map into the pooling layer for dimensionality reduction and set the pooling window size as max pooling, and then send the dimensionality-reduced feature map into the fully connected layer. The fully connected layer obtains the weight matrix according to the dimensionality-reduced feature map;

[0027] Multiply the dimensionality-reduced feature map by the weight matrix and add the bias vector to obtain the output vector of the fully connected layer. Send the output vector into the output layer to output the vibration risk factor of the cultural relic display cabinet.

[0028] As a preferred solution of the anti-seismic prediction method for cultural relic display cabinets based on big data analysis according to the present invention, it includes: performing anti-seismic prediction on the cultural relic display cabinet with the vibration risk factor, and dividing the prediction results into short-term prediction and medium- and long-term prediction, including:

[0029] Perform short-term and medium- and long-term prediction on the vibration risk factor in a time series manner;

[0030] When the vibration risk factor is between [0, 0.3) and remains within this range within a future time window, it indicates that the cultural relic display case will not be affected by vibration, and there is no need to pre-adjust the cultural relics in the display case;

[0031] When the vibration risk factor is between [0.3, 0.7) and remains within this range within a future time window, it indicates that the cultural relic display case will experience slight vibration but will not affect the cultural relics themselves, and it is necessary to pre-strengthen the stability of the cultural relic display case;

[0032] When the vibration risk factor ≥ 0.7 and remains within this range within multiple future time windows, it indicates that the cultural relic display case will experience large vibrations and will affect the cultural relics themselves, and it is necessary to pre-transfer the cultural relics in the cultural relic display case or reinforce the cultural relic display case;

[0033] Among them, short-term prediction refers to a future time window, and long-term prediction refers to multiple future time windows.

[0034] As a preferred solution of the cultural relic display case earthquake prevention prediction method based on big data analysis according to the present invention, wherein: by performing reliability evaluation on the above two prediction results, including:

[0035] Divide each time window into K groups, each time take K - 1 groups as training data until only 1 group remains in the K groups, and use the remaining 1 group as verification data;

[0036] Use the Pearson correlation coefficient r to measure the consistency between the predicted range and the actual range, and obtain the reliability evaluation results of the two prediction results.

[0037] In the second aspect, the present invention provides a cultural relic display case earthquake prevention prediction system based on big data analysis, which includes:

[0038] A cultural relic display case data processing module, configured to obtain the physical parameters and historical vibration data of the cultural relic display case, preprocess the physical parameters and historical vibration data of the cultural relic display case, and generate a weighted multi-dimensional dynamic time series matrix;

[0039] A cultural relic display case vibration risk generation module, configured to initialize a deep learning model, send the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimize the deep learning model, and generate a vibration risk factor of the cultural relic display case;

[0040] A cultural relic display case vibration risk prediction module, configured to perform earthquake prevention prediction on the cultural relic display case based on the vibration risk factor, divide the prediction results into short-term prediction and medium- and long-term prediction, and perform reliability evaluation on the above two prediction results to ensure the accuracy of the earthquake prevention prediction of the cultural relic display case.

[0041] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the processor executes the computer program, any step of the above method is implemented.

[0042] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by a processor, any step of the above method is implemented.

[0043] Compared with the prior art, the beneficial effects of the invention are as follows:

[0044] 1. By combining the static physical parameters of the cultural relic display cabinet with historical vibration data and using a weighted multi-dimensional dynamic time series matrix to construct a deep learning model, it is possible to capture complex temporal features and vibration patterns from vibration signals, optimize the capture ability of vibration amplitude, frequency, and periodic changes, improve the calculation accuracy of vibration risk factors, and thus achieve a more reliable earthquake prevention prediction effect;

[0045] 2. Conduct a dynamic trend analysis of the vibration risk factors to ensure that the prediction results cover the risk changes in different future time windows. Compared with the traditional method that is limited to a prediction model with a single time scale, the present invention optimizes the convolutional and pooling layers through deep learning, can capture the vibration characteristics within medium and long time series, and further improves the globality and robustness of the prediction, helping cultural relic managers take scientific protection measures before minor or major vibrations occur;

