Pulsatile pressure frequency domain analysis method based on transient schlieren feature point time series data

Abstract

The application discloses a kind of pulsating pressure frequency domain analysis method based on transient schlieren feature point timing data, it is related to the cross technical field of intelligent fluid mechanics and aircraft aerodynamic test, comprising: S1, the transient schlieren image sequence related to isolation section is obtained by high-speed schlieren acquisition system, and the transient schlieren image sequence is preprocessed, to obtain corresponding flow field feature point timing data;S2, based on the acquisition timing of S1, the trajectory of the timing position of feature point in x direction is constructed;Based on timing position trajectory, the oscillation displacement sequence of each time is calculated, and the corresponding oscillation displacement difference sequence is obtained by differentiating processing to oscillation displacement sequence;S3, the time-domain signal in oscillation displacement difference sequence is converted into frequency-domain signal after fast fourier transform, and the corresponding power spectral density PSD curve is obtained.This application compared with traditional sensor measurement method, no flow field interference, can be adapted to complex configuration, and has greater application potential to high temperature scene.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of intelligent fluid dynamics and aircraft aerodynamic testing. More specifically, this invention relates to a method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data. Background Technology

[0002] During supersonic flight or complex aerodynamic conditions, unsteady oscillations of structures such as strong shock waves and shear layers in the flow field can induce pulsating pressure. The frequency domain characteristics (dominant oscillation frequency, frequency band distribution, peak power, etc.) directly determine the structural strength, vibration and noise levels, and combustion stability of the aircraft. Accurately obtaining the frequency domain characteristics of pulsating pressure is a key prerequisite for aircraft structural design verification and flight safety assurance, and has significant theoretical research value and engineering application significance.

[0003] Existing frequency domain analysis methods for pulsating pressure have the following insurmountable technical shortcomings, making them unable to meet engineering requirements under complex working conditions:

[0004] 1. Traditional pressure sensor measurement method: It is necessary to drill holes in the aircraft wall to install sensors, which not only interferes with the integrity of the flow field, but is also limited by complex configurations (such as narrow spaces in isolation sections), high-temperature conditions (sensors are prone to failure when the wall temperature is >300℃), and the number of installation points is limited, making it impossible to achieve wide-field monitoring of pulsating pressure distribution; at the same time, there are problems of difficult wiring and high data acquisition costs in the wind tunnel.

[0005] 2. CFD numerical simulation method: It relies on complex turbulence models and boundary condition settings, and its accuracy in reproducing the dynamic evolution characteristics of unsteady flow fields is insufficient. In particular, for the pulsating pressure induced by strong shock waves, the frequency domain prediction error usually exceeds 10%, making it difficult to provide reliable engineering reference data.

[0006] 3. Existing image-based analysis methods: Although non-contact detection has been achieved, a clear correlation mechanism between the time series data of flow field feature points and pulsating pressure has not been established, and a systematic technical process of "data preprocessing - feature extraction - frequency domain inversion" is lacking.

[0007] With the development of high-speed schlieren imaging technology, transient schlieren images can capture the dynamic evolution information of flow field feature points non-contactly, at high frame rates, and with a wide field of view, providing rich raw data for frequency domain analysis of pulsating pressure. However, how to accurately extract dynamic information directly related to pulsating pressure from the time-series data of feature points and achieve efficient frequency domain feature inversion through scientific signal processing has become a pressing technical challenge. Therefore, it is urgent to propose a systematic and high-precision frequency domain analysis method to overcome existing technical bottlenecks and adapt to the needs of aircraft pulsating pressure detection under complex operating conditions. Summary of the Invention

[0008] One object of the present invention is to solve at least the above-mentioned problems and / or defects, and to provide at least the advantages described below.

[0009] To achieve these objectives and other advantages of the present invention, a method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data is provided, comprising:

[0010] S1. Acquire transient schlieren image sequences related to the isolation section through a high-speed schlieren acquisition system, extract feature points from the transient schlieren image sequences, and preprocess the transient schlieren image sequences based on the feature points to obtain the corresponding flow field feature point time series data.

