A low-frequency petrophysical measurement method based on continuous frequency-sweep vibration

Abstract

The application discloses a low-frequency rock physical measurement method based on continuous frequency-sweep vibration and relates to the technical field of low-frequency rock physical measurement. Under the conditions of a given temperature, confining pressure and axial pressure, a reference frequency-sweep signal is designed, continuous frequency-sweep vibration is repeatedly applied through an exciter, and actual input force and rock sample deformation response are synchronously collected; the signals are aligned and superposed according to a common trigger time, a sliding short-time window Fourier transform is performed on the superposed signals, effective signals are identified according to a reference frequency-sweep track and noise is suppressed, a time domain signal is reconstructed through inverse transform, the amplitude envelope and instantaneous phase are extracted through Hilbert transform, the amplitude ratio and phase difference of force and deformation changing with frequency-sweep time are calculated, and rock sample stiffness and phase response are obtained through instrument correction. The application has the beneficial effects that the frequency sampling density and weak signal signal-to-noise ratio can be improved, the wide-frequency measurement time can be shortened, the amplitude and phase extraction fluctuation can be reduced, and a basis is provided for rock frequency-variable mechanical parameter measurement.

Description

Technical Field

[0001] This invention relates to the field of low-frequency rock physics measurement technology in seismic exploration, and specifically to a low-frequency rock physics measurement method based on continuous frequency sweep vibration. Background Technology

[0002] The elastic parameters and energy dissipation characteristics of rocks are related to frequency, temperature, pressure, and pore fluid state. Low-frequency stress-strain vibration experiments, by applying small dynamic loads to rock samples and measuring the amplitude and phase relationship between load and deformation, can be used to study elastic dispersion and attenuation within the seismic frequency band, providing experimental basis for seismic data interpretation and reservoir rock physics modeling.

[0003] Existing low-frequency measurements typically employ discrete fixed-frequency excitation, which involves sequentially applying sinusoidal vibrations at several preset frequency points, recording multiple cycles after the response at each frequency point stabilizes, and then calculating the mechanical parameters at that frequency. This method requires repeated frequency switching and waiting for the system to stabilize, resulting in a long broadband measurement time. When a limited number of frequency points are selected, it is difficult to fully describe the continuously changing parameters within the frequency band, and local dispersion characteristics may also be missed.

[0004] Continuous frequency sweep excitation can cover a frequency band in a single recording, but its frequency varies with time, and force and deformation signals are non-stationary. Directly using the amplitude statistics of the entire recording segment cannot obtain the amplitude ratio and phase difference that change with the sweep time; performing a Fourier transform on only the entire signal segment is also not conducive to distinguishing the effective signal distributed along the sweep trajectory from noise such as fixed-frequency interference and impulse interference. Especially when the rock sample deformation is weak, noise can cause significant fluctuations in the amplitude envelope and instantaneous phase.

[0005] Furthermore, the exciter driving voltage is not equal to the actual dynamic force acting on the rock sample; triggering deviations in repeated frequency sweep recordings may cause misalignment superposition; sensor amplitude-frequency response, channel fixation delay, and loading component flexibility also affect the measured amplitude ratio and phase difference. When the analysis frequency band is close to the resonant frequency of the measurement system after loading, inertia and resonance effects may also cause the measurement results to deviate from the quasi-static dynamic stiffness of the rock sample itself. Therefore, a complete processing flow needs to be established, from frequency sweep design, synchronous measurement, recording alignment, time-frequency domain noise reduction to amplitude and phase extraction and instrument calibration. Summary of the Invention

[0006] The purpose of this invention is to provide a low-frequency rock physics measurement method based on continuous frequency sweep vibration. Under set temperature, confining pressure, and axial pressure conditions, the method utilizes continuous frequency sweep excitation, synchronous acquisition of actual input force and rock sample deformation, repeated recording, alignment and superposition, and time-frequency domain noise reduction to determine the rock sample amplitude ratio, phase difference, and related mechanical parameters. By synchronously acquiring, aligning and superimposing, performing sliding short-time window Fourier transform, time-frequency domain noise suppression, and inverse transform on the actual input force and rock sample deformation, the denoised time-domain signal is obtained, and the amplitude envelope ratio and instantaneous phase difference characterize the mechanical response of the rock sample during the frequency sweep process.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A low-frequency rock physics measurement method based on continuous frequency sweep vibration includes the following steps:

