Linear array data processing method for surface pulse pressure wave number-frequency spectrum measurement

Abstract

The invention discloses a linear array data processing method for surface pulse pressure wave number-frequency spectrum measurement. A one-dimensional linear array is adopted to carry out a pulsating pressure measurement test, a model surface time domain signal is obtained, an array signal cross-spectrum matrix is calculated, and a flow direction wave number-frequency spectrum result is obtained through calculation by taking the cross-spectrum matrix as input. The method is directly obtained through a turbulent boundary layer pulsating pressure flow direction wave number-frequency spectrum physical definition formula and has direct relevance with the physical definition formula, the cross spectrum matrix method is applied, the calculation and analysis capacity is greatly improved, parallel calculation can be adopted in cross spectrum matrix calculation, a direct discrete algorithm and an average discrete algorithm, and the calculation speed is greatly improved.

Description

Technical Field

[0001] The invention relates to the field of wind tunnel tests, and in particular to a line array data processing method for surface pulse pressure wave number-frequency spectrum measurement. Background Art

[0002] The pulsating pressure of the turbulent boundary layer is a problem of great concern in the fields of air, space, sea and ground rapid transportation systems. On the one hand, it will directly radiate noise outward, and on the other hand, it will cause flow-induced vibration of the wall structure, resulting in structural fatigue and the formