Wavelet basis function selection method and device for seismic velocity pulse recognition

Abstract

The invention discloses a wavelet basis function selection method and device for ground vibration speed pulse recognition, and the method comprises the steps: carrying out the integral calculation of ground vibration record data, and obtaining an energy concentration interval of an original speed time history through calculation; performing iterative analysis on the original speed time history by using different wavelet basis functions in sequence to obtain time history distribution data and an energy concentration interval corresponding to each wavelet basis function; performing down-sampling on the energy concentration interval corresponding to each wavelet basis function to obtain a down-sampled interval of each wavelet basis function; calculating a first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original speed time history; calculating a second correlation coefficient between the speed pulse time history identified by each wavelet basis function and the original speed time history; and selecting a wavelet basis function by using the size sorting of the first correlation coefficient and the second correlation coefficient. According to the invention, the fidelity of speed pulse recognition can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intersection of earthquake engineering and signal processing, and in particular to a method and device for selecting a wavelet basis function for seismic velocity pulse identification. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. No statement herein is admitted to be prior art by virtue of its inclusion in this section.

[0003] Ground motion velocity pulses are common seismic characteristics of strong earthquakes, typically occurring in near-fault ground motions and particularly prominent in the fault hanging wall. Velocity pulses manifest as transient, high-amplitude velocity variations caused by fault slip and the concentrated propagation of energy toward the surface. These pulses exhibit significant low-frequency characteristics and a short duration, yet they significantly influence the dynamic response of engineering structures. In recent years, with the increasing frequency of earthquakes, analysis of strong motion records has revealed that some near-fault ground motions contain velocity pulses of interest in seismology and earthquake engineering. These pulses possess abundant medium- and long-period components, large pulse peaks, and significant engineering hazards, making their identification and study crucial.

[0004] The appearance of velocity pulses is often associated with directional effects near faults. This effect can lead to a significant enhancement of the velocity pulse in a specific direction. In addition, although velocity pulses often appear in near-fault records, not all near-fault ground motions contain velocity pulse characteristics. In some cases, even in the absence of significant fault slip effects, similar velocity pulse phenomena may be observed due to local site effects or other nonlinear seismic responses. However, the hidden nature of velocity pulse characteristics makes them difficult to identify through intuitive methods, which poses a challenge to the accurate detection of pulses.

[0005] Various methods for velocity pulse identification and extraction have emerged in the prior art. For example, one method based on empirical mode decomposition extracts pulse characteristics by decomposing the eigenmode functions of seismic signals; one method employs the Hilbert-Huang transform (HHT) combined with empirical mode decomposition and Hilbert spectrum analysis to perform instantaneous frequency analysis on non-stationary signals; and one method employs wavelet analysis to identify velocity pulses. This method offers computational efficiency advantages for processing large-scale records and facilitates the establishment of a unified database. Wavelet analysis, widely used for its computational advantages in standardized identification, primarily involves selecting the Daubechies wavelet (db4) as the mother wavelet, performing a continuous wavelet transform on the velocity time history, and extracting the wavelet with the largest coefficient absolute value. The ratio of the peak ground velocity (PGV) before and after pulse extraction and the ratio of the cumulative energy are then used as indicators of pulse characteristics. A single pulse index is established using regression analysis; a larger index value indicates a more pronounced pulse characteristic.

[0006] Despite progress, existing methods still have some shortcomings. For example, the impact of different wavelet bases on pulse recognition has not been systematically studied. In practical applications, wavelet basis functions are usually selected based on expert experience. However, different wavelet basis functions have different characteristics, and the choice of wavelet basis significantly affects the effectiveness of velocity pulse recognition, resulting in a need to improve the fidelity of velocity pulse recognition. Summary of the Invention

[0007] An embodiment of the present invention provides a wavelet basis function selection method for seismic velocity pulse identification, which is used to assist in selecting appropriate wavelet basis functions in seismic velocity pulse identification and improve the fidelity of velocity pulse identification. The method includes:

[0008] Obtaining earthquake motion record data;

[0009] Integrate the earthquake motion data and calculate the energy concentration interval of the original velocity history;

[0010] Iteratively analyzing the original velocity time history using different wavelet basis functions in sequence to obtain time history distribution data and energy concentration interval corresponding to each wavelet basis function; the energy concentration interval represents a preset percentage of the total energy;

[0011] Downsampling the energy concentration interval corresponding to each wavelet basis function to obtain the downsampled interval of each wavelet basis function; wherein the data density of the downsampled interval of each wavelet basis function is the same as the energy concentration interval of the original velocity time history;

[0012] Calculate the first correlation coefficient between the downsampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history;

[0013] Using each wavelet basis function in turn to identify the velocity pulse time history of the original velocity time history, and calculating the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history;

[0014] The wavelet basis function is selected by sorting the first correlation coefficient and the second correlation coefficient.

[0015] An embodiment of the present invention further provides a wavelet basis function selection device for seismic velocity pulse identification, which is used to assist in selecting appropriate wavelet basis functions in seismic velocity pulse identification and improve the fidelity of velocity pulse identification. The device includes:

[0016] A seismic record data acquisition module, used to acquire seismic record data;

[0017] The original velocity time history processing module is used to perform integration calculation on the earthquake motion record data and calculate the energy concentration interval of the original velocity time history;

[0018] A wavelet basis function processing module is used to iteratively analyze the original velocity time history using different wavelet basis functions in sequence to obtain time history distribution data and energy concentration interval corresponding to each wavelet basis function; the energy concentration interval represents a preset percentage of the total energy;

[0019] A downsampling processing module is used to downsample the energy concentration interval corresponding to each wavelet basis function to obtain a downsampled interval of each wavelet basis function; wherein the data density of the downsampled interval of each wavelet basis function is the same as the energy concentration interval of the original velocity time history;

[0020] The correlation calculation module is used to calculate the first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original speed time history; use each wavelet basis function in turn to identify the speed pulse time history of the original speed time history, and calculate the second correlation coefficient between the speed pulse time history identified by each wavelet basis function and the original speed history; and select the wavelet basis function based on the order of the first correlation coefficient and the second correlation coefficient.

[0021] An embodiment of the present invention further provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the wavelet basis function selection method for ground motion velocity pulse identification is implemented.

[0022] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for selecting wavelet basis functions for ground motion velocity pulse identification is implemented.

[0023] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method for selecting wavelet basis functions for ground motion velocity pulse identification is implemented.

[0024] In the embodiment of the present invention, a first correlation coefficient is calculated between the downsampled energy concentration interval corresponding to each wavelet basis function and the energy concentration interval of the original velocity time history. A second correlation coefficient is calculated between the velocity pulse time history identified by each wavelet basis function and the original velocity time history. The correlation between different wavelet basis identification results and the original velocity time history is utilized to optimize the selection of wavelet basis functions, thereby assisting in selecting appropriate wavelet basis functions in seismic velocity pulse identification. This effectively improves the accuracy of velocity pulse identification, ensures that the extracted pulse features are more consistent with the physical characteristics of actual seismic motion, and enhances the fidelity of velocity pulse identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0026] Figure 1 Schematic diagram of the flow of a method for selecting a wavelet basis function for ground velocity pulse identification according to an embodiment of the present invention;

[0027] Figure 2 A diagram showing a specific example of a method for selecting a wavelet basis function for ground velocity pulse identification according to an embodiment of the present invention;

[0028] Figure 3 FIG. 1 is another specific example of a method for selecting a wavelet basis function for identifying ground velocity pulses according to an embodiment of the present invention;

[0029] Figure 4 FIG. 1 is another specific example of a method for selecting a wavelet basis function for identifying ground velocity pulses according to an embodiment of the present invention;

[0030] Figure 5 The velocity pulse time course identified by multiple wavelet basis functions in the embodiment of the present invention is Figure 1 ;

[0031] Figure 6 The velocity pulse time course identified by multiple wavelet basis functions in the embodiment of the present invention is Figure 2 ;

[0032] Figure 7Schematic diagram of a wavelet basis function selection device for ground velocity pulse identification according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0034] In order to clearly describe the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, words such as "first" and "second" are used to distinguish between identical or similar items with basically the same functions and effects. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order.

[0035] Although existing methods for identifying seismic velocity pulses have made some progress, they still have some shortcomings. For example, the impact of different wavelet basis functions on pulse identification has not been systematically studied, and the choice of wavelet basis function may significantly affect the recognition effect. In addition, the impact of the directional effect of velocity pulses on the recognition results also needs further exploration. Based on strong vibration recording data, the embodiment of the present invention uses wavelet analysis to study the recognition effect of velocity pulses, explores the directionality of velocity pulse identification, and proposes an improved wavelet method based on the correlation between the wavelet basis and the velocity time history to further improve the accuracy and reliability of velocity pulse identification.

[0036] Figure 1 FIG. 1 is a flow chart of a method for selecting a wavelet basis function for ground velocity pulse identification according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0037] Step 101: Acquire earthquake motion record data;

[0038] Step 102: perform integration calculation on the earthquake motion record data and calculate the energy concentration interval of the original velocity time history;

[0039] Step 103: Iteratively analyze the original velocity time history using different wavelet basis functions in sequence to obtain time history distribution data and energy concentration interval corresponding to each wavelet basis function; the energy concentration interval represents a preset percentage of the total energy;

[0040] Step 104: downsample the energy concentration interval corresponding to each wavelet basis function to obtain a downsampled interval of each wavelet basis function; wherein the downsampled interval of each wavelet basis function has the same data density as the energy concentration interval of the original velocity time history;

[0041] Step 105: Calculate the first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history;

[0042] Step 106: Use each wavelet basis function in turn to identify the velocity pulse time history of the original velocity time history, and calculate the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history;

[0043] Step 107: Select a wavelet basis function by sorting the first correlation coefficient and the second correlation coefficient.

[0044] Below Figure 1 The wavelet basis function selection method for seismic velocity pulse identification is explained in detail.

[0045] Step 101 first obtains earthquake record data.

[0046] Raw seismic motion data refers to the ground motion data directly recorded by seismic monitoring instruments when an earthquake occurs without any artificial processing or modification. It is presented in the form of a time series and records the changes in physical quantities such as displacement, velocity or acceleration of ground motion at different times during the earthquake.

[0047] The seismic recording data acquired in the embodiment of the present invention mainly includes acceleration time history data.

[0048] Step 102 integrates and calculates the recorded seismic data to obtain an energy concentration interval for the original velocity time series. This energy concentration interval represents the range within which the energy of the velocity time series, filtered from the original velocity time series, reaches a predetermined range of the total energy of the original velocity time series. Specifically, in this embodiment of the present invention, the energy concentration interval is the shortest interval that accounts for a predetermined percentage of the total energy.

[0049] For example, using the trapezoidal method to integrate the velocity history, recorded as the original velocity history, according to E = v 2 (t) / 2 calculates the energy, draws the energy curve, and determines the energy concentration range.

[0050] Figure 2 FIG. 1 is a specific example of a method for selecting a wavelet basis function for ground velocity pulse identification according to an embodiment of the present invention. Figure 2 As shown, the method includes:

[0051] Step 201: Baseline correction is performed on the earthquake record data, and low-frequency trend data in the earthquake record data is removed by filtering to obtain processed earthquake record data; the frequency of the low-frequency trend data is lower than a preset value;

[0052] Step 202: Perform a secondary integration on the processed earthquake motion record data to obtain the original velocity time history;

[0053] Step 203: Using the sliding spectrum window method, perform energy accumulation calculation on the original velocity time history to obtain the energy concentration area of ​​the original velocity time history.

[0054] During implementation, the ground motion recording data is first baseline corrected, the low-frequency trend in the signal is removed by filtering, and then the original velocity time history is obtained by secondary integration.

