Configurable far-field long-period ground motion intelligent identification system based on multi-domain features

Abstract

The application discloses a configurable far-field long-period seismic motion intelligent identification system based on multi-domain features and belongs to the technical field of earthquake engineering and structural seismic analysis. In view of the problems of high labor cost, strong subjectivity, single index misjudgment and criterion inadaptability to different structural requirements of the existing far-field long-period seismic motion screening, the system realizes automatic analysis, standardized processing and multi-domain feature extraction of seismic motion data, adopts a layered discrimination mechanism of weighted comprehensive scoring combined with far-field conditions, period core conditions and auxiliary long-period conditions, supports self-defined configuration of an identification threshold, and covers the whole process from data import to report export. The application greatly improves screening efficiency and identification accuracy and can adapt to the seismic analysis requirements of various long-period structures.

Description

Technical Field

[0001] This invention relates to the fields of earthquake engineering, strong ground motion data processing and structural seismic analysis, specifically a configurable far-field long-period ground motion intelligent identification system based on multi-domain features. Background Technology

[0002] With the rapid development of building and transportation infrastructure construction, the application scale of long-period sensitive structures such as high-rise buildings, super high-rise buildings, long-span bridges, seismic isolation structures, and large-span spatial structures continues to expand. Far-field long-period ground motions are characterized by long propagation distances, a high proportion of low-frequency (long-period) vibration energy, and significant duration. They are prone to resonance effects with long-period structures, causing large structural displacement responses and cumulative damage. Therefore, in structural seismic design, dynamic time history analysis, and performance-based seismic assessment, it is necessary to accurately select far-field long-period ground motion records that are suitable for the requirements of the target structure.

[0003] Currently, the screening of far-field long-period ground motions largely relies on manual operation: engineers need to manually download ground motion records from public strong ground motion databases such as PEER, manually organize metadata such as the number of sampling points, sampling interval, acceleration unit, magnitude, and epicentral distance, calculate characteristic parameters using third-party software or self-written programs, and then determine whether it is a far-field long-period ground motion based on personal experience. This method has drawbacks such as cumbersome processing procedures, low batch processing efficiency, and inconsistent judgment results due to the influence of subjective experience.

[0004] Existing automatic identification methods have significant limitations: one type of method relies solely on metadata such as magnitude and epicentral distance to determine far-field attributes, but far-field ground motions do not necessarily possess significant long-period characteristics; another type of method uses only single indicators such as dominant period, PGV / PGA ratio, or significant duration to determine long-period attributes, which are easily affected by factors such as ground motion spectrum distribution and site conditions, leading to misjudgments and omissions. Furthermore, existing methods often employ fixed identification thresholds, which cannot adapt to the differentiated needs of different types of long-period structures regarding the scope of long-period focus and feature thresholds, thus limiting their applicability.

[0005] In addition, most existing ground motion processing software only supports single functions such as waveform display, response spectrum calculation, and parameter output. They lack a closed-loop system that covers the entire process from ground motion data import, file parsing, standardization processing, feature extraction, automatic discrimination to result output, making it difficult to meet the actual needs of efficient screening of batch ground motions in engineering scenarios. Summary of the Invention

[0006] To address the problems existing in the background technology, this invention proposes a configurable far-field long-period ground motion intelligent identification system based on multi-domain features. This system solves the problems of cumbersome manual operation processes, low batch processing efficiency, inconsistent judgment results due to subjective experience, susceptibility to misjudgments or omissions due to single indicators or metadata-based judgments, inability of fixed thresholds to adapt to the differentiated screening needs of different types of long-period structures, and the lack of full-process closed-loop functionality and weak engineering applicability of existing processing tools. This invention achieves fully automated identification from ground motion data import to result export, resulting in a highly accurate, flexibly configurable, and adaptable far-field long-period ground motion screening service that meets the needs of various long-period structures.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A configurable far-field long-period ground motion intelligent identification system based on multi-domain features includes: The ground motion data import module is used to receive the raw ground motion data to be identified; The ground motion file parsing module is used to parse the raw data of the ground motion to be identified, and obtain sampling parameters, acceleration units, acceleration time history and time series corresponding to the acceleration time history; The seismic motion normalization module is used to generate a normalized time history dataset; The multi-domain feature extraction module is used to extract various types of ground motion feature parameters in the time domain, frequency domain, and response spectrum domain from the standardized time history dataset. The standard configuration module is used to configure the set of recognition thresholds for far-field long-period ground motion. The far-field long-period discrimination module is used to determine the condition compliance of the ground motion to be identified based on the ground motion characteristic parameters and the set of identification thresholds, calculate a comprehensive score, and output the identification result.