[0011] S2. Based on the acquisition time sequence of S1, construct the temporal position trajectory of the feature points in the x direction;

[0012] The oscillation displacement sequence at each moment is calculated based on the temporal position trajectory, and the corresponding oscillation displacement difference sequence is obtained by differential processing of the oscillation displacement sequence.

[0013] S3. The time-domain signal in the oscillating displacement difference sequence is converted into a frequency-domain signal by fast Fourier transform, and the corresponding power spectral density (PSD) curve is obtained by amplitude spectrum calculation and normalization. The frequency corresponding to the peak value in the PSD curve reflects the main oscillation frequency of the pulsating pressure.

[0014] Preferably, in S1, the high-speed schlieren acquisition system has a acquisition frame rate Fs of 9000 FPS;

[0015] The feature points are extracted using the following method:

[0016] Strong shock wave targets were detected within the isolation section using a target detection algorithm.

[0017] The corner points of the detection frame corresponding to the contact points between the strong shock wave target and the isolation section wall are used as feature points, and the chordal displacement of the aircraft at the feature points is directly related to the pulsating pressure.

[0018] Preferably, in S1, the preprocessing includes:

[0019] Based on the established abnormal frame judgment criteria, abnormal frames in the transient schlieren image sequence are removed, and the transient schlieren image sequence is completed using linear interpolation.

[0020] Gaussian filtering is used to smooth the temporal coordinate data of the feature points in the x-direction of the completed transient schlieren image sequence in order to obtain the corresponding temporal data of the flow field feature points.

[0021] The Gaussian kernel size of the Gaussian filter is 5×1, and the standard deviation σ is set to 0.03~0.05.

[0022] Preferably, in S2, the time-series position trajectory is formed by arranging the feature points in the time-series data of the flow field feature points according to their x-axis coordinates in a time sequence, and the time-series position trajectory X(t) is characterized by the following expression:

[0023] X(t)=[x(tᵢ)]

[0024] In the above formula, x(tᵢ) is the x-coordinate of the feature point in the i-th frame of the image at time t, and i=1,2,…n, where n is the total number of frames;

[0025] The oscillation displacement Δxᵢ of adjacent frames is calculated using the following formula to obtain the oscillation displacement sequence at time t:

[0026] Δxᵢ=x(tᵢ)-x(t i-1 )

[0027] By directly characterizing the dynamic oscillation amplitude of the feature point relative to the reference position using an oscillating displacement sequence, the conclusion is drawn that the feature point and the changing trend of pulsating pressure are linearly positively correlated.

[0028] Preferably, in S2, the differential processing refers to performing a first-order backward differential processing on the oscillating displacement sequence to obtain the oscillating displacement difference sequence Δx'(t) at time t, and Δx'(t) = [Δx'ᵢ]:

[0029] Δx'ᵢ=Δxᵢ -Δxᵢ₋1

[0030] In the above formula, Δx'ᵢ is the oscillation displacement difference between adjacent frames, and when i=1, Δx'1=Δx1.

[0031] Preferably, in S3, the frequency resolution Δf during the fast Fourier transform is obtained by the following formula:

[0032] Δf=Fs / M

[0033] In the above formula, Fs is the acquisition frame rate of the high-speed schlieren acquisition system, M is the signal length, and M=n-1, where n is the total number of frames.

[0034] The present invention has at least the following beneficial effects:

[0035] Firstly, this invention provides a systematic non-contact analysis process, namely, a complete technical process of "feature point time series data acquisition - abnormal frame removal - Gaussian filtering - trajectory construction - oscillation displacement analysis - differential processing - FFT transformation - PSD feature extraction", which realizes the inversion analysis from transient schlieren images to pulsating pressure frequency domain features, and solves the core problem of the lack of a correlation mechanism between flow structure and unsteady pulsating pressure in the existing technology.