[0055] In a preferred embodiment, to further reduce drift, the least squares method is used to perform baseline adjustment, ultimately obtaining a corrected velocity time history. Baseline correction is performed on the seismic record data, and low-frequency trend data in the seismic record data is removed by filtering to obtain processed seismic record data. Specifically, the following steps may be performed:

[0056] The least square method is used to adjust the baseline. Based on the adjusted baseline, the low-frequency trend data is removed to obtain the processed seismic motion record data.

[0057] In one embodiment, the energy accumulation calculation is performed on the original velocity time history using a sliding spectrum window method to obtain the energy concentration area of ​​the original velocity time history, which may include:

[0058] The original velocity time history is slid with a preset sliding window width and a preset sliding step length, and the signal energy is calculated segment by segment to generate a cumulative energy curve;

[0059] According to the cumulative energy curve, the time region where the total energy reaches a preset range is extracted as the energy concentration region of the original speed time course.

[0060] For example, the sliding spectrum window method is used to calculate the cumulative energy of the signal, and the shortest time interval [t1, t2] with 90% energy is selected through the cumulative energy curve. This interval is used to represent the energy concentration area of ​​the original velocity time history.

[0061] Specifically, calculating the cumulative energy of a signal using the sliding window method may include:

[0062] (1) Determine the sliding window parameters: According to the sampling frequency and target resolution of the original velocity time history signal, set the sliding window width and sliding step size to ensure a balance between time resolution and computational efficiency. The target resolution refers to the desired analysis accuracy in the time or frequency domain. Determining the target resolution requires a compromise between time resolution and frequency resolution.

[0063] (2) Segmented signal processing: The original velocity time history signal is divided into several time periods with the set sliding window width as the unit. Each signal segment covers the range of the sliding window, and the signal energy is calculated segment by segment.

[0064] (3) Calculate the signal energy within the window: For each speed signal within the sliding window, calculate the energy according to the following formula:

[0065]

[0066] Among them, E i is the energy calculation result in the window during the i-th sliding, v(t) is the velocity time signal, t i and t i+1 The start and end time of the current sliding window.

[0067] (4) Generate cumulative energy curve: Accumulate the energy calculated by all sliding windows to generate the cumulative energy curve of the signal, which is used to reflect the energy distribution characteristics of the signal over time.

[0068] (5) Determine the energy concentration interval: By analyzing the cumulative energy curve, extract the shortest time interval [t1, t2] where the total energy reaches 90% as the energy concentration area of ​​the original speed time history, providing a basis for subsequent steps.

[0069] In step 103, different wavelet basis functions are used in turn to iteratively analyze the original velocity time history to obtain time history distribution data and energy concentration interval corresponding to each wavelet basis function; the energy concentration interval represents a preset percentage of the total energy.

[0070] When implemented, step 103 includes performing the following processing for each wavelet basis function:

[0071] (1) Setting the wave basis function processing parameters, which include the number of iterations and the scale parameter;

[0072] (2) performing continuous wavelet transform on the original velocity time history signal and decomposing the signal to obtain the time history distribution data corresponding to each wavelet basis function; wherein the wavelet coefficients are generated when the signal is decomposed;

[0073] (3) Calculate the cumulative energy curve of the decomposed wavelet coefficients using the sliding spectrum window method;

[0074] (4) According to the cumulative energy curve, determine the energy concentration interval corresponding to each wavelet basis function.

[0075] For example, 14 wavelet basis functions are selected for time-history iterative wavelet analysis, as follows:

[0076] (1) Selection of wavelet basis functions: 14 common wavelet basis functions (such as Daubechies, Symlets, etc.) are selected, and time-history iterative wavelet analysis is performed for each wavelet basis function.

[0077] (2) Setting parameters: The number of iterations for each wavelet basis function is set to 10 to ensure that the calculation accuracy meets the analysis requirements while taking into account both calculation efficiency and result stability. More than 10 iterations can meet the accuracy requirements. There is basically no difference in numerical value between the results of 10 and 20 iterations. The only difference is in accuracy, that is, the higher the number of iterations, the higher the accuracy of the results. Since the embodiment of the present invention mainly focuses on the numerical results of the wavelet basis functions and has a general requirement for accuracy, 10 iterations are used in the embodiment of the present invention.

[0078] (3) Decomposition of time-course signals: For wavelet transform, the selected wavelet basis function is used to decompose the signal and extract its multi-scale time-course distribution characteristics.

[0079] (4) Calculate the cumulative energy: For the decomposed wavelet coefficients, use the sliding spectrum window method to calculate its cumulative energy curve and analyze the distribution of energy over time.

[0080] (5) Extracting energy duration intervals: According to the cumulative energy curve, the shortest time interval [t1', t2'] containing 90% of the energy is selected as the energy concentration area for wavelet basis analysis.

[0081] (6) Repeating steps: Repeat the above analysis steps for all 14 wavelet basis functions to generate the time distribution characteristics and energy duration interval of each wavelet basis function.

[0082] Among them, by performing wavelet transformation in the interval [t1', t2'], the velocity time history signal value after wavelet basis transformation can be obtained.

[0083] In step 104, the energy concentration interval corresponding to each wavelet basis function is downsampled to obtain the downsampled interval of each wavelet basis function; wherein the downsampled interval of each wavelet basis function has the same data density as the energy concentration interval of the original velocity time history.

[0084] During implementation, considering that the wavelet basis function has a higher time resolution and denser data points in the time domain transform (for example, 1024 Hz), while the sampling frequency of the earthquake motion record is relatively low (for example, 200 Hz or 50 Hz), that is, the number of nodes covered by the interval [t1, t2] is not as dense as that of the interval [t1', t2']. Therefore, the time intervals [t1, t2] and [t1', t2'] are equally divided and processed so that the number of nodes in the interval [t1, t2] is equal to the number of nodes in the interval [t1', t2'] after the equal division, thereby achieving consistency of data dimensions.