[0008] Specifically, the execution flow of the seismic motion standardization processing module is as follows: (1) Remove invalid and outlier data from the acceleration time history to obtain the valid acceleration time history; (2) Based on the conversion relationship between acceleration units and international standard acceleration units, the units of effective acceleration time history are uniformly converted to international standard acceleration units; (3) The acceleration time history after unit unification is processed by removing the mean and removing the linear trend in sequence to obtain the acceleration time history after baseline correction; (4) The acceleration time history after baseline correction is numerically integrated to obtain the velocity time history, and the displacement time history is numerically integrated to obtain the displacement time history. The acceleration time history, velocity time history and displacement time history after baseline correction together constitute the standardized time history dataset.

[0009] Specifically, the time-domain features extracted by the multi-domain feature extraction module include peak ground acceleration, peak ground velocity, peak ground displacement, the ratio of peak ground velocity to peak ground acceleration, and significant duration; the frequency-domain features extracted include Fourier dominant period and low-frequency energy proportion; and the response spectrum domain features extracted include response spectrum dominant period, long-period spectral area, and long-period energy proportion.

[0010] Specifically, the significant duration is calculated based on the Arias intensity and is the time difference between the time point when the Arias intensity reaches a first preset proportion of the total Arias intensity and the time point when the Arias intensity reaches a second preset proportion of the total Arias intensity.

[0011] Specifically, the formula for calculating the proportion of long-period energy is as follows: ; in, For the proportion of long-cycle energy, For period The corresponding pseudo-velocity response spectrum value, ~ To pre-set a long-term period, ~ To analyze the periodic interval.

[0012] Specifically, the recognition standard configuration module has several sets of recognition thresholds corresponding to different application scenarios, and supports custom adjustment of each threshold in the recognition threshold set.

[0013] Specifically, the execution rules of the far-field long-period discrimination module are as follows: (1) Determine whether the ground motion to be identified meets the far-field condition; (2) Determine whether the ground motion to be identified meets the core periodic condition. If it does not meet the condition, output the non-far-field long-period ground motion identification result. (3) Count the number of auxiliary long-period conditions satisfied by the ground motion to be identified; (4) The scores of the sub-items corresponding to the far-field condition, the core period condition, and each auxiliary long-period condition are weighted and integrated to obtain a comprehensive score; If the ground motion to be identified simultaneously meets the far-field condition, the core period condition, and at least one auxiliary long-period condition, and the comprehensive score reaches the comprehensive score threshold, then the far-field long-period ground motion identification result is output; if the period parameter of the ground motion to be identified is close to the core period condition threshold and meets the preset number of auxiliary long-period conditions, then the boundary-type record identification result is output; otherwise, the non-far-field long-period ground motion identification result is output.

[0014] Specifically, the far-field condition is: the magnitude of the ground motion to be identified is greater than or equal to the minimum far-field magnitude in the identification threshold set, and the observation distance of the ground motion to be identified is greater than or equal to the minimum far-field distance in the identification threshold set. The core condition for the period is that the dominant period of the response spectrum or the dominant period of the Fourier transform of the ground motion to be identified is greater than or equal to the dominant period threshold in the identification threshold set. The auxiliary long-period conditions include three categories: the ratio of peak ground velocity to peak ground acceleration is greater than or equal to the peak ground velocity to peak ground acceleration ratio threshold in the identification threshold set; the long-period energy proportion is greater than or equal to the long-period energy proportion threshold in the identification threshold set; and the significant duration is greater than or equal to the significant duration threshold in the identification threshold set.

[0015] Specifically, the formula for calculating the comprehensive score is as follows: ; in, For comprehensive scoring, The scores are for the sub-items corresponding to the far-field conditions. The scores are the sub-items corresponding to the core conditions of the cycle. The score corresponds to the ratio of peak ground velocity to peak ground acceleration. The sub-scores correspond to the proportion of long-cycle energy. For the sub-scores corresponding to significant duration, ~ These are the corresponding weighting coefficients.

[0016] Specifically, the system also includes a result visualization and file generation module and a graphical user interface (GUI) module; The results visualization and file generation module is used to display standardized acceleration time history, velocity time history, displacement time history, pseudo-velocity response spectrum, pseudo-acceleration response spectrum, summary table of seismic motion characteristic parameters, judgment condition compliance status, comprehensive score, judgment basis and engineering significance interpretation, and generate and export multi-domain characteristic parameter files and recognition report files. The multi-domain feature parameter file records peak ground acceleration, peak ground velocity, peak ground displacement, the ratio of peak ground velocity to peak ground acceleration, significant duration, Fourier dominant period, response spectrum dominant period, low-frequency energy proportion, long-period energy proportion, and long-period spectrum area; the identification report file records basic earthquake information, information on the set of identification thresholds used, identification results, judgment criteria, and threshold settings; The graphical user interface (GUI) module provides an interactive operation entry point for the other modules. It includes a data input area, a seismic information input area, an identification standard configuration area, a threshold adjustment area, a time history display area, a response spectrum display area, a parameter display area, a discrimination result display area, and a report export trigger control.