[0009] S1, under given temperature, confining pressure and axial pressure conditions, design and determine the reference sweep frequency signal;

[0010] S2, continuous frequency sweep vibration is applied to the rock sample under test through the exciter, and the actual input force value acting on the rock sample is collected by the dynamic force sensor. The axial deformation response of the rock sample under test was simultaneously acquired using a deformation sensor. ;

[0011] S3, repeat step S2 at least N times;

[0012] S4, The force signal and deformation signal collected in step S3 are aligned according to the common trigger time and then superimposed;

[0013] S5. Perform a sliding short-window Fourier transform on the superimposed signal obtained in step S4 to obtain the time spectrum.

[0014] S6 identifies noise and valid signals within the time spectrum and suppresses the noise;

[0015] S7. Perform an inverse Fourier transform on the signal obtained in step S6 to obtain the time-domain signal.

[0016] S8, based on the processed force signal and deformation signal obtained in S7, calculate the amplitude ratio and phase difference of the two as a function of the sweep frequency time. Further, in step S1, when the reference sweep frequency signal uses a linear sweep frequency, the reference sweep frequency signal and its instantaneous frequency are expressed as follows:

[0017] ;

[0018] Where A is the reference amplitude. A window function is used to ensure a smooth transition during the start and end phases of the signal. and These are the start frequency and the end frequency, respectively. Let T be the time variable within the duration of the frequency sweep signal, and T be the frequency sweep duration. This is the initial phase.

[0019] The reference signal is used to define the relationship between the sweep time and the frequency, and the input quantity used for the mechanical parameter calculation is the force value measured by the dynamic force sensor.

[0020] Further, in step S3, the number of repetitions N is an integer greater than 4; repeated acquisitions are performed under the same sweep parameters and experimental conditions, and the records from the force and deformation channels are checked as paired data. The same time correction is applied to each record, and then coherent superposition is performed separately. When using normalized average superposition, the two channels use the same set of records and the same weights to maintain their original amplitude ratio and phase relationship.

[0021] Furthermore, in step S4, the alignment and superposition requires comparison and correction at the start time of each repeated frequency sweep, and then the input force values ​​are respectively... and axial deformation response Superposition is performed to avoid signal distortion caused by misalignment; after superposition, the average is calculated to obtain the average input force value. and mean axial deformation response As shown below;

[0022] ;

[0023] ;

[0024] Where j is the record number and N is the maximum number of repetitions. This represents the actual input force value for the j-th repeated test. The axial deformation response of the rock sample collected during the j-th repeated test is given.

[0025] Furthermore, in step S5, the sliding short-window Fourier transform refers to using a time window function to divide the force signal and deformation signal obtained in S4 into multiple continuous segments, and then performing a Fourier transform on each segment, thereby transforming the one-dimensional time domain signal into a two-dimensional time-frequency domain signal.

[0026] Furthermore, in step S6, the identification of noise and valid signals first involves analyzing the time-frequency domain characteristics of the reference swept frequency signal to confirm the valid signal characteristics, and then applying the average input force value obtained in step S4. and mean axial displacement deformation response Time-frequency domain filtering is performed to suppress noise, achieve time-frequency separation of the effective signal, and obtain the effective average input force value after noise reduction. and effective average axial displacement deformation response .

[0027] Furthermore, in step S7, the inverse Fourier transform refers to transforming the effective signal in the time-frequency domain to the time domain through the inverse Fourier transform, to obtain the force value and deformation signal after time-domain denoising.

[0028] Furthermore, in step S8, the amplitude ratio is obtained by first calculating the force signal amplitude envelope and the deformation amplitude envelope, then calculating their ratio, and finally obtaining the sample stiffness after instrument calibration.

[0029] Further, in step S8, the phase difference is obtained by performing Hilbert transforms on the force signal and the deformation signal respectively, and then obtaining their instantaneous phase difference; the phase difference obtained in step S6... and Perform Hilbert transforms on each, and obtain the corresponding results. and Then, calculate its amplitude envelope ratio. and instantaneous phase difference :

[0030] ;

[0031] ;

[0032] Where unwrap is the phase unwinding calculation, and atan2 is the arctangent function for calculating the phase difference.

[0033] The sliding short-window Fourier transform (SFT) performs Fourier transforms on adjacent and overlapping signal segments by moving the window function along the time axis, obtaining complex time-frequency spectra that simultaneously contain amplitude and phase information. The effective signal region is determined based on the time-frequency trajectory of the reference swept signal and the time-frequency distribution of the measured input force, suppressing out-of-region noise while preserving the complex information of the effective signal. Subsequently, an inverse transform is performed on the complex spectrum within each window, and continuous-time records are recovered using an overlap-addition and normalization method matched to the analysis window and inter-window step size.

[0034] Hilbert analytical signals were constructed from the recovered force and deformation signals, respectively. The magnitude was taken as the amplitude envelope, and the argument as the instantaneous phase. The amplitude ratio was obtained by dividing the force amplitude envelope by the deformation amplitude envelope at the same moment. The phase difference was obtained by expanding and subtracting the instantaneous phases of the two channels. After corrections for sensor sensitivity, channel amplitude and phase response, and necessary loading system flexibility, the stiffness and phase response of the rock sample were obtained.

[0035] For frequency sweep records with monotonically changing instantaneous frequencies, within the effective time period that satisfies the conditions of local narrow band and slowly varying response, the time-domain amplitude ratio and phase difference are mapped to the corresponding frequencies according to the known frequency sweep law. When the linear viscoelastic interpretation is applicable and the instrument phase has been corrected, the complex dynamic stiffness and stiffness loss factor can be further obtained; when the deformation gauge length and rock sample geometry are clear, the corresponding complex elastic modulus can be calculated.

[0036] The present invention has the following beneficial effects:

[0037] (1) Continuous frequency sweeping covers the preset frequency band in a single excitation, reducing discrete frequency point switching and repeated waiting processes, and obtaining response information continuously distributed along the frequency sweeping trajectory. Under the condition of achieving the same effective frequency band and data quality requirements, it is beneficial to shorten the broadband test time and increase the sampling density of the parameter curve.

[0038] (2) By simultaneously measuring the actual input force and rock sample deformation, the influence of the exciter output changing with frequency can be incorporated into the calculation of the ratio of force to deformation; combined with sensor amplitude and phase correction, the system error caused by replacing the input force with the drive command and the channel delay can be reduced.

[0039] Co-triggered and paired time corrections ensure the coherence of the repetitive sweep signals. For random noise with approximately independent and zero mean between repetitions, averaging and superposition can reduce the noise amplitude and improve the identifiability of weak deformation responses; time-frequency domain processing further utilizes the characteristics of the effective signal distribution along the sweep trajectory to suppress residual out-of-trajectory noise after superposition.

[0040] (3) First, the complex time spectrum is denoised and the time-domain signal is reconstructed. Then, the envelope and instantaneous phase are extracted, which can reduce the disturbance of noise on the amplitude ratio and phase difference, and form a parameter curve corresponding to the sweep time and frequency. Preserving complete force and deformation records can also support repeatability checks, reselection of processing parameters, and related pulse compression analysis. The degree of improvement mentioned above is determined by noise characteristics, sweep rate, instrument calibration, and experimental stability. Attached Figure Description

[0041] Figure 1 This is a flowchart of the present invention;

[0042] Figure 2 Examples of input force and axial deformation displacement response waveforms for traditional discrete harmonic testing at different frequencies are shown. The left figure is the 1Hz test, and the right figure is the 80Hz test.

[0043] Figure 3 The curves for a) stiffness as a function of frequency and b) phase difference as a function of frequency are obtained from traditional discrete harmonic testing.

[0044] Figure 4The original record of continuous frequency sweep and its time spectrum are shown, where a) and b) are the time domain curve and time frequency domain curve of the input force, respectively, and c) and d) are the time domain curve and time frequency domain curve of the axial deformation displacement response, respectively.