[0085] In one embodiment, downsampling the energy concentration interval corresponding to each wavelet basis function to obtain the downsampled interval of each wavelet basis function may include:

[0086] According to the ratio of the time resolution of the wavelet basis function transformation result to the sampling frequency of the earthquake motion record, the ratio of the wavelet basis function transformation result to be downsampled is determined;

[0087] According to the time range difference and density difference between the energy concentration interval of the original velocity time history and the energy concentration interval corresponding to the wavelet basis function, the number of nodes that need to be retained in the energy concentration interval corresponding to the wavelet basis function is determined;

[0088] According to the ratio of downsampling required for the wavelet basis function transformation result, the nodes on the energy concentration interval corresponding to the wavelet basis function are uniformly downsampled, and the number of nodes required to be retained is retained at equal intervals.

[0089] During implementation, considering that the wavelet basis function has a high time resolution in the time domain transform, in order to match the ground motion record, the specific downsampling process includes:

[0090] (1) Analyze the time resolution difference: Calculate the difference between the time resolution of the wavelet basis function transform result (e.g., 1024 Hz) and the sampling frequency of the earthquake motion record (e.g., 200 Hz or 50 Hz), and determine the scaling factor that needs to be downsampled.

[0091] The downsampling scaling factor can be calculated using the following formula:

[0092] The downsampling scale factor = the time resolution of the wavelet basis function transformation result / the sampling frequency of the earthquake motion record.

[0093] For example, the sampling frequency of the earthquake record is 200 Hz, and the sampling frequency of the wavelet transform is 1024 Hz, which means that the signal after the wavelet transform needs to be downsampled by a ratio of about 1024 / 200=5.12.

[0094] (2) Select the target number of nodes: Based on the time range and density difference between the intervals [t1, t2] and [t1', t2'], calculate the number of nodes that need to be retained in [t1', t2'] so that it is consistent with the number of nodes on [t1, t2].

[0095] For example, calculate as follows:

[0096] 1. Calculate the time range Δt1 = t2 - t1 for the interval [t1, t2].

[0097] 2. Calculate the time range Δt2 = t′2 - t′1 for the interval [t′1, t′2].

[0098] 3. Calculate the number of nodes on the interval [t′1, t′2]: N2 = f2 × Δt2.

[0099] 4. Divide the interval [t1, t2] evenly so that the number of nodes is N2, that is, each time interval is

[0100] 5. Generate new nodes according to equal interval division to complete the adjustment of time series data.

[0101] (3) Uniform downsampling: Uniformly downsample the nodes in the interval [t1', t2'], retaining the target nodes at equal intervals according to the calculated scaling factor, thereby adjusting the density of the time series. In this step, the interval [t1, t2] is divided equally so that the number of nodes in this interval is consistent with the time interval [t1', t2'] after the wavelet transform, solving the problem of inconsistent dimensions of time series data at different time resolutions.

[0102] (4) Signal smoothing: To reduce information loss caused by downsampling, the velocity time-history signal is low-pass filtered before downsampling to remove high-frequency components and avoid aliasing. In this step, linear interpolation is performed on the node sequence after averaging over the interval [t1, t2] to obtain a new velocity time-history signal. This ensures that the signal maintains smoothness and consistency over the new node sequence.

[0103] (5) Generate downsampled data: Output the downsampled time series data, ensuring that the number of its nodes is consistent with the number of nodes on [t1, t2], thereby achieving data dimension matching.

[0104] In step 105 , a first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history is calculated; the first correlation coefficient reflects the correlation between the wavelet basis and the original velocity time history.

[0105] In the embodiment of the present invention, the correlation coefficient between the transformed velocity time history and the original velocity time history is the Pearson correlation coefficient. The first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history can be calculated according to the following formula:

[0106]

[0107] Among them, R is the first correlation coefficient, which ranges from [-1,1] and is used to measure the correlation between two sequences;

[0108] x i is the original speed time history signal value after the energy concentration interval of the original speed time history is evenly divided;

[0109] y i is the velocity time history signal value after wavelet basis transformation on the down-sampling interval of the wavelet basis function;

[0110] and x i 、y i The average value of

[0111] n is the number of corresponding nodes.

[0112] The Pearson correlation coefficient evaluates the strength and direction of the linear relationship between two variables by measuring the ratio of their covariance to their standard deviation. The main goal of the formula is to provide a correlation measure that is comparable across different datasets and fields by standardizing the covariance and eliminating unit and scale differences.

[0113] Among them, R=1 indicates a complete positive correlation (that is, the wavelet basis function has the highest fidelity), and R=0 indicates no linear correlation.

[0114] In step 106 , each wavelet basis function is used in turn to identify the velocity pulse time history of the original velocity time history, and a second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history is calculated.

[0115] During implementation, the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history can be calculated according to the following formula:

[0116]

[0117] Among them, r is the second correlation coefficient, v original (t) is the original velocity time history, v wavelet (t) is the velocity pulse time course identified by the wavelet basis function, v original The mean of (t), v wavelet The second correlation coefficient, r, reflects the degree of similarity between the velocity pulses identified by the wavelet basis function and the original velocity time history. A higher second correlation coefficient indicates that the identified pulses are more consistent with the original velocity time history in terms of trend, morphology, and other characteristics, indicating that the recognition effect of the wavelet basis function is likely to be more ideal.

[0118] molecular This part is the covariance, which represents the sum of the products of the deviations of the two signals at each moment. Covariance can measure whether two signals increase or decrease at the same time, or whether their changing trends are consistent. If the two signals change in the same direction (i.e., when one signal increases, the other signal also increases, or when one signal decreases, the other signal also decreases), the covariance will be positive. If the two signals change in opposite directions (i.e., when one signal increases, the other signal decreases), the covariance will be negative.

[0119] Denominator This component is the product of the standard deviation of the original velocity time-history signal and the wavelet-transformed signal. The standard deviation reflects the degree of fluctuation of each signal and normalizes the covariance, thus making the result of this second correlation coefficient independent of the signal scale. A large standard deviation indicates large signal fluctuations, and the calculated result of this second correlation coefficient will be subject to significant fluctuations. A smaller standard deviation indicates less signal variation, and the calculated result of this second correlation coefficient will be more stable.