[0017] In summary, the beneficial technical effects of the present invention are as follows: 1. Reduced labor costs and improved screening efficiency: This invention integrates a complete closed-loop function from automatic parsing of ground motion files, standardization processing, multi-domain feature extraction, automatic judgment to result output. It eliminates the need for manual organization of ground motion metadata and manual calculation of feature parameters, enabling rapid screening of batch ground motions. This significantly improves efficiency compared to traditional manual screening methods, while avoiding inconsistencies in judgment results caused by differences in human experience, resulting in higher reliability of identification results.

[0018] 2. Reduced risk of misjudgment and missed judgment: This invention extracts three types of core feature parameters in the time domain, frequency domain, and response spectrum domain, comprehensively covering the temporal vibration attributes, spectral distribution attributes, and structural dynamic response attributes of ground motion. It avoids the limitations of relying solely on metadata such as magnitude / distance or a single long-period indicator for judgment, and can comprehensively reflect the far-field and long-period characteristics of ground motion. The recognition accuracy is significantly improved compared to the single indicator judgment method.

[0019] 3. Meets the actual needs of screening long-period structures: The far-field long-period discrimination module of this invention adopts a four-layer progressive discrimination rule of "far-field condition verification → hard threshold verification of core period conditions → auxiliary long-period condition verification → comprehensive score verification", so as to prioritize the output of far-field long-period ground motions that meet the strict conditions.

[0020] 4. Adapting to diverse screening needs across multiple scenarios: The standard configuration module of this invention can be built with recognition thresholds for various types of structures, such as high-rise buildings, super high-rise buildings, long-span bridges, and seismic isolation structures. It also supports user-defined adjustment of various judgment thresholds, which can flexibly adapt to the long-term sensitivity range of different structural types, as well as the judgment requirements of different scenarios such as conventional engineering design and rigorous scientific research analysis. The applicable scenario coverage is higher than that of existing methods that use fixed thresholds. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the system framework of the present invention; Figure 2 This is a schematic diagram of the batch recognition results and result filtering interface according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the interface for batch result export, report packaging, single record selection, and result recognition in an embodiment of the present invention; Figure 4This is an explanation of the engineering significance of the embodiments of the present invention, as well as a schematic diagram showing the time history of acceleration, velocity, and displacement; Figure 5 This is a schematic diagram illustrating the pseudo-velocity response spectrum and pseudo-acceleration response spectrum according to an embodiment of the present invention; Figure 6 This is a schematic diagram showing the recognition results and multi-domain feature parameters of an embodiment of the present invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention clearer and easier to understand, the invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0023] Example like Figure 1 As shown in this embodiment, the configurable far-field long-period ground motion intelligent identification system based on multi-domain features includes a ground motion data import module, a ground motion file parsing module, a ground motion standardization processing module, a multi-domain feature extraction module, an identification standard configuration module, a far-field long-period discrimination module, a result visualization and file generation module, and a graphical user interface (GUI) module. It is applied to the screening of single or batch far-field long-period ground motions in scenarios such as seismic design of high-rise buildings, performance evaluation of long-span bridges, and dynamic analysis of seismic isolation structures. The specific implementation process of each module is as follows: The seismic ground motion data import module receives the selected raw seismic ground motion data to be identified through the data input area of ​​the graphical user interface (GUI), supporting the import of three types of data sources: (1) Importing publicly available strong ground motion database format records: Select to import AT2 / TXT format files from publicly available strong ground motion databases such as PEER. Upload one or more AT2 / TXT format files in batches using the file selection control in the GUI. The system will then transfer the uploaded files to the ground motion file parsing module one by one.

[0024] (2) Importing a table-format time history file: Select to import a custom time history file in CSV format. After uploading the CSV file, the system will automatically read all column names of the CSV file and display them in the GUI. The user selects the corresponding time column, acceleration column and acceleration unit. After confirmation, the system will transmit the file and column selection information to the ground motion file parsing module.

[0025] (3) Import of simulated ground motion data generated by the system: In system testing or functional demonstration scenarios, the system generates simulated acceleration time histories by using non-stationary envelope, multi-frequency components and smooth random disturbances based on the duration, sampling interval, target dominant period, target peak ground acceleration and random seed set by the user, and scales them according to the target peak ground acceleration; the same parameters and random seed can generate the same results, and the simulated data is only used for system testing and functional demonstration.