[0045] Figure 5 The waveform and time spectrum after continuous frequency sweep recording, alignment and superposition, and time-frequency denoising are shown. Among them, a) and b) are the time domain curve and time-frequency curve of the input force after processing, respectively, and c) and d) are the time domain curve and time-frequency curve of the axial deformation displacement response after processing, respectively.

[0046] Figure 6 The results of a) stiffness as a function of frequency and b) phase difference extraction are compared before and after denoising using the continuous frequency sweep method. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] This invention discloses a low-frequency rock physics measurement method based on continuous frequency sweep vibration. An embodiment of this invention is described below with reference to the accompanying drawings. Figure 1 A complete workflow from reference frequency sweep design to amplitude ratio and phase difference calculation is presented. Steps S1 to S8 constitute the main measurement and processing procedures, with correlation pulse compression serving as an extended processing method to preserve the waveform. The listed sweep frequency band, duration, and sampling rate are used to illustrate the implementation process and should be selected based on the effective frequency band of the system after loading and the response characteristics of the rock sample to be tested. The specific process is as follows:

[0049] S1, Test condition setting and reference sweep signal design.

[0050] The rock sample to be tested is installed in the loading system, and the temperature, confining pressure, and axial pressure are set. In this embodiment, the axial pressure refers to the prestress corresponding to the axial static load, and the continuously swept frequency dynamic load is superimposed on this static load. After the temperature, pressure, and static deformation reach the stability requirements within the preset observation time, the frequency sweep measurement begins. If it is necessary to study the influence of pore fluid, the pore pressure and fluid state can be recorded and controlled simultaneously.

[0051] The reference sweep signal is generated by the following formula:

[0052] ;

[0053] Where A is the reference amplitude. A window function is used to ensure a smooth transition during the start and end phases of the signal. and These are the start frequency and the end frequency, respectively. Let T be the time variable within the duration of the frequency sweep signal, and T be the frequency sweep duration. This is the initial phase.

[0054] The reference signal is used to define the relationship between the sweep time and the frequency, and the input quantity used for the mechanical parameter calculation is the force value measured by the dynamic force sensor.

[0055] This example uses a linear frequency sweep from 1 Hz to 100 Hz for a duration of T = 12 s, with a sweep rate of (100−1) / 12 = 8.25 Hz / s. Window function Cosine gradients are used at both ends of the signal to reduce transients caused by sudden start and stop of excitation; the gradient segment and the adjacent interval affected by it are not used for the final amplitude and phase parameter interpretation.

[0056] For rock samples with lower frequencies or significant long-term memory effects, it is necessary to extend the sweep time or reduce the sweep rate of the corresponding frequency band to ensure that the local signal contains sufficient vibration periods and that the rock sample response changes smoothly with the excitation frequency. The sampling interval of the entire Fourier spectrum recorded by the 12-second sweep is 1 / T, approximately 0.083 Hz.

[0057] S2, continuous frequency sweep excitation and synchronous acquisition of force and deformation.

[0058] A reference signal is output from a frequency sweep signal generator, which drives a vibrator via a power driver to apply continuous frequency sweep vibrations to the rock sample. A dynamic force sensor records the actual input force. The axial displacement of the rock sample was recorded synchronously using a deformation sensor. , where j is the record number. The two channels share a sampling clock and are started by the same trigger signal; at the same time, the reference signal, trigger position, sampling rate, temperature, pressure and instrument calibration information are saved.

[0059] In this example, the sampling rate is 1 kHz. The force and deformation channels are configured with analog anti-aliasing low-pass filters before analog-to-digital conversion. The passband should cover the effective analysis frequency band, and the stopband should meet the suppression requirements of the sampling system. Recording should cover the entire frequency sweep process, and background noise and attenuation tails should be retained before and after the sweep to estimate noise, check response stability, and handle boundary effects. Digital denoising after analog-to-digital conversion is used to process the acquired signal but cannot recover information lost due to sampling aliasing.