[0046] 3. Evaluate the reliability of the prediction results, systematically solve the problem of high error caused by the lack of a result verification mechanism in the existing methods. Since the traditional model is easily interfered by external noise when classifying vibration risks, through multiple rounds of grouped training and verification, the sensitivity of the model to abnormal data is effectively reduced, making the prediction of vibration risk factors more stable, thereby reducing the risk of wrong decisions and comprehensively ensuring the safety of cultural relics. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Among them:

[0048] Figure 1 is the overall flowchart of the anti-seismic prediction method for cultural relic display cabinets based on big data analysis according to an embodiment of the present invention;

[0049] Figure 2Flow chart of the normalization process of the sliding window for the anti-seismic prediction method of cultural relic display cabinets based on big data analysis according to an embodiment of the present invention;

[0050] Figure 3 Short-term model prediction comparison chart for the anti-seismic prediction method of cultural relic display cabinets based on big data analysis according to an embodiment of the present invention;

[0051] Figure 4 Medium- and long-term model prediction comparison chart for the anti-seismic prediction method of cultural relic display cabinets based on big data analysis according to an embodiment of the present invention. Detailed implementation manners

[0052] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific implementation manners of the present invention with reference to the accompanying drawings of the specification. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0054] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that mutually excludes other embodiments.

[0055] The present invention is described in detail with reference to schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views showing the device structure will be enlarged locally not in accordance with the general scale, and the schematic diagrams are only examples and should not limit the scope of protection of the present invention herein. In addition, in actual production, three-dimensional spatial dimensions including length, width, and depth should be included.

[0056] Meanwhile, in the description of the present invention, it should be noted that the orientation or positional relationships indicated by terms such as "upper, lower, inner, and outer" are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention. In addition, the terms "first, second, or third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0057] Unless otherwise clearly specified and defined in the present invention, the terms "installation, connection, and coupling" shall be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may also be a mechanical connection, an electrical connection, or a direct connection, or may be indirectly connected through an intermediate medium, or may be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention may be understood according to specific circumstances.

[0058] Embodiment 1

[0059] Referring to Figure 1 and Figure 2 , this is the first embodiment of the present invention. This embodiment provides a method for predicting the earthquake resistance of a cultural relic display case based on big data analysis, including:

[0060] S1. Obtain the physical parameters and historical vibration data of the cultural relic display case, preprocess the physical parameters and historical vibration data of the cultural relic display case, and generate a weighted multi-dimensional dynamic time series matrix;

[0061] Specifically, the physical parameters refer to the fixed parameters related to the cultural relic display case itself, which usually do not change with time. They mainly include the load parameters of the cultural relic display case (the weight of the cultural relic itself and its distribution in the display case), the connection point characteristic parameters (the stiffness and strength of the connection between the display case components), and the material parameters (elastic modulus, damping ratio), etc.;

[0062] Specifically, the historical vibration data refers to the time series characteristics of vibrations, which are usually obtained by sensors, monitoring devices, or historical earthquake records. They mainly include the vibration amplitude (the amplitude range of the vibration acceleration, velocity, or displacement), the vibration frequency (the spectral distribution of the vibration, especially the proportion of the low-frequency and high-frequency vibration components), the vibration period (the periodic change characteristics of each vibration, that is, the regularity of the vibration source), etc.;

[0063] Furthermore, the physical parameters of the cultural relic display case are used as static data, and the historical vibration data of the cultural relic display case are used as dynamic data;

[0064] Even further, the static data is processed in a normalized manner to obtain a static feature vector, and the static feature vector is used as the global information of the time series matrix, running through all time steps;

[0065] Specifically, the normalization processing method adopts the Min-Max normalization and Z-Score standardization selection method. Among them, for data with a known range, the Min-Max normalization is used to linearly map the data to the interval [0, 1], while for data with an unknown range, the Z-Score standardization is used to transform the data into zero mean and unit variance;

[0066] Exemplarily, assume that the density range of the cultural relic display case is 500 - 800 kg / m 3 , and the density of a certain cultural relic display case is 600 kg / m 3 , then after Min - Max normalization, it is:

[0067]