[0026] The seismic motion file parsing module receives the raw data and configuration information transmitted by the seismic motion data import module, and uses corresponding parsing logic for different types of raw data to extract sampling parameters, acceleration units, acceleration time histories and corresponding time series; For AT2 / TXT format files, the system performs parsing as follows: (1) File header field recognition: Read the file header text content line by line, and identify the NPTS and DT fields by keyword matching. NPTS is the total number of sampling points of the ground motion record, and DT is the sampling interval. For PEER AT2 / TXT format records, the system reads the acceleration time history in the file in g units.

[0027] (2) Acceleration time history extraction: After skipping the header line of the file, read all acceleration values ​​in the subsequent data area and store them as the original acceleration time history in one-dimensional array format.

[0028] (3) Time series generation: Based on the sampling interval DT and the total number of sampling points NPTS, a time series corresponding one-to-one with the acceleration time history is generated. The calculation formula for the time series is: ; in, For the first The time corresponding to each sampling point The sampling point number, The sampling interval is... This represents the number of sampling points; (4) Parameter output: Pack and transmit the obtained sampling point number NPTS, sampling interval DT, acceleration unit, original acceleration time history and time series to the ground motion standardization processing module.

[0029] For CSV format files, the system performs parsing as follows: (1) Column data extraction: Based on the “time column” and “acceleration column” selected in the data import stage, extract the corresponding column data to generate a time series array and a raw acceleration time history array.

[0030] (2) Sampling parameter calculation: The average value of the difference between two adjacent time points is calculated as the sampling interval DT, and the total length of the statistical time series is calculated as the total number of sampling points NPTS. When there is a significant inconsistency between adjacent time intervals, the system performs linear interpolation on the time series and acceleration time history according to the sampling interval to form an equally spaced time history.

[0031] (3) Unit labeling: The user is prompted to confirm the unit of acceleration in the GUI pop-up window. The user can select from g, m / s², etc. 2 cm / s 2After selecting the corresponding unit in gal, the system marks the selected acceleration unit.

[0032] (4) Parameter output: Pack and transmit the obtained sampling point number NPTS, sampling interval DT, acceleration unit, original acceleration time history and time series to the ground motion standardization processing module.

[0033] The seismic motion normalization module receives the raw acceleration time history, time series, sampling parameters, and acceleration unit information output by the seismic motion file parsing module, and performs normalization processing according to the following process to generate a normalized time history dataset: (1) Removal of invalid and abnormal data: Traverse the original time series and acceleration time history, remove null values, non-numerical data and infinite values, and retain valid data with finite values ​​for both time and acceleration.

[0034] (2) Unit conversion: Based on the identified acceleration units, the units of the effective acceleration time history are converted to the international standard acceleration unit m / s². 2 .

[0035] (3) Baseline correction: The acceleration time history after unit unification is subjected to mean removal and linear trend removal processing in sequence. Mean removal processing is to calculate the arithmetic mean of the entire acceleration time history, and subtract the arithmetic mean from the acceleration value of each sampling point to eliminate the constant offset error of the strong vibration observation instrument. Linear trend removal processing is to fit the linear trend term of the acceleration time history using the least squares method, and subtract the trend term value at the corresponding time from the acceleration value of each sampling point to eliminate the linear offset error caused by instrument baseline drift, and obtain the baseline corrected acceleration time history.

[0036] (4) Integrating to generate velocity and displacement time histories: The acceleration time histories after baseline correction are numerically integrated for the first time to obtain the velocity time histories and remove their linear trends; then the velocity time histories are numerically integrated for the second time to obtain the displacement time histories and remove their linear trends. Finally, the baseline-corrected acceleration time histories, velocity time histories, and displacement time histories together constitute a standardized time histories dataset, which is then passed to the multi-domain feature extraction module.

[0037] The multi-domain feature extraction module receives a standardized time-history dataset and extracts various types of seismic motion feature parameters from three dimensions: time domain, frequency domain, and response spectrum domain. Specifically, these include: (1) Temporal feature extraction; The time-domain characteristics include five items: peak ground acceleration, peak ground velocity, peak ground displacement, the ratio of peak ground velocity to peak ground acceleration, and significant duration. Peak ground acceleration (PGA) is the maximum absolute value of acceleration at all sampling points in the normalized acceleration time history, and is calculated using the following formula: ; in, For a moment The corresponding standardized acceleration value.

[0038] Peak ground velocity (PGV) is the maximum absolute value of the velocities at all sampling points in the velocity time history, and is calculated using the following formula: ; in, For a moment The corresponding speed value.

[0039] Peak ground displacement (PGD) is the maximum absolute value of the displacement at all sampling points in the displacement time history, and is calculated using the following formula: ; in, For a moment The corresponding displacement value.

[0040] The ratio of peak ground velocity to peak ground acceleration is denoted as... It is used to characterize the relative proportion of long-period components in earthquake motion.