[0060] Figure 4 a) and c) in the figure show the original force and displacement waveforms, respectively, and the vibration waveforms that change gradually at both ends and with the frequency sweep process can be observed. Figure 4The bright diagonal bands in b) and d) correspond to the trajectory of the sweep energy moving towards higher frequencies over time; the horizontal axis of each subplot is time (s), the vertical axes of a) and c) are force (kN) and displacement (mm) respectively, and the vertical axes of b) and d) are frequency (Hz). Figure 4 In b), the horizontal stripe around 50 Hz reflects the difference in time-frequency distribution between fixed-frequency interference and the effective swept-frequency signal, which can be used as a basis for subsequent noise identification.

[0061] S3, Quality check for repeated data collection and recording.

[0062] Keeping the experimental conditions, reference frequency sweep pattern, driving amplitude, triggering method, and sampling parameters consistent, repeat step S2 to obtain no less than N sets of valid force-deformation pair records, where N is an integer greater than 4. In this example, N = 5, using 5 sets of valid records to illustrate the subsequent superposition processing. Sufficient intervals should be reserved between each frequency sweep to allow for transient decay and device stabilization.

[0063] The range, waveform integrity, reference sweep trajectory, and test conditions of the force and deformation channels are checked sequentially. Records with exciter overload, sensor saturation, missed triggers, or significant changes in test conditions are removed in pairs and retested until the valid repeatable records meet the set requirements. The allowable range of peak value, root mean square value, and repeatable waveform differences are determined by instrument calibration and pre-experimentation; obviously abnormal records are not directly mixed into the superposition.

[0064] S4, jointly triggers alignment and coherent overlay.

[0065] First, coarsely align the records according to their common trigger times. Then, estimate the residual time shift using methods such as the cross-correlation peak position between the actual input force waveform and the selected reference force waveform, and superimpose them to reduce the influence of random noise.

[0066] After superposition, the average value is calculated to obtain the average input force value. and mean axial deformation response As shown below;

[0067] ;

[0068] ;

[0069] Where j is the record number and N is the maximum number of repetitions. This represents the actual input force value for the j-th repeated test. The axial deformation response of the rock sample collected during the j-th repeated test is given.

[0070] S5, Sliding Short-Window Fourier Transform.

[0071] right and Remove the mean or slow drift in the same way, and set the analysis window g and the window sliding step size H on the same time grid. Perform a Fourier transform on the discrete signal within each window, with the window center corresponding to the time coordinate.

[0072] The window length should take into account the low-frequency vibration period, time positioning, and the change in sweep frequency within the window. A window that is too short will make it difficult to distinguish low-frequency components, while a window that is too long will broaden the sweep energy along the frequency axis. The window length, overlap rate, and effective frequency band are determined by combining the sweep rate and the measured trajectory width, and the applicability of the parameters is checked through reconstruction consistency and local steady-state comparison. The force and deformation channels use the same analysis window, step size, and time reference.

[0073] S6, effective signal recognition and time-frequency domain noise suppression.

[0074] Perform the same transformation as in step S5 on the reference sweep signal to obtain the expected sweep trajectory; combine this with the measured force-time spectrum to determine the effective signal region. The region width covers the main frequency band broadening caused by the analysis window, sweep rate, and allowable response delay, and a smooth transition boundary is used to reduce ringing introduced by hard truncation. Fixed-frequency interference, asynchronous pulses, or random background noise outside the region are suppressed according to their degree of separation from the trajectory.

[0075] Figure 5 Figures b) and d) show that the processed effective energy is mainly distributed along the frequency sweep trajectory. Compare this with... Figure 4 By comparing b) and d) in the data, we can check whether the noise outside the trajectory has been reduced, whether the main trajectory is continuous, and whether the effective signal has been mistakenly deleted in the filtering area.

[0076] S7, Inverse Fourier Transform and Time-Domain Signal Reconstruction

[0077] Perform an inverse Fourier transform on the complex spectrum of each time window in step S6 to obtain the denoised local time segments; place each segment back to its corresponding time position according to the original step size, multiply by the synthesis window, and sum them up; then normalize by the sum of the overlaps of the analysis window and the synthesis window products to obtain the continuous force signal. and displacement signal The synthesis window and step size should ensure that the normalized denominator within the retention interval is non-zero. Figure 5 a) and c) in the figure are the reconstructed force and displacement waveforms, which can be used to check the noise reduction and waveform continuity.