[0068] Among them, min(x) represents the minimum value of x within the range, max(x) represents the maximum value of x within the range; x is the specific data;

[0069] Exemplarily, assume that the mean value of the load data of the cultural relic display case is 100 kg, and the standard deviation is 15 kg. Among them, the load of a certain cultural relic display case is 115 kg, then after Z - Score standardization, it is:

[0070]

[0071] Among them, μ is the mean value, and σ is the standard deviation;

[0072] It should be noted that through normalization, static data is transformed into dimensionless and uniformly - ranged feature vectors, reducing the influence caused by the difference in the order of magnitude of different physical quantities;

[0073] Specifically, taking the static feature vector as the global information of the time - series matrix and running through all time steps means that the static feature vector is repeatedly assigned to each row (i.e., each time step) of the time - series matrix. The purpose is to enable the model to always perceive the global attributes independent of time when processing the time - series matrix, such as the material, shape, weight distribution, etc. of the cultural relic display case;

[0074] Furthermore, the historical vibration data of the cultural relic display case is indexed by time steps to construct the original time - series matrix. Taking the columns of the original time - series matrix as units, standardization processing with a sliding window is performed on the data of each column to generate a standardized dynamic time - series matrix;

[0075] Specifically, in the constructed original time - series matrix, each column represents a specific vibration feature (such as vibration acceleration), and indexing by time steps allows the values of these features to change with time. Moreover, during subsequent processing, standardized processing with a sliding window can be performed on each specific feature value separately;

[0076] Specifically, the standardized processing with a sliding window is similar to the previous normalization processing, but the difference is that the previous normalization processing is mainly used to reduce the influence between the differences in the order of magnitude of different physical quantities, while the standardized processing with a sliding window is mainly used for segmenting data in the time series, referring to Figure 2 , and the operation is as follows:

[0077] S101. Set the size of the sliding window to W (i.e., consider the data of consecutive W time steps each time).

[0078] S102. Slide the window at a fixed step size, for example, slide W time steps each time.

[0079] S103. When the window slides each time, select the data of all time steps within the window for processing.

[0080] Exemplarily, assume the window size W = 3 and the time series is [x 1 , x 2 , x 3 , x 4 , x 5 . Then the following segments are obtained:

[0081] [x 1 , x 2 , x 3 . [x 2 , x 3 , x 4 . [x 3 , x 4 , x 5

[0082] Furthermore, based on each time step, merge the static feature vectors and the dynamic feature vectors in the dynamic time series matrix to obtain multiple multi-dimensional feature vectors, calculate the contribution degree of each multi-dimensional feature vector to obtain the weight list of the multi-dimensional feature vectors, and multiply the weight list of the multi-dimensional feature vectors by the normalized dynamic time series matrix to obtain the weighted multi-dimensional dynamic time series matrix.

[0083] Specifically, the merging method is to add the eigenvalues of the static feature vector and the dynamic feature vector at this time step to obtain a multi-dimensional feature vector, and so on until all time steps are completed, and multiple multi-dimensional feature vectors can be obtained.

[0084] It should be noted that the contribution degree of the multi-dimensional feature vector can be calculated by statistical methods (such as information gain, entropy method, etc.) to obtain the weight of each multi-dimensional feature vector, which will not be elaborated here.

[0085] Furthermore, add a period label to the weighted multi-dimensional dynamic time series matrix to capture the periodic changes when the cultural relic display cabinet vibrates, and at the same time consider the first derivative of the acceleration when the cultural relic display cabinet vibrates to capture the speed and displacement changes when the cultural relic display cabinet vibrates.