[0036] Secondly, compared with traditional sensor measurement methods, the method of this invention is free from flow field interference and can be adapted to complex configurations; in addition, it has great application potential for high-temperature scenarios that cannot be measured by sensors and has strong engineering applicability.

[0037] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the overall technical process of the present invention;

[0039] Figure 2 This is a comparative diagram showing the effectiveness verification of the method between the present invention and the prior art.

[0040] Figure 3 for Figure 2 Enlarged view of the area within the dashed box. Detailed Implementation

[0041] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0042] This invention provides a frequency domain analysis method for pulsating pressure based on transient schlieren feature point time-series data. It is particularly suitable for key components such as the isolator and nozzle of air-breathing aircraft. Under conditions of high temperature, complex configuration, and limited sensor deployment, it achieves non-contact, high-precision inversion of pulsating pressure frequency domain features, providing core technical support for aircraft structural design optimization and flow control. The specific implementation path involves feature point time-series data preprocessing, oscillation displacement calculation, frequency domain transformation, and feature extraction to achieve high-precision inversion of pulsating pressure frequency domain features. Figure 1 As shown, the specific steps are as follows:

[0043] Step 1: Obtain time series data of flow field feature points

[0044] Data Source and Definition: Taking the isolator section of an air-breathing aircraft as the research object, and focusing on the scenario of strong shock wave-induced pulsating pressure under Mach 4.0 conditions, the input is a sequence of transient schlieren images of the isolator section (acquired by a high-speed schlieren acquisition system, with an acquisition frame rate of 9000 FPS and a total number of frames N=3000). High-precision detection of strong shock wave targets within the isolator section is achieved using a deep learning target detection algorithm. Then, the corner points of the detection frame corresponding to the contact points between the strong shock wave and the isolator section wall are extracted as core feature points. The chordal displacement of the aircraft at these feature points is directly related to the pulsating pressure.

[0045] Abnormal frame removal: Set the criteria for judging abnormal frames: If the feature point coordinates are missing in a certain frame or the difference between the x coordinate of the feature point in the frame and the average x coordinate of the adjacent 5 frames exceeds 5 pixels (considered as a coordinate mutation), it is marked as an abnormal frame. Linear interpolation method (based on the normal coordinate data of the 3 frames before and after the abnormal frame) is used to fill in the missing or mutation data to ensure the continuity of time series data.

[0046] Gaussian filtering: To eliminate the interference of image noise on the feature point coordinates, Gaussian filtering is used to smooth the temporal coordinate data of the feature points in the x-direction. The filtering parameters are set as follows: Gaussian kernel size 5×1 (adapted to one-dimensional temporal data), standard deviation σ=0.03~0.05. High-frequency noise is suppressed through convolution operation, while preserving the true dynamic evolution trend of the feature points.

[0047] Step 2: Constructing the temporal trajectory of feature points and calculating oscillation displacement

[0048] Temporal trajectory construction: based on the acquisition time sequence of transient schlieren images (t1~t... 3000 The preprocessed feature points are arranged in the x-direction coordinates according to the time series to construct the temporal position trajectory of the feature points in the x-direction: X(t)=[x(t1), x(t2), ..., x(t... n [], where n=3000 is the total number of frames, and x(tᵢ) is the x-coordinate of the feature point in the i-th frame. The x-coordinate positions of the core feature points in the corresponding temporal schlieren image are shown in Table 1:

[0049] Table 1: x-coordinates of core feature points in temporal schlieren images

[0050]

[0051] Oscillatory displacement analysis: Calculate the oscillatory displacement sequence at each time step: Δxᵢ=x(tᵢ)-x(t i-1 (i=1~n), this sequence directly reflects the dynamic oscillation amplitude of the feature point relative to the reference position, and is linearly positively correlated with the changing trend of pulsating pressure.