[0078] S8, Envelope amplitude ratio, instantaneous phase difference and instrument calibration

[0079] Taking displacement response as an example, for the denoised... and Perform Hilbert transforms on each side to obtain and Then, calculate its amplitude envelope ratio. and instantaneous phase difference :

[0080] ;

[0081] ;

[0082] Where unwrap is the phase unwinding calculation, and atan2 is the arctangent function for calculating the phase difference.

[0083] During periods when the displacement envelope is close to the noise level, the ends gradually change, there is strong interference crossing, or the phase estimation is unstable, no division ratio calculation or loss interpretation is performed. Figure 6 Figures a) and b) show the apparent dynamic stiffness and instantaneous phase difference obtained from the envelope ratio, respectively. The blue line represents the result before denoising, and the red line represents the result after denoising. Compared with the blue curve before denoising, the rapid fluctuations of the red curve after denoising are significantly reduced, reflecting the improvement in the stability of amplitude and phase extraction. The horizontal axis in the figure represents the sweep time, and the lower axis represents the corresponding frequency. The smoothness of the curve is used to evaluate the processing effect; the final value of material loss is still based on the effective frequency band result after channel and system calibration.

[0084] Discrete Harmonic Comparison and Result Check

[0085] When using the discrete harmonic method for comparison, the force and displacement waveforms after stabilization are recorded one by one at the selected frequency. Figure 2 The left and right graphs show the records at approximately 1 Hz and approximately 80 Hz, respectively. The blue line represents the actual input force, and the red line represents the displacement response. Both graphs have time (s) on the horizontal axis, force (kN) on the left vertical axis, and displacement (mm) on the right vertical axis. Force values ​​are read from the left axis and displacement values ​​from the right axis. The amplitude ratio and relative time position of the two waveforms are used to calculate the apparent stiffness and phase difference at that frequency. Figure 3 Figures a) and b) summarize the stiffness and phase results at discrete frequency points, respectively. The horizontal axis of both figures is frequency (Hz), and the vertical axis is apparent stiffness (kN / mm) and phase difference (°), respectively.

[0086] The control test should use the same rock sample and maintain consistent temperature, confining pressure, axial static load, pore fluid state, deformation measurement gauge length, and sensor calibration; the dynamic amplitude should be the same or verified to be within the same linear response range. Figure 6 The frequency sweep results are in Figure 3 Compare values ​​taken near the corresponding frequency points, and simultaneously check the degree of dispersion of repeated measurements and the effect of channel correction. Figures 2 to 6 The study demonstrates two measurement methods and the characteristics before and after processing.

[0087] The above implementation method obtains the amplitude ratio and phase difference of rock sample force and deformation through reference frequency sweep design, paired synchronous acquisition, repeated recording alignment and superposition, complex time spectrum denoising, and time domain reconstruction. Under applicable conditions, it further obtains frequency-varying stiffness, modulus, and loss parameters. Different types of exciters and deformation sensors can be selected according to the test frequency band and environmental requirements. When using other monotonic frequency sweep laws, a time-frequency mapping is established based on its known instantaneous frequency function, and the processing parameters are adjusted accordingly.

[0088] This invention also discloses a measurement system for implementing the above-described method, comprising an environmental and static load control section, a frequency sweep excitation section, a synchronous measurement section, and a data processing section. The environmental and static load control section includes a pressure vessel, a confining pressure loading assembly, a temperature control assembly, and an axial static load mechanism; the frequency sweep excitation section includes a frequency sweep signal generator, a power driver, and a vibrator; the synchronous measurement section includes a dynamic force sensor, a deformation sensor, and a synchronous data acquisition unit; the data processing section includes a processor and a storage unit for storing reference signals, raw records, and processing results. The processor executes steps S4 to S8 and can coordinate the frequency sweep excitation and repeated acquisition. Each component uses existing devices capable of meeting the required frequency band, measurement range, and synchronization accuracy.

[0089] The above embodiments are used to illustrate the technical solution of the present invention. Those skilled in the art can make corresponding adjustments to the device selection, sweep frequency parameters, and processing parameters within the scope defined by the claims, and these adjustments can be implemented according to the above process.