[0086] ​Specifically, the addition period label uses the Fast Fourier Transform (FFT) to extract the main frequency component in the weighted multi-dimensional dynamic time series matrix as the period, and the first derivative of the acceleration during the vibration of the cultural relic display cabinet is considered in terms of the period change;

[0087] It should be noted that since the first derivative of the acceleration is the rate of change of the acceleration with respect to time, representing the change in velocity during the vibration process, substituting it with the period here is equivalent to performing a low-pass filtering process on the vibration data from the time domain to the period domain, only focusing on the average rate of change within a specific period and being insensitive to the minor fluctuations within the period. Moreover, through the calculation method of the period benchmark, the cumulative displacement and velocity effects of the display cabinet under periodic vibration can be captured, which is more suitable for evaluating whether the cultural relic display cabinet will generate resonance or gradual offset phenomena due to periodic vibration;

[0088] S2. Initialize the deep learning model, send the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimize the deep learning model, and generate the vibration risk factor of the cultural relic display cabinet;

[0089] Furthermore, initialize the input layer dimension in the deep learning model, the convolution kernel in the convolutional layer (the convolution kernel size is composed of the number of time windows and the number of features), the pooling window size in the pooling layer, the weight matrix and bias vector in the fully connected layer, and the output layer as a single neuron, and the activation function is the Tanh function;

[0090] Even further, the input layer receives the weighted multi-dimensional dynamic time series matrix, sets its own dimension size to be the same as that of the weighted multi-dimensional dynamic time series matrix according to the dimension size of the weighted multi-dimensional dynamic time series matrix, and then sends the multi-dimensional dynamic time series matrix into the convolutional layer. The convolutional layer determines the number of features of the convolution kernel and the number of time windows respectively according to the number of feature vectors and the number of time steps in the weighted multi-dimensional dynamic time series matrix;

[0091] It should be noted that in the solution of the present invention, only the case where the number of time steps is the same as the number of feature vectors is considered;

[0092] Even further, the corresponding feature map is obtained through the convolution kernel, the feature map is sent into the pooling layer for dimensionality reduction and the pooling window size is set to max pooling, and then the dimensionality-reduced feature map is sent into the fully connected layer. The fully connected layer obtains the weight matrix according to the dimensionality-reduced feature map;

[0093] Specifically, each convolution kernel will generate a feature map which includes the batch size and the number of elements, and multiple convolution kernels can capture different types of vibration features;

[0094] Specifically, the dimensionality reduction method uses max pooling to reduce each local area of the feature map to a maximum value, so as to retain the feature of the time period with the greatest vibration influence;

[0095] It should be noted that the weight matrix is obtained from the size of the feature map after dimensionality reduction, and the size of the feature map is obtained from the number of feature channels of the convolutional kernel, the number of time windows, and the input feature dimension;

[0096] Furthermore, multiply the feature map after dimensionality reduction by the weight matrix and add the bias vector to obtain the output vector of the fully connected layer. By sending the output vector into the output layer, the vibration risk factor of the cultural relic display cabinet is output;

[0097] It should be noted that the bias vector mainly avoids the model output depending on zero-point symmetry. For example, the vibration feature itself is very small, and the linear mapping of the weight matrix may not be sufficient to shift the output. After adding the bias vector b, the output can be shifted;

[0098] S3. Perform earthquake prevention prediction on the cultural relic display cabinet with the vibration risk factor, and divide the prediction results into short-term prediction and medium- and long-term prediction. By evaluating the reliability of the above two prediction results, the accuracy of the earthquake prevention prediction of the cultural relic display cabinet is ensured;

[0099] Further, perform short-term and medium- and long-term predictions on the vibration risk factor in a time series manner, where the short-term prediction is as follows:

[0100] When the vibration risk factor is between [0, 0.3) and remains within this range in the next time window, it indicates that the cultural relic display cabinet will not be affected by vibration and there is no need to pre-adjust the cultural relics in the display cabinet;

[0101] When the vibration risk factor is between [0.3, 0.7) and remains within this range in the next time window, it indicates that the cultural relic display cabinet will have slight vibration but will not affect the cultural relics themselves, and it is necessary to pre-strengthen the stability of the cultural relic display cabinet;

[0102] The medium- and long-term prediction is as follows:

[0103] When the vibration risk factor ≥ 0.7 and remains within this range in multiple future time windows, it indicates that the cultural relic display cabinet will have large vibration and will affect the cultural relics themselves, and it is necessary to pre-transfer the cultural relics in the cultural relic display cabinet or reinforce the cultural relic display cabinet;

[0104] It should be noted that the introduction of the time series better conforms to the dynamic change law of the vibration risk in the actual application scenario;