[0041] Significant duration is calculated based on Arias intensity. First, the cumulative Arias intensity time history is calculated using the following formula: ; in, for The cumulative Arias intensity over time. It is the acceleration due to gravity. The integral time variable is used; the total Arias intensity is the cumulative Arias intensity value at the end of the earthquake duration. Significant duration is the time difference between the time point when the Arias intensity reaches the first preset proportion of the total Arias intensity and the time point when the Arias intensity reaches the second preset proportion of the total Arias intensity.

[0042] (2) Frequency domain feature extraction; Frequency domain characteristics include Fourier dominant periodicity There are two items: low-frequency energy ratio; The Fourier amplitude spectrum is obtained by performing a Fast Fourier Transform (FFT) on the standardized acceleration time history. The frequency corresponding to the peak value of the amplitude spectrum is found, and the Fourier dominant period is the reciprocal of that frequency. The low-frequency energy ratio is the ratio of the sum of the Fourier spectrum energies in the frequency range of 0 to 0.5 Hz to the sum of the Fourier spectrum energies in all positive frequency ranges, where the frequency range corresponds to low-frequency components with a period of not less than 2 s.

[0043] (3) Extraction of reaction spectral domain features; Reaction spectral characteristics include the dominant period of the reaction spectrum. Three items: long-period spectral area, long-period energy percentage, etc. Based on a 5% damping ratio, 120 period points were selected within the analysis period range of 0.1s to 10.0s to calculate the pseudo-velocity response spectrum of a single-degree-of-freedom linear system under different natural vibration periods; the period corresponding to the peak value of the pseudo-velocity response spectrum is the dominant period of the response spectrum.

[0044] The long-period spectral area is the integral area of ​​the pseudo-velocity response spectrum within the long-period range of 2.0 s to 10.0 s, and is calculated using the following formula: ; in, For long-period spectral area, ~ It has a long period range of 2.0s to 10.0s.

[0045] The long-period energy proportion is the ratio of the area of ​​the long-period spectrum from 2.0s to 10.0s to the total integrated area of ​​the pseudo-velocity response spectrum within the analysis period range of 0.1s to 10.0s. The calculation formula is as follows: ; in, For the proportion of long-cycle energy, For period The corresponding pseudo-velocity response spectrum value, ~ The period range is from 2.0s to 10.0s. ~ The analysis period range is from 0.1s to 10.0s.

[0046] All extracted multi-domain feature parameters are packaged and synchronously transmitted to the far-field long-period discrimination module and the result visualization and file generation module.

[0047] The standard configuration module is used to configure the threshold set for identifying far-field long-period ground motions. The system has multiple preset threshold sets corresponding to different application scenarios. Each threshold set includes seven types of thresholds: far-field minimum magnitude, far-field minimum distance, dominant period threshold, peak ground velocity to peak ground acceleration ratio threshold, long-period energy proportion threshold, significant duration threshold, and comprehensive score threshold. Application scenarios include: (1) Default scenario: Applicable to general long-period structure screening scenarios. As an example configuration, the thresholds are set as follows: minimum far-field magnitude 6.5, minimum far-field distance 100km, dominant period threshold 2.0s, and the ratio of peak ground velocity to peak ground acceleration. Threshold 0.15s, long-cycle energy proportion threshold 0.35, significant duration threshold 20s, comprehensive score threshold 0.6; (2) High-rise building scenario: Applicable to the seismic analysis screening of high-rise buildings with a height of 100m~250m; (3) Super high-rise building scenario: Applicable to the seismic analysis screening of super high-rise buildings with a height of 250m or more; (4) Long-span bridge scenario: Applicable to the seismic analysis screening of long-span bridges with a main span of more than 300m; (5) Seismic isolation structure scenario: Applicable to seismic analysis and screening of buildings and bridges using foundation isolation technology; (6) Scientific research screening scenario: Applicable to scientific research analysis scenarios with strict requirements.

[0048] The preset threshold values ​​in the above scenarios can be adjusted according to actual engineering needs or research objectives.

[0049] The system allows users to customize one or more thresholds in the recognition threshold set. After the adjustment is completed, when a single recognition or batch recognition is triggered again, the system will use the updated threshold to perform the judgment without restarting the system.

[0050] The far-field long-period discrimination module receives seismic motion feature parameters output by the multi-domain feature extraction module, the set of recognition thresholds output by the recognition standard configuration module, and basic earthquake information input by the user, including magnitude and observation distance, and performs discrimination according to the following hierarchical logic: S1, far-field condition determination; The magnitude input by the user is compared with the minimum far-field magnitude, and the observation distance is compared with the minimum far-field distance. If both the magnitude and the observation distance are not lower than the corresponding thresholds, the far-field condition is met. The score for each far-field condition is the average of the normalized magnitude score and the normalized observation distance score. Each normalized score is the ratio of the actual value to the corresponding threshold and is limited to the range of 0 to 1.