Claims

1. A low-frequency rock physics measurement method based on continuous frequency sweep vibration, characterized in that, Includes the following steps: S1, under given temperature, confining pressure and axial pressure conditions, design and determine the reference sweep frequency signal; S2, continuous frequency sweep vibration is applied to the rock sample under test through the exciter, and the actual input force value acting on the rock sample is collected by the dynamic force sensor. The axial deformation response of the rock sample under test was simultaneously acquired using a deformation sensor. ; S3, repeat step S2 at least N times; S4, The force signal and deformation signal collected in step S3 are aligned according to the common trigger time and then superimposed; S5. Perform a sliding short-window Fourier transform on the superimposed signal obtained in step S4 to obtain the time spectrum. S6 identifies noise and valid signals within the time spectrum and suppresses the noise; S7. Perform an inverse Fourier transform on the signal obtained in step S6 to obtain the time-domain signal. S8, based on the processed force signal and deformation signal obtained in S7, calculate the amplitude ratio and phase difference of the two as a function of the sweep frequency time.

2. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 1, characterized in that, The reference frequency sweep signal in step S1, when using linear frequency sweep, is expressed as follows: ； Where A is the reference amplitude. A window function is used to ensure a smooth transition during the start and end phases of the signal. and These are the start frequency and the end frequency, respectively. Let T be the time variable within the duration of the frequency sweep signal, and T be the frequency sweep duration. This is the initial phase.

3. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 2, characterized in that, In step S3, the number of repetitions N is an integer greater than 4. The data is repeatedly collected under the same frequency sweep parameters and experimental conditions. The records of the force and deformation channels are checked as paired data. The same time correction is applied to each record, and then coherent superposition is performed separately. When normalized average superposition is used, the same set of records and the same weights are used for the two channels to improve the signal-to-noise ratio.

4. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 3, characterized in that, In step S4, the alignment and superposition requires comparison and correction at the start time of each repeated frequency sweep, and then the input force values ​​are respectively... and axial deformation response Superimpose the signals to avoid signal distortion caused by misaligned superposition. After superposition, the average value is calculated to obtain the average input force value. and mean axial deformation response As shown below; ； ； Where j is the record number and N is the maximum number of repetitions. This represents the actual input force value for the j-th repeated test. The axial deformation response of the rock sample collected during the j-th repeated test is given.

5. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 4, characterized in that, The sliding short-window Fourier transform in step S5 refers to using a time window function to divide the force and deformation signals obtained in S4 into multiple continuous segments, and then performing a Fourier transform on each segment to transform the one-dimensional time domain signal into a two-dimensional time-frequency domain signal.

6. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 5, characterized in that, In step S6, noise and valid signal identification first involves analyzing the time-frequency domain characteristics of the reference swept frequency signal to confirm valid signal characteristics, and then applying the average input force value obtained in step S4. and mean axial displacement deformation response Time-frequency domain filtering is performed to suppress noise, achieve time-frequency separation of the effective signal, and obtain the effective average input force value after noise reduction. and effective average axial displacement deformation response .

7. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 6, characterized in that, The inverse Fourier transform in step S7 refers to transforming the effective signal in the time-frequency domain to the time domain through the inverse Fourier transform, so as to obtain the force value and deformation signal after time-domain denoising.

8. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 7, characterized in that, In step S8, the amplitude ratio is obtained by first calculating the force signal amplitude envelope and the deformation amplitude envelope, then calculating their ratio, and finally obtaining the sample stiffness after instrument calibration.

9. The low-frequency rock physics measurement method based on continuous frequency sweep vibration as described in claim 8, characterized in that, In step S8, the phase difference is obtained by performing Hilbert transforms on the force signal and the deformation signal respectively, and then obtaining their instantaneous phase difference; the phase difference obtained in step S6... and Perform Hilbert transforms on each, and obtain the corresponding results. and Then, calculate its amplitude envelope ratio. and instantaneous phase difference : ； ； Where unwrap is the phase unwinding calculation, and atan2 is the arctangent function for calculating the phase difference.

Images