[0120] The calculation formula of the second correlation coefficient is used to measure the original signal v original (t) and the wavelet transformed signal v wavelet The wavelet transform is a method that can perform analysis in both the time and frequency domains. By comparing the correlation between the original signal and the wavelet transformed signal, we can evaluate the effect of the wavelet transform on the signal or the effectiveness of signal feature extraction. If the correlation coefficient r approaches 1, it indicates a close relationship between the original signal and the wavelet transformed signal, indicating high fidelity. If r approaches 0, it indicates a weak or no significant linear relationship between them.

[0121] To improve evaluation accuracy, the present invention specifically selects the energy-concentrated interval of the original velocity time series when calculating correlations. This is the shortest time interval containing 90% of the energy. This effectively filters out high-frequency noise or subtle long-term variations, thereby focusing on the primary energy contribution of the velocity pulse. By calculating the correlation coefficient between the velocity pulse time series extracted by each wavelet basis and the original velocity time series within this interval, the fidelity of different wavelet basis functions in preserving signal characteristics can be more accurately quantified, thereby improving the stability and reliability of velocity pulse identification.

[0122] Step 107: Select a wavelet basis function by sorting the first correlation coefficient and the second correlation coefficient.

[0123] During implementation, the first correlation coefficients corresponding to the wavelet basis functions may be relatively close. When the first correlation coefficients are similar, the second correlation coefficients are continuously observed and the optimal wavelet basis function is selected according to their sizes.

[0124] Figure 3 FIG. 1 is another specific example of a method for selecting a wavelet basis function for identifying ground velocity pulses according to an embodiment of the present invention. Figure 3 As shown, the method includes:

[0125] (1) Data preprocessing: including baseline correction, velocity time course energy curve processing, and calculation of 90% energy interval using the sliding spectrum window method.

[0126] First, baseline correction is performed on the raw ground motion (primarily consisting of acceleration time history data). Low-frequency trends are removed from the signal through filtering, and then the velocity history is obtained by quadratic integration. To further reduce drift, the baseline is adjusted using the least squares method, ultimately resulting in the corrected velocity history. The cumulative energy of the signal is calculated using a sliding window method. The shortest time interval [t1, t2] with 90% energy is selected from the cumulative energy curve. This interval represents the energy concentration region of the raw velocity history.

[0127] (2) Wavelet basis selection and energy duration analysis: iterative calculation of 14 wavelet bases, wavelet basis energy curve processing, and sliding spectrum window method to calculate wavelet basis energy interval.

[0128] Fourteen wavelet basis functions were selected and all were iterated 10 times to obtain the time domain distribution of the wavelet basis function. The sliding window method was also used to select the shortest time interval [t1', t2'] with 90% energy duration.

[0129] (3) Data consistency adjustment: wavelet basis and velocity time history data density analysis, dividing the interval to ensure dimensional consistency.

[0130] Considering that the wavelet basis function has a higher time resolution and denser data points (1024 Hz) in the time domain transform, while the sampling frequency of the earthquake motion record is relatively low (200 Hz or 50 Hz), that is, the number of nodes covered by the interval [t1, t2] is not as dense as that of the interval [t1', t2'], the time interval [t1, t2] is divided equally so that the number of nodes in the interval [t1, t2] is equal to the number of nodes in the interval [t1', t2'] after the equal division, thereby achieving consistency in data dimensions.

[0131] (4) Interpolation and correlation calculation: signal interpolation and smoothing processing, calculation of the correlation between wavelet basis and velocity time history, and calculation of the correlation between pulse time history and velocity time history.

[0132] The node sequence evenly divided in the interval [t1, t2] is piecewise linearly interpolated on the original velocity time history to obtain a new velocity time history signal.

[0133] Finally, the correlation between the nodes in the interval [t1, t2] and the interval [t1', t2'] is calculated to obtain the first correlation coefficient.

[0134] The first correlation coefficient is calculated repeatedly by changing different wavelet basis functions to obtain a sorted list of correlation coefficients under different wavelet basis functions.

[0135] The second correlation coefficient between the velocity pulse time history identified by different wavelet basis functions and the original velocity time history is calculated to quantify the fidelity of each wavelet basis function in extracting the velocity pulse time history.

[0136] (5) Comparison and result output: Repeat the wavelet basis type output results, generate the correlation coefficients of various types of wavelet basis, and select the optimal wavelet basis function.

[0137] The correlation calculation results of two earthquake motion data under 14 different types of wavelet basis functions are shown in Table 1. It can be seen that for different earthquake motions with velocity pulse characteristics, different wavelet basis functions have different recognition effects on velocity pulses.

[0138] For the earthquake motion data numbered 01, the correlation coefficient of the sym7 wavelet basis function reaches 0.6951, which shows the best recognition effect. Figure 4 FIG. 1 is another specific example of a method for selecting a wavelet basis function for ground velocity pulse identification according to an embodiment of the present invention. Figure 4 The upper middle figure (1) is the time domain interval with 90% energy duration corresponding to the sym7 wavelet basis function. Figure 4 The lower middle figure (2) is the time interval with 90% energy duration corresponding to the original speed time history. It can be seen that the change trends of the upper and lower figures are highly consistent.

[0139] For the earthquake motion data numbered 02, the db7 wavelet basis function has the best recognition effect.

[0140] Table 1 The first correlation coefficients between different wavelet bases and original velocity time history

[0141]

[0142] The above analysis demonstrates that different wavelet basis functions provide varying recognition results for different ground motions. The db4 wavelet basis function, recommended in existing Baker identification methods, performs well for extracting some velocity pulse records. However, the shape of some velocity pulse records cannot be identified using the fourth-order Daubechies method. Therefore, in the embodiments of the present invention, the appropriate wavelet basis function is selected based on the calculation of correlation coefficients before using wavelet analysis for velocity pulse identification. This provides an efficient and accurate method for effectively identifying velocity pulses.