[0052] Differential processing: To highlight the rate of change characteristics of oscillating displacement (which directly corresponds to the time-domain variation characteristics of pulsating pressure), a first-order backward differential processing is performed on the oscillating displacement sequence, i.e., Δx'ᵢ=Δxᵢ -Δxᵢ₋1 (i=1~n), to obtain the oscillating displacement differential sequence Δx'(t), i.e., Δx'(t)=[Δx'ᵢ]; where, in the first frame (i.e. when i=1), Δx'1=Δx1, that is, the differential data of the first frame is completed according to the first value. The dynamic change information of the signal is enhanced by differential processing, providing high-quality input data for subsequent frequency domain transformation.

[0053] Step 3: Frequency Domain Feature Inversion of Pulsating Pressure

[0054] A Fast Fourier Transform (FFT) is performed on the differential oscillation displacement sequence Δx'(t) to convert the time-domain signal into a frequency-domain signal. The signal length is taken as the effective length of the differential sequence, M = N - 1 = 2999; the frequency resolution is Δf = Fs / M ≈ 3.00 Hz. The complex form of the frequency-domain signal Y(f) is calculated using the FFT algorithm, and its amplitude spectrum |Y(f)| is calculated (as shown in Table 2). The amplitude spectrum is then normalized (divided by the signal length M). The resulting power spectral density (PSD) curve can intuitively reflect the energy distribution at different frequencies. The frequency corresponding to the peak value in the PSD curve is the main oscillation frequency of the pulsating pressure.

[0055] Table 2 Power spectral density of this method

[0056]

[0057] Step 4: Verification of Method Validity

[0058] To verify the accuracy, reliability and engineering applicability of the method of the present invention, three sets of comparative experiments were set up. The experimental conditions were uniform: air-breathing aircraft isolation section, Mach number 4.0, strong shock wave induced pulsating pressure condition, test duration ≈ 0.33s (3000 frames of images).

[0059] Comparative experimental design:

[0060] Control group 1: Traditional pressure sensor measurement method, using a high-temperature resistant piezoelectric pressure sensor (measurement range 0~1MPa, frequency response 0~10kHz), three sensors are installed by drilling holes at the corresponding positions of the feature points on the wall of the isolation section, and the frequency domain features are obtained by collecting the time domain signal of the pulsating pressure through FFT.

[0061] Control group 2: SPOD (Spectral Proper Orthogonal Decomposition) modal analysis method, based on the flow field velocity field data of transient schlieren image sequence, extracts the main oscillation mode and corresponding frequency of the flow field through SPOD algorithm.

[0062] Experimental group: The frequency domain analysis method based on feature point time series data proposed in this invention.

[0063] Figure 2 The image shows a comparison of the calculation results from the three methods above. Figure 3 As can be seen from the enlarged schematic diagram, the main frequencies calculated by the traditional pressure sensor measurement method, the SPOD modal analysis method, and this method are 500Hz, 510Hz, and 510Hz, respectively.

[0064] The results show that the relative error between the dominant frequency obtained by this method and the measurement results of traditional pressure sensors is only 2.0%, and it is in high agreement with the SPOD modal analysis results, fully verifying the high accuracy and reliability of the method in frequency domain analysis. Compared with traditional pressure sensor measurement, this method does not require physical sensor deployment, fundamentally avoiding the intrusive interference of sensors on the supersonic flow field. At the same time, it solves the technical problem of sensor high-temperature failure under extreme conditions and can be adapted to harsh testing environments where sensors cannot work. Compared with the SPOD modal analysis method, this method achieves efficient and intelligent extraction of flow field frequency domain features, greatly simplifies the analysis process and improves computational efficiency. Moreover, relying on the non-invasive advantage of schlieren visual observation, it directly correlates the shock wave spatial structure and oscillation frequency characteristics, providing a new quantitative means with accuracy, engineering practicality and physical interpretability for the study of the shock wave-load coupling mechanism of supersonic flow field, highlighting the unique application value and innovative significance of this method in the field of supersonic flow testing.