[0051] S2, Determination of core conditions for the cycle; The larger of the dominant period of the response spectrum and the dominant period of the Fourier transform is used as the periodicity index, which is compared with the dominant period threshold. If the periodicity index is not lower than the threshold, the core periodicity condition is met; if the periodicity index is not lower than 90% of the dominant period threshold but lower than the dominant period threshold, the boundary periodicity condition is met. The core periodicity score is the ratio of the periodicity index to the dominant period threshold, and is limited to the range of 0 to 1.

[0052] S3, Auxiliary long-period condition determination; The satisfaction status of the three types of auxiliary long-period conditions is determined sequentially: (1) Condition for the ratio of peak ground velocity to peak ground acceleration: If Not lower than the set of recognition thresholds If the threshold is met, the condition is satisfied, and the sub-score is the ratio of the actual ratio to the corresponding threshold, limited to the range of 0 to 1; (2) Long-cycle energy ratio condition: If the long-cycle energy ratio is not lower than the corresponding threshold, then this condition is met; its sub-score is the ratio of the actual ratio to the corresponding threshold, and is limited to the range of 0 to 1. (3) Significant duration condition: If the significant duration is not less than the corresponding threshold, then the condition is met; its sub-score is the ratio of the actual significant duration to the corresponding threshold, and is limited to the range of 0 to 1; The number of auxiliary long-period conditions that are satisfied is statistically analyzed.

[0053] S4. Calculation of overall score; The scores for each sub-item are weighted and merged according to preset weights to obtain the comprehensive score. The calculation formula is as follows: ; in, The score is a comprehensive rating, ranging from 0 to 1. The scores are for the sub-items corresponding to the far-field conditions. The scores are the sub-items corresponding to the core conditions of the cycle. The score corresponds to the ratio of peak ground velocity to peak ground acceleration. The sub-scores correspond to the proportion of long-cycle energy. For the sub-scores corresponding to significant duration, the value range of each sub-score is 0~1; ~ These are the preset weighting coefficients for each sub-item's score. In this embodiment... to The weights are 0.25, 0.25, 0.20, 0.20, and 0.10 respectively, with the sum of each weight being 1.

[0054] S5. Final recognition result output; Based on the judgment result, the output is divided into three categories: (1) If the far-field condition, the periodic core condition, and at least one auxiliary long-period condition are met at the same time, and the comprehensive score is not lower than the comprehensive score threshold, then the far-field long-period ground motion identification result is output. (2) If the aforementioned strict judgment conditions are not met, but the far-field conditions are met, and the larger value of the dominant period of the response spectrum and the dominant period of the Fourier is in the range of 90% to 100% of the dominant period threshold, and at least two auxiliary long period conditions are met, then the far-field long period boundary type record identification result is output. (3) For other cases, output the non-far-field long-period ground motion identification results. The condition compliance status, sub-item score, comprehensive score, and judgment basis of the judgment process are synchronously transmitted to the result visualization and file generation module.

[0055] The results visualization and document generation module receives standardized time history datasets, multi-domain feature parameters, discrimination process data and final recognition results, and performs visualization and document generation functions. The following content is displayed in the corresponding area of ​​the graphical user interface: like Figure 4 As shown, (1) Time history curve display area: draws line graphs of standardized acceleration time history, velocity time history and displacement time history, and displays the engineering significance of the current record; like Figure 5 As shown, (2) Reaction spectrum display area: draw line graphs of pseudo velocity response spectrum and pseudo acceleration response spectrum, and mark the corresponding positions of the dominant period of the response spectrum; like Figure 6 As shown, (3) Feature parameter display area: all extracted multi-domain feature parameters are displayed in tabular form, including peak ground acceleration, peak ground velocity, peak ground displacement, ratio of peak ground velocity to peak ground acceleration, significant duration, Fourier dominant period, response spectrum dominant period, low frequency energy ratio, long period spectrum area, and long period energy ratio. like Figure 3 and Figure 6 As shown, (4) Judgment result display area: Displays three final labels: far-field long-period ground motion, far-field long-period boundary type record, and non-far-field long-period ground motion, and displays the compliance status, corresponding threshold, continuous sub-item score and comprehensive score of each judgment condition in tabular form; at the same time, it generates an engineering significance interpretation based on the current identification standard preset, response spectrum dominant period, Fourier dominant period, peak ground velocity to peak ground acceleration ratio, long-period energy ratio, significant duration and comprehensive score.