[0143] In order to further improve the accuracy of speed pulse recognition and explore the performance of different wavelet basis functions in speed pulse recognition, the embodiment of the present invention also quantifies the fidelity of each wavelet basis function in extracting the speed pulse time history by calculating the second correlation coefficient between the speed pulse time history identified by different wavelet basis functions and the original speed time history, thereby providing a scientific basis for the selection of wavelet basis functions.

[0144] In the embodiment of the present invention, 10 commonly used wavelet basis functions are used to identify the velocity pulse of the strong vibration record data No. 01, and the second correlation coefficient is calculated. The calculation results are shown in Table 2 below.

[0145] Table 2 The second correlation coefficient of the identified velocity pulse and the original velocity time history of different wavelet bases

[0146] Serial number Wavelet basis corr indicator is_pulse Tp(s) 1 db1 0.8681 0.9999 1 5.7875 2 db2 0.8542 1.0000 1 9.6150 3 db3 0.8656 1.0000 1 8.2813 4 db4 0.8444 0.9997 1 8.9740 5 db5 0.8218 0.9994 1 9.6150 6 db6 0.8260 0.9994 1 8.8138 7 db7 0.8122 0.9988 1 8.3272 8 db8 0.8082 0.9982 1 8.6475 9 db9 0.8099 0.9987 1 8.4717 10 db10 0.8128 0.9986 1 8.7400

[0147] Where corr represents the correlation coefficient;

[0148] indicator represents the pulse indicator, which takes values ​​between 0 and 1. Records with scores higher than 0.85 and lower than 0.15 are classified as pulse and non-pulse, respectively;

[0149] is_pulse indicates whether it is a speed pulse, 1 for yes, 0 for no;

[0150] Tp(s) represents the speed pulse period.

[0151] As shown in Table 2, the db1 wavelet basis function has the highest second-order correlation coefficient between velocity pulses and velocity histories, at 0.8681. From a correlation perspective, db1 is considered the wavelet basis function with the best recognition performance in this experiment.

[0152] Figure 5 The velocity pulse time course identified by multiple wavelet basis functions in the embodiment of the present invention is Figure 1 , Figure 6 The velocity pulse time course identified by multiple wavelet basis functions in the embodiment of the present invention is Figure 2 , Figure 5 、 Figure 6 The velocity pulse time histories identified by the four wavelet basis functions db1, db3, db8, and db9 in Table 2 are shown. It can be seen that the correlation coefficients for db1 and db3 are similar and high, while the correlation coefficients for db8 and db9 are low. This shows that different wavelet bases have significant differences in capturing pulse characteristics.

[0153] refer to Figure 5 、 Figure 6 In the pulse time history diagrams identified by different wavelet basis functions, db1 is the best choice for capturing short-term pulses, db3 is more suitable for capturing oscillation characteristics, and db8 and db9 are more suitable for long-term pulse signals. Figure 5 、 Figure 6A graphical comparison clearly shows that as the wavelet order increases (e.g., from DB3 to DB9), the pulse time history diagram becomes smoother, better representing the changing trend over long time scales. In contrast, lower-order wavelet basis functions (e.g., DB1) focus more on short-timescale details and better capture instantaneous pulse characteristics. Furthermore, a comparison of the extracted pulse signal and the residual signal reveals less fluctuation in the DB1 residual signal, indicating that the DB1 wavelet transform is able to better capture the key pulse characteristics. This is also corroborated by the large correlation coefficient between velocity pulses and velocity time histories identified by DB1.

[0154] Therefore, in a specific embodiment, different wavelet basis functions are used to identify velocity pulses in different scenarios. The order of the wavelet basis function is proportional to the time scale. That is, a lower-order wavelet basis function (such as db1) is used to identify short-time-scale velocity pulses, and a higher-order wavelet basis function (such as db9) is used to identify long-time-scale velocity pulses. For example, the db1 function is used to identify short-time pulses, where short-time refers to a time less than a first set threshold, which is determined based on actual conditions. The db3 function is used to identify oscillation characteristics. The db8 and db9 functions are used to identify long-duration pulse signals, where long-duration refers to a time greater than a second set threshold, which is determined based on actual conditions.

[0155] In this embodiment of the present invention, different wavelet basis functions can effectively identify velocity pulses in some strong earthquake records. Selecting a wavelet basis function with a higher correlation with the original velocity time history is crucial for ensuring the accuracy of the identification results. In this study, db1 demonstrated a higher correlation in the correlation calculation, demonstrating superior identification performance, rather than the previously widely assumed db4. Through correlation analysis, this embodiment of the present invention can quantify the applicability of different wavelet basis functions.

[0156] This analysis method, based on time-history correlation, not only intuitively measures the applicability of different wavelet basis functions but also provides a scientific basis for selecting the appropriate wavelet basis function. By comparing and ranking the correlation coefficients of different wavelet basis functions, we can better understand which wavelet basis function is most suitable for a specific type of seismic motion data, thereby improving the accuracy and stability of velocity pulse identification. Therefore, when using wavelet analysis for velocity pulse identification, first calculating the correlation between different wavelet basis functions to select the optimal wavelet basis function is crucial for improving the effective identification of seismic velocity pulses.

[0157] In summary, the beneficial effects of the embodiments of the present invention are:

[0158] 1. The method proposed in the embodiment of the present invention can demonstrate high efficiency in identifying seismic velocity pulses. Selecting the wavelet basis function using the method proposed in the embodiment of the present invention can reduce computational complexity and simplify the process. The embodiment of the present invention provides an efficient and reliable identification method for engineering practice.

[0159] 2. Improve the accuracy of velocity pulse identification: By optimizing the wavelet basis selection and utilizing wavelet basis functions that are highly correlated with the original velocity time history, the accuracy of velocity pulse identification is effectively improved, ensuring that the extracted pulse characteristics are more consistent with the physical characteristics of actual ground motion.