[0065] The above solution is merely an illustration of a preferred example and is not limited thereto. When implementing this invention, appropriate substitutions and / or modifications can be made according to the user's needs.

[0066] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for the present invention. Other modifications can be readily made by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and examples shown and described herein.

Claims

1. A method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data, characterized in that, include: S1. Acquire transient schlieren image sequences related to the isolation section through a high-speed schlieren acquisition system, extract feature points from the transient schlieren image sequences, and preprocess the transient schlieren image sequences based on the feature points to obtain the corresponding flow field feature point time series data. S2. Based on the acquisition time sequence of S1, construct the temporal position trajectory of the feature points in the x direction; The oscillation displacement sequence at each moment is calculated based on the temporal position trajectory, and the corresponding oscillation displacement difference sequence is obtained by differential processing of the oscillation displacement sequence. S3. The time-domain signal in the oscillating displacement difference sequence is converted into a frequency-domain signal by fast Fourier transform, and the corresponding power spectral density (PSD) curve is obtained by amplitude spectrum calculation and normalization. The frequency corresponding to the peak value in the PSD curve reflects the main oscillation frequency of the pulsating pressure.

2. The method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data as described in claim 1, characterized in that, In S1, the high-speed schlieren acquisition system has a frame rate Fs of 9000 FPS; The feature points are extracted using the following method: Strong shock wave targets were detected within the isolation section using a target detection algorithm. The corner points of the detection frame corresponding to the contact points between the strong shock wave target and the isolation section wall are used as feature points, and the chordal displacement of the aircraft at the feature points is directly related to the pulsating pressure.

3. The method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data as described in claim 1, characterized in that, In S1, the preprocessing includes: Based on the established abnormal frame judgment criteria, abnormal frames in the transient schlieren image sequence are removed, and the transient schlieren image sequence is completed using linear interpolation. Gaussian filtering is used to smooth the temporal coordinate data of the feature points in the x-direction of the completed transient schlieren image sequence in order to obtain the corresponding temporal data of the flow field feature points. The Gaussian kernel size of the Gaussian filter is 5×1, and the standard deviation σ is set to 0.03~0.

05.

4. The method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data as described in claim 1, characterized in that, In S2, the temporal position trajectory is formed by arranging the feature points in the time series data of the flow field feature points according to their x-axis coordinates in a time sequence. The temporal position trajectory X(t) is characterized by the following expression: X(t)=[x(tᵢ)] In the above formula, x(tᵢ) is the x-coordinate of the feature point in the i-th frame of the image at time t, and i=1,2,…n, where n is the total number of frames; The oscillation displacement Δxᵢ of adjacent frames is calculated using the following formula to obtain the oscillation displacement sequence at time t: Δxᵢ=x(tᵢ)-x(t i-1 ) By directly characterizing the dynamic oscillation amplitude of the feature point relative to the reference position using an oscillating displacement sequence, the conclusion is drawn that the feature point and the changing trend of pulsating pressure are linearly positively correlated.

5. The method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data as described in claim 4, characterized in that, In S2, the differential processing refers to performing a first-order backward differential processing on the oscillating displacement sequence using the following formula to obtain the oscillating displacement difference sequence Δx'(t) at time t, where Δx'(t) = [Δx'ᵢ]: Δx'ᵢ=Δxᵢ -Δxᵢ₋1 In the above formula, Δx'ᵢ is the oscillation displacement difference between adjacent frames, and when i=1, Δx'1=Δx1.

6. The method for frequency domain analysis of pulsating pressure based on transient schlieren feature point time-series data as described in claim 1, characterized in that, In S3, the frequency resolution Δf during the Fast Fourier Transform is obtained by the following formula: Δf=Fs / M In the above formula, Fs is the acquisition frame rate of the high-speed schlieren acquisition system, M is the signal length, and M=n-1, where n is the total number of frames.

Images