[0105] Further, to verify the accuracy of the vibration risk factor range in short-term prediction and medium- and long-term prediction, each time window is divided into K groups. Each time, K - 1 groups are taken as training data until only 1 group remains in the K groups, and the remaining 1 group is used as verification data;

[0106] Furthermore, the training data is used as the model optimization parameters, and the verification data is used as the model prediction results. The Pearson correlation coefficient r is used to measure the consistency between the predicted range and the actual range, and a reliability assessment of the short-term prediction and medium- and long-term prediction results is obtained;

[0107] Specifically, if the Pearson correlation coefficient r is close to 1, it indicates that the vibration risk factor range predicted by the model is very close to the actual value, with high reliability, that is, the verification data is basically consistent with the actual range; if the Pearson correlation coefficient r is far from 1, it indicates that the model prediction is inconsistent with the actual situation, that is, the verification data differs greatly from the actual range, and adjustment is required by optimizing the model parameters.

[0108] Further, this embodiment also provides an anti-seismic prediction system for cultural relic display cabinets based on big data analysis, including:

[0109] A cultural relic display cabinet data processing module, configured to obtain the physical parameters and historical vibration data of the cultural relic display cabinet, preprocess the physical parameters and historical vibration data of the cultural relic display cabinet, and generate a weighted multi-dimensional dynamic time series matrix;

[0110] A cultural relic display cabinet vibration risk generation module, configured to initialize a deep learning model, send the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimize the deep learning model, and generate the vibration risk factor of the cultural relic display cabinet;

[0111] A cultural relic display cabinet vibration risk prediction module, configured to perform anti-seismic prediction on the cultural relic display cabinet with the vibration risk factor, divide the prediction results into short-term prediction and medium- and long-term prediction, and ensure the accuracy of the anti-seismic prediction of the cultural relic display cabinet by performing a reliability assessment on the above two prediction results.

[0112] This embodiment also provides a computer device applicable to the case of the anti-seismic prediction method for cultural relic display cabinets based on big data analysis, including:

[0113] A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the anti-seismic prediction method for cultural relic display cabinets based on big data analysis as proposed in the above embodiment.

[0114] The computer device may be a terminal, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be implemented through WIFI, carrier networks, NFC (Near Field Communication), or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0115] This embodiment also provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method for predicting the earthquake resistance of cultural relic display cabinets based on big data analysis as proposed in the above embodiment.

[0116] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. For technical details not described in detail in this embodiment, reference can be made to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0117] Embodiment 2

[0118] Refer to Figure 3 and Figure 4 , which is the second embodiment of the present invention. This embodiment provides a method for predicting the earthquake resistance of cultural relic display cabinets based on big data analysis, including: in order to verify the accuracy and reliability of the vibration risk factors in the solution of the present invention in prediction, a systematic experiment is conducted by comparing with the existing prediction method based on the traditional rule model, and the focus of the experiment is on the accuracy and stability of short-term and medium- and long-term vibration prediction;

[0119] The hardware environment is deployed as follows:

[0120] Data acquisition device: Triaxial acceleration sensor (model: ADXL345, accuracy ±0.002g);

[0121] Data processing device: Server (CPU: Intel Xeon Gold 6230, memory: 128GB, GPU: NVIDIA Tesla V100 32GB);

[0122] The software environment is deployed as follows:

[0123] Deep learning framework: PyTorch 1.10;

[0124] Data analysis tools: Python 3.9, using Pandas, NumPy, and Matplotlib libraries;

[0125] Traditional model: Rule-based threshold grading method (for comparison);

[0126] The test data includes the physical parameters of the cultural relic showcase: load (100 kg), material (tempered glass), connection point stiffness (50 N / mm); vibration historical data: vibration acceleration, period, and frequency within 3 days, sampling frequency 100 Hz;

[0127] The experimental process is as follows:

[0128] First, standardize the historical vibration data, process the vibration sequence using the sliding window technique (window size 20 steps, step size 5), normalize the physical parameters to generate static feature vectors, combine the dynamic time series matrix with the static feature vectors to form a weighted multi-dimensional dynamic time series matrix; subsequently, initialize the deep learning model, the input layer accepts the weighted dynamic time series, the convolutional layer is set to 15 convolutional kernels (size 5×5), the pooling layer uses max pooling, the output layer has a single neuron, and the activation function is Tanh, which is used to generate the vibration risk factor; finally, divide the experiment into two parts: short-term prediction and medium- and long-term prediction; use the Pearson correlation coefficient to evaluate the consistency between the prediction result and the actual value, and in addition, evaluate the model prediction performance through non-linear fitting, input the same data set into the rule-based threshold grading model, and compare the prediction results;

[0129] The experimental results are as follows:

[0130] From Figure 3 it can be seen that in the short-term prediction, the true value shows a gradually increasing trend with the sample index (time), representing the change law of the vibration risk factor, that is, the target value predicted by the model. Among them, the method of the present invention (deep learning model) reduces the fluctuation of the true value compared with the traditional method (rule model) and is relatively smooth, showing good fitting ability; referring to Figure 4 , in the medium- and long-term prediction, the deep learning fitting results show obvious advantages in periodic trend fitting, local change capture, and noise processing, and the predicted value is highly consistent with the true trend, and the average error is greatly reduced. However, the traditional rule model is weak in dealing with complex periodic changes and cannot meet the accuracy requirements of medium- and long-term prediction;

[0131] In summary, the present invention realizes the accurate prediction of the vibration risk of cultural relic showcases through a deep learning model, overcomes the limitations of traditional rule models, and provides more reliable technical support for risk assessment and prevention.

[0132] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The solutions in the embodiments of the present application can be implemented in various computer languages. For example, object-oriented programming languages such as Java and interpreted scripting languages such as JavaScript, etc.

[0133] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0134] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0135] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, so that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0136] Although the preferred embodiments of the present application have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once they learn of the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present application.

[0137] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. A method for earthquake-proof prediction of cultural relics display cabinets based on big data analysis, characterized in that: include: Acquiring physical parameters and historical vibration data of the cultural relics display cabinet, preprocessing the physical parameters and historical vibration data of the cultural relics display cabinet, and generating a weighted multi-dimensional dynamic time series matrix; Initializing a deep learning model, sending the weighted multidimensional dynamic time series matrix into the deep learning model, optimizing the deep learning model, and generating a vibration risk factor for the cultural relics display cabinet; The earthquake risk factor is used to make earthquake-proof predictions for the cultural relics display cabinets, and the prediction results are divided into short-term predictions and medium- and long-term predictions. The reliability of the above two prediction results is evaluated to ensure the accuracy of the earthquake-proof predictions for the cultural relics display cabinets.

2. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis according to claim 1, characterized in that: The physical parameters and historical vibration data of the cultural relics display cabinet are obtained, and the physical parameters and historical vibration data of the cultural relics display cabinet are preprocessed to generate a weighted multi-dimensional dynamic time series matrix, including: The physical parameters of the cultural relics display case are taken as static data, and the historical vibration data of the cultural relics display case is taken as dynamic data; The static data is processed in a normalized manner to obtain a static feature vector, and the static feature vector is used as the global information of the time series matrix throughout all time steps; The historical vibration data of the cultural relic display cabinet is indexed by the time step to construct an original time series matrix, and each column of data is subjected to sliding window standardization processing with the column of the original time series matrix as a unit to generate a dynamic time series matrix after standardization processing; Taking each time step as a benchmark, the static feature vectors and the dynamic feature vectors in the dynamic time series matrix are merged to obtain multiple multidimensional feature vectors, and the contribution of each multidimensional feature vector is calculated to obtain a weight list of the multidimensional feature vectors. The weight list of the multidimensional feature vectors is multiplied by the standardized dynamic time series matrix to obtain a weighted multidimensional dynamic time series matrix.

3. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis as claimed in claim 2, characterized in that: Also includes: Periodic labels are added to the weighted multidimensional dynamic time series matrix to capture periodic changes of the cultural relics display cabinet during vibration, and the first-order derivative of the acceleration of the cultural relics display cabinet during vibration is considered to capture the changes in velocity and displacement of the cultural relics display cabinet during vibration.

4. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis as claimed in claim 2, characterized in that: Initialize the deep learning model, including: Initialize the dimension of the input layer; Initialize the convolution kernel in the convolution layer. The convolution kernel size is composed of the number of time windows and the number of features. Initialize the pooling window size in the pooling layer; Initialize the weight matrix and bias vector in the fully connected layer; Initialize the output layer to a single neuron and the activation function to the Tanh function.

5. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis according to claim 2 or 4, characterized in that: The weighted multi-dimensional dynamic time series matrix is ​​sent to the deep learning model, and the deep learning model is optimized to generate the vibration risk factor of the cultural relics display cabinet, including: The input layer sets its own input dimension to the same value as the dimension of the weighted multidimensional dynamic time series matrix, and sends the multidimensional dynamic time series matrix to the convolution layer, and the convolution layer determines the number of features and the number of time windows of the convolution kernel according to the number of feature vectors and the number of time steps in the weighted multidimensional dynamic time series matrix; The corresponding feature map is obtained through the convolution kernel, and the feature map is sent to the pooling layer for dimensionality reduction and the pooling window size is set to the maximum pooling. Then the reduced-dimensional feature map is sent to the fully connected layer, and the fully connected layer obtains a weight matrix according to the reduced-dimensional feature map; The reduced-dimensional feature map is multiplied by the weight matrix, and a bias vector is added to obtain an output vector of the fully connected layer. The output vector is sent to the output layer to output the vibration risk factor of the cultural relic display cabinet.

6. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis according to claim 5 is characterized in that: The earthquake risk factors are used to predict the earthquake resistance of the cultural relics display cabinets, and the prediction results are divided into short-term prediction and medium- and long-term prediction, including: To make short-term and medium-term forecasts of the shock risk factors in a time series manner; When the vibration risk factor is between [0, 0.3) and remains in this range in a future time window, it means that the cultural relic display case will not be affected by the vibration and there is no need to adjust the cultural relics in the display case in advance; When the vibration risk factor is between [0.3, 0.7) and remains in this range in a future time window, it means that the cultural relic display case will vibrate slightly but will not affect the cultural relics themselves, and the stability of the cultural relic display case needs to be strengthened in advance; When the vibration risk factor is ≥ 0.7 and remains within this range in multiple future time windows, it means that the cultural relic display cabinet will experience a large vibration and will affect the cultural relics themselves, and it is necessary to transfer the cultural relics in the cultural relic display cabinet in advance or reinforce the cultural relic display cabinet; Among them, short-term prediction refers to a time window in the future, and long-term prediction refers to multiple time windows in the future.

7. The earthquake-proof prediction method for cultural relics display cabinets based on big data analysis according to claim 2 or 6, characterized in that: The reliability of the above two prediction results is evaluated, including: Divide each time window into K groups, and take K-1 groups as training data each time, until only one group is left in the K groups, and use the remaining group as verification data; The Pearson correlation coefficient r is used to measure the consistency between the predicted range and the actual range, and the reliability evaluation results of the two prediction results are obtained.

8. A system for earthquake-proof prediction of cultural relics display cabinets based on big data analysis, based on the earthquake-proof prediction method of cultural relics display cabinets based on big data analysis according to any one of claims 1 to 7, characterized in that: include: The cultural relic display cabinet data processing module is configured to obtain physical parameters and historical vibration data of the cultural relic display cabinet, pre-process the physical parameters and historical vibration data of the cultural relic display cabinet, and generate a weighted multi-dimensional dynamic time series matrix; The cultural relics display cabinet vibration risk generation module is configured to initialize the deep learning model, input the weighted multi-dimensional dynamic time series matrix into the deep learning model, optimize the deep learning model, and generate the vibration risk factor of the cultural relics display cabinet; The cultural relics display cabinet vibration risk prediction module is configured to perform earthquake-proof prediction on the cultural relics display cabinet based on the vibration risk factor, and divide the prediction results into short-term prediction and medium- and long-term prediction, and ensure the accuracy of earthquake-proof prediction on the cultural relics display cabinet by performing reliability evaluation on the above two prediction results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images