[0160] 3. Considering the directional effect of ground motion: In view of the directional dependence of near-fault ground motion, the embodiments of the present invention can perform a comprehensive analysis based on three-component acceleration records in different directions, making the identification results more comprehensive and accurate, and helping to reveal the influence of the ground motion propagation direction on the pulse characteristics.

[0161] 4. By accurately identifying the velocity pulse characteristics, it provides a scientific basis for the in-depth analysis of seismic motion characteristics, provides data support for structural seismic design and earthquake disaster assessment, and helps to improve the seismic performance of engineering structures.

[0162] The present invention also provides a wavelet basis function selection device for seismic velocity pulse identification, as described in the following embodiments. Because the principles underlying the device are similar to those of the wavelet basis function selection method for seismic velocity pulse identification, the implementation of the device can be referenced to the implementation of the wavelet basis function selection method for seismic velocity pulse identification, and any repetitive details will not be repeated.

[0163] Figure 7 FIG. 1 is a schematic diagram of a wavelet basis function selection device for ground velocity pulse identification according to an embodiment of the present invention. Figure 7 As shown, the apparatus 700 includes:

[0164] The earthquake record data acquisition module 701 is used to acquire earthquake record data;

[0165] The original velocity time history processing module 702 is used to perform integration calculation on the earthquake record data and calculate the energy concentration interval of the original velocity time history;

[0166] The wavelet basis function processing module 703 is used to iteratively analyze the original velocity time history using different wavelet basis functions in sequence to obtain the time history distribution data and energy concentration interval corresponding to each wavelet basis function; the energy concentration interval represents a preset percentage of the total energy;

[0167] A downsampling processing module 704 is configured to downsample the energy concentration interval corresponding to each wavelet basis function to obtain a downsampled interval of each wavelet basis function; wherein the downsampled interval of each wavelet basis function has the same data density as the energy concentration interval of the original velocity time history;

[0168] The correlation calculation module 705 is used to calculate the first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history; use each wavelet basis function in turn to identify the velocity pulse time history of the original velocity time history, and calculate the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history; and select the wavelet basis function by sorting the first correlation coefficient and the second correlation coefficient.

[0169] In one embodiment, the original speed time history processing module 702 is specifically configured to:

[0170] Performing baseline correction on the seismic motion record data and removing low-frequency trend data in the seismic motion record data by filtering to obtain processed seismic motion record data; wherein the frequency of the low-frequency trend data is lower than a preset value;

[0171] The processed earthquake motion record data is quadratically integrated to obtain the original velocity time history;

[0172] The energy accumulation calculation of the original velocity time history is performed using the sliding spectrum window method to obtain the energy concentration area of ​​the original velocity time history.

[0173] In one embodiment, the original speed time history processing module 702 is specifically configured to:

[0174] The least square method is used to adjust the baseline. Based on the adjusted baseline, the low-frequency trend data is removed to obtain the processed seismic motion record data.

[0175] In one embodiment, the original speed time history processing module 702 is specifically configured to:

[0176] The original velocity time history is slid with a preset sliding window width and a preset sliding step length, and the signal energy is calculated segment by segment to generate a cumulative energy curve;

[0177] According to the cumulative energy curve, the time region where the total energy reaches a preset range is extracted as the energy concentration region of the original speed time course.

[0178] In one embodiment, the wavelet basis function processing module 703 is specifically used to:

[0179] For each wavelet basis function, the following processing is performed:

[0180] Set the wave basis function processing parameters, which include the number of iterations and scale parameters;

[0181] Performing continuous wavelet transform on the original velocity time history signal and decomposing the signal to obtain the time history distribution data corresponding to each wavelet basis function; wherein the wavelet coefficients are generated when the signal is decomposed;

[0182] The cumulative energy curve is calculated for the decomposed wavelet coefficients using the sliding spectrum window method;

[0183] According to the cumulative energy curve, the energy concentration interval corresponding to each wavelet basis function is determined.

[0184] In one embodiment, the downsampling processing module 704 is specifically configured to:

[0185] According to the ratio of the time resolution of the wavelet basis function transformation result to the sampling frequency of the earthquake motion record, the ratio of the wavelet basis function transformation result to be downsampled is determined;

[0186] According to the time range difference and density difference between the energy concentration interval of the original velocity time history and the energy concentration interval corresponding to the wavelet basis function, the number of nodes that need to be retained in the energy concentration interval corresponding to the wavelet basis function is determined;

[0187] According to the ratio of downsampling required for the wavelet basis function transformation result, the nodes on the energy concentration interval corresponding to the wavelet basis function are uniformly downsampled, and the number of nodes required to be retained is retained at equal intervals.

[0188] In one embodiment, the correlation calculation module 705 is specifically configured to:

[0189] The first correlation coefficient between the downsampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history is calculated according to the following formula:

[0190]

[0191] Where R is the first correlation coefficient; x i is the original speed time signal value after the energy concentration interval of the original speed time is evenly divided; y i is the velocity time history signal value after wavelet basis transformation on the down-sampling interval of the wavelet basis function; and x i 、y i The average value of ; n is the number of corresponding nodes.

[0192] In one embodiment, the correlation calculation module 705 is specifically configured to:

[0193] The second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history is calculated according to the following formula:

[0194]

[0195] Among them, r is the second correlation coefficient, v original (t) is the original velocity time history, v wavelet (t) is the velocity pulse time course identified by the wavelet basis function, v original The mean of (t), v wavelet The mean of (t).

[0196] An embodiment of the present invention further provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the wavelet basis function selection method for ground motion velocity pulse identification is implemented.

[0197] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for selecting wavelet basis functions for ground motion velocity pulse identification is implemented.

[0198] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method for selecting wavelet basis functions for ground motion velocity pulse identification is implemented.

[0199] In the embodiment of the present invention, a first correlation coefficient is calculated between the downsampled energy concentration interval corresponding to each wavelet basis function and the energy concentration interval of the original velocity time history. A second correlation coefficient is calculated between the velocity pulse time history identified by each wavelet basis function and the original velocity time history. The correlation between different wavelet basis identification results and the original velocity time history is utilized to optimize the selection of wavelet basis functions, thereby assisting in selecting appropriate wavelet basis functions in seismic velocity pulse identification. This effectively improves the accuracy of velocity pulse identification, ensures that the extracted pulse features are more consistent with the physical characteristics of actual seismic motion, and enhances the fidelity of velocity pulse identification.