[0056] The system supports users triggering file generation via an export control, generating two types of downloadable files: (1) Multi-domain feature parameter file: stored in CSV format. At a minimum, the following information should be recorded: file name, event name, magnitude, distance type, observation distance, peak ground acceleration, peak ground velocity, peak ground displacement, ratio of peak ground velocity to peak ground acceleration, significant duration, Fourier dominant period, response spectrum dominant period, low-frequency energy proportion, long-period spectrum area, long-period energy proportion, final identification label, risk level, comprehensive score, and preset identification criteria. Batch identification results should also record the data source, sampling parameters, processing status, and error reasons.

[0057] (2) Identification report file: TXT, PDF or Word format is optional. The content includes file name, data source, event name, magnitude, distance type, observation distance, identification standard and thresholds used, multi-domain characteristic parameters, judgment condition compliance status, continuous sub-item score, comprehensive score, final identification result, judgment basis, explanation of engineering significance, as well as acceleration time history, velocity time history, displacement time history, pseudo-velocity response spectrum and pseudo-acceleration response spectrum charts.

[0058] like Figure 2 and Figure 3 As shown, the graphical user interface (GUI) module adopts a Streamlit-based web interface to provide operation entry points for each functional module. The interface is set with fixed functional sections, and the operation logic of each section is as follows: (1) Data input area: includes file upload control and simulated ground motion generation control, supports users to upload AT2 / TXT and CSV format files in batches, or input simulated ground motion parameters to generate test data; (2) Earthquake information input area: It provides input or display controls for event name, magnitude, observation distance, and distance type. Distance types include fault rupture distance (Rrup), epicentral distance (Repi), focal distance (Rhyp), Joyner-Boore distance (Rjb), and other observation distances. For PEER records, the magnitude, Rrup, and event name can be automatically matched by searching the report. It also supports entering earthquake information through batch metadata tables. (3) Identify the standard configuration area: It includes a drop-down selection box for preset identification standards, which includes default standards, high-rise buildings, super high-rise buildings, long-span bridges, seismic isolation structures, and strict scientific research screening. It also includes threshold adjustment controls, allowing users to select presets or manually adjust various thresholds. (4) Results display area: It is divided into time history display sub-area, response spectrum display sub-area, characteristic parameter display sub-area, and discrimination result display sub-area, which supports users to switch and view the analysis results of different ground motions; (5) Report Export Area: Includes feature parameter export button and recognition report export button, supports users to select export file format and export range (current single record / all batch data).

[0059] Therefore, the configurable far-field long-period ground motion intelligent identification system based on multi-domain features provided by this invention adopts a closed-loop logic throughout the entire process, including automatic parsing of ground motion files, signal standardization preprocessing, multi-domain feature fusion extraction in the time domain, frequency domain, and response spectrum domain, configurable criterion matching for multiple scenarios, hierarchical progressive intelligent discrimination, and result visualization export. It also employs a three-category feature cross-validation, a strict target-first and boundary-type supplementary judgment order, and a threshold mechanism with multiple scenario presets and custom adjustments. This solves the problems of low efficiency, strong subjectivity, easy misjudgment and omission of single indicators, and the inability of fixed thresholds to adapt to the differentiated needs of various long-period structures in traditional manual screening. It improves batch screening efficiency, judgment consistency, and result interpretability, reduces manual costs and reliance on experience, and provides ground motion screening support for seismic analysis and performance evaluation of long-period sensitive structures such as high-rise buildings, super high-rise buildings, long-span bridges, and seismic isolation structures.

[0060] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A configurable far-field long-period ground motion intelligent identification system based on multi-domain features, characterized in that, include: The ground motion data import module is used to receive the raw ground motion data to be identified; The ground motion file parsing module is used to parse the raw data of the ground motion to be identified, and obtain sampling parameters, acceleration units, acceleration time history and time series corresponding to the acceleration time history; The seismic motion normalization module is used to generate a normalized time history dataset; The multi-domain feature extraction module is used to extract various types of ground motion feature parameters in the time domain, frequency domain, and response spectrum domain from the standardized time history dataset. The standard configuration module is used to configure the set of recognition thresholds for far-field long-period ground motion. The far-field long-period discrimination module is used to determine the condition compliance of the ground motion to be identified based on the ground motion characteristic parameters and the set of identification thresholds, calculate a comprehensive score, and output the identification result.

2. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 1, characterized in that, The execution flow of the seismic motion standardization processing module is as follows: (1) Remove invalid and outlier data from the acceleration time history to obtain the valid acceleration time history; (2) Based on the conversion relationship between acceleration units and international standard acceleration units, the units of effective acceleration time history are uniformly converted to international standard acceleration units; (3) The acceleration time history after unit unification is processed by removing the mean and removing the linear trend in sequence to obtain the acceleration time history after baseline correction; (4) The acceleration time history after baseline correction is numerically integrated to obtain the velocity time history, and the displacement time history is numerically integrated to obtain the displacement time history. The acceleration time history, velocity time history and displacement time history after baseline correction together constitute the standardized time history dataset.

3. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 1, characterized in that, The time-domain features extracted by the multi-domain feature extraction module include peak ground acceleration, peak ground velocity, peak ground displacement, the ratio of peak ground velocity to peak ground acceleration, and significant duration; the frequency-domain features extracted include Fourier dominant period and low-frequency energy proportion; and the response spectrum domain features extracted include response spectrum dominant period, long-period spectral area, and long-period energy proportion.

4. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 3, characterized in that, The significant duration is calculated based on the Arias intensity and is the time difference between the time point when the Arias intensity reaches a first preset proportion of the total Arias intensity and the time point when the Arias intensity reaches a second preset proportion of the total Arias intensity.

5. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 3, characterized in that, The formula for calculating the proportion of long-period energy is: ； in, For the proportion of long-cycle energy, For period The corresponding pseudo-velocity response spectrum value, ~ To pre-set a long-term period, ~ To analyze the periodic interval.

6. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 1, characterized in that, The recognition standard configuration module has several sets of recognition thresholds for different application scenarios, and supports custom adjustment of each threshold in the recognition threshold set.

7. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 1, characterized in that, The execution rules of the far-field long-period discrimination module are as follows: (1) Determine whether the ground motion to be identified meets the far-field condition; (2) Determine whether the ground motion to be identified meets the core periodic condition. If it does not meet the condition, output the non-far-field long-period ground motion identification result. (3) Count the number of auxiliary long-period conditions satisfied by the ground motion to be identified; (4) The scores of the sub-items corresponding to the far-field condition, the core period condition, and each auxiliary long-period condition are weighted and integrated to obtain a comprehensive score; If the ground motion to be identified simultaneously meets the far-field condition, the period core condition, and at least one auxiliary long-period condition, and the comprehensive score reaches the comprehensive score threshold, then the far-field long-period ground motion identification result is output; if the dominant period of the response spectrum or the dominant period of the Fourier transform of the ground motion to be identified does not meet the period core condition, but is within the preset proportion range of the dominant period threshold in the set of identification thresholds, and simultaneously meets the preset number of auxiliary long-period conditions, then the boundary-type record identification result is output; otherwise, the non-far-field long-period ground motion identification result is output.

8. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 7, characterized in that, The far-field condition is: the magnitude of the ground motion to be identified is greater than or equal to the minimum far-field magnitude in the identification threshold set, and the observation distance of the ground motion to be identified is greater than or equal to the minimum far-field distance in the identification threshold set. The core condition for the period is that the dominant period of the response spectrum or the dominant period of the Fourier transform of the ground motion to be identified is greater than or equal to the dominant period threshold in the identification threshold set. The auxiliary long-period conditions include three categories: the ratio of peak ground velocity to peak ground acceleration is greater than or equal to the peak ground velocity to peak ground acceleration ratio threshold in the identification threshold set; the long-period energy proportion is greater than or equal to the long-period energy proportion threshold in the identification threshold set; and the significant duration is greater than or equal to the significant duration threshold in the identification threshold set.

9. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 7, characterized in that, The formula for calculating the comprehensive score is as follows: ； in, For comprehensive scoring, The sub-scores are for the far-field conditions. The scores are the sub-items corresponding to the core conditions of the cycle. The score corresponds to the ratio of peak ground velocity to peak ground acceleration. The sub-scores correspond to the proportion of long-cycle energy. For the sub-scores corresponding to significant duration, ~ These are the corresponding weighting coefficients.

10. The configurable far-field long-period ground motion intelligent identification system based on multi-domain features according to claim 1, characterized in that, The system also includes a result visualization and document generation module and a graphical user interface (GUI) module; The results visualization and file generation module is used to display standardized acceleration time history, velocity time history, displacement time history, pseudo-velocity response spectrum, pseudo-acceleration response spectrum, summary table of seismic motion characteristic parameters, judgment condition compliance status, comprehensive score, judgment basis and engineering significance interpretation, and generate and export multi-domain characteristic parameter files and recognition report files. The multi-domain feature parameter file records peak ground acceleration, peak ground velocity, peak ground displacement, the ratio of peak ground velocity to peak ground acceleration, significant duration, Fourier dominant period, response spectrum dominant period, low-frequency energy proportion, long-period energy proportion, and long-period spectrum area; the identification report file records basic earthquake information, information on the set of identification thresholds used, identification results, judgment criteria, and threshold settings; The graphical user interface (GUI) module provides an interactive operation entry point for the other modules. It includes a data input area, a seismic information input area, an identification standard configuration area, a threshold adjustment area, a time history display area, a response spectrum display area, a parameter display area, a discrimination result display area, and a report export trigger control.

Images