[0200] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0201] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0202] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0203] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0204] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A wavelet basis function selection method for ground motion velocity pulse identification, characterized in that: include: Obtaining earthquake motion record data; Integrate the earthquake motion data and calculate the energy concentration interval of the original velocity history; Different wavelet basis functions are used in turn to iteratively analyze the original velocity time history, and the time history distribution data and energy concentration interval corresponding to each wavelet basis function are obtained; The energy concentration interval represents a preset percentage of the total energy; Downsampling the energy concentration interval corresponding to each wavelet basis function to obtain the downsampled interval of each wavelet basis function; wherein the data density of the downsampled interval of each wavelet basis function is the same as the energy concentration interval of the original velocity time history; Calculate the first correlation coefficient between the downsampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history; Using each wavelet basis function in turn to identify the velocity pulse time history of the original velocity time history, and calculating the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history; The wavelet basis function is selected by sorting the first correlation coefficient and the second correlation coefficient.

2. The method according to claim 1, wherein By integrating the earthquake motion data, the energy concentration interval of the original velocity history is obtained, including: Performing baseline correction on the seismic motion record data and removing low-frequency trend data in the seismic motion record data by filtering to obtain processed seismic motion record data; wherein the frequency of the low-frequency trend data is lower than a preset value; The processed earthquake motion record data is quadratically integrated to obtain the original velocity time history; The energy accumulation calculation of the original velocity time history is performed using the sliding spectrum window method to obtain the energy concentration area of ​​the original velocity time history.

3. The method according to claim 2, wherein Baseline correction is performed on the seismic motion record data. Low-frequency trend data in the seismic motion record data is removed by filtering to obtain processed seismic motion record data, including: The least square method is used to adjust the baseline. Based on the adjusted baseline, the low-frequency trend data is removed to obtain the processed seismic motion record data.

4. The method according to claim 2, wherein Using the sliding spectrum window method, the energy accumulation calculation of the original velocity time history is performed to obtain the energy concentration area of ​​the original velocity time history, including: The original velocity time history is slid with a preset sliding window width and a preset sliding step length, and the signal energy is calculated segment by segment to generate a cumulative energy curve; According to the cumulative energy curve, the time region where the total energy reaches a preset range is extracted as the energy concentration region of the original speed time course.

5. The method according to claim 1, wherein Different wavelet basis functions are used in turn to iteratively analyze the original velocity time history, and the time history distribution data and energy concentration interval corresponding to each wavelet basis function are obtained, including: For each wavelet basis function, the following processing is performed: Set the wave basis function processing parameters, which include the number of iterations and scale parameters; Performing continuous wavelet transform on the original velocity time history signal and decomposing the signal to obtain the time history distribution data corresponding to each wavelet basis function; wherein the wavelet coefficients are generated when the signal is decomposed; The cumulative energy curve is calculated for the decomposed wavelet coefficients using the sliding spectrum window method; According to the cumulative energy curve, the energy concentration interval corresponding to each wavelet basis function is determined.

6. The method according to claim 1, wherein The energy concentration interval corresponding to each wavelet basis function is downsampled to obtain the downsampled interval of each wavelet basis function, including: According to the ratio of the time resolution of the wavelet basis function transformation result to the sampling frequency of the earthquake motion record, the ratio of the wavelet basis function transformation result to be downsampled is determined; According to the time range difference and density difference between the energy concentration interval of the original velocity time history and the energy concentration interval corresponding to the wavelet basis function, the number of nodes that need to be retained in the energy concentration interval corresponding to the wavelet basis function is determined; According to the ratio of downsampling required for the wavelet basis function transformation result, the nodes on the energy concentration interval corresponding to the wavelet basis function are uniformly downsampled, and the number of nodes required to be retained is retained at equal intervals.

7. The method according to claim 1, wherein Calculate the first correlation coefficient between the downsampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history, including: The first correlation coefficient between the downsampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history is calculated according to the following formula: Where R is the first correlation coefficient; x i is the original speed time signal value after the energy concentration interval of the original speed time is evenly divided; y i is the velocity time history signal value after wavelet basis transformation on the down-sampling interval of the wavelet basis function; and x i 、y i The average value of ; n is the number of corresponding nodes.

8. The method according to claim 1, wherein Calculating the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history, including: The second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history is calculated according to the following formula: Among them, r is the second correlation coefficient, v original (t) is the original velocity time history, v wavelet (t) is the velocity pulse time course identified by the wavelet basis function, v original The mean of (t), v wavelet The mean of (t).

9. A wavelet basis function selection device for seismic velocity pulse identification, characterized in that: include: A seismic record data acquisition module, used to acquire seismic record data; The original velocity time history processing module is used to perform integration calculation on the earthquake motion record data and calculate the energy concentration interval of the original velocity time history; The wavelet basis function processing module is used to iteratively analyze the original velocity time history using different wavelet basis functions in sequence to obtain the time history distribution data and energy concentration interval corresponding to each wavelet basis function; The energy concentration interval represents a preset percentage of the total energy; A downsampling processing module is used to downsample the energy concentration interval corresponding to each wavelet basis function to obtain a downsampled interval of each wavelet basis function; wherein the data density of the downsampled interval of each wavelet basis function is the same as the energy concentration interval of the original velocity time history; A correlation calculation module is used to calculate the first correlation coefficient between the down-sampled interval of each wavelet basis function and the energy concentration interval of the original velocity time history; Each wavelet basis function is used in turn to identify the velocity pulse time history of the original velocity time history, and the second correlation coefficient between the velocity pulse time history identified by each wavelet basis function and the original velocity time history is calculated; the wavelet basis function is selected by sorting the first correlation coefficient and the second correlation coefficient.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

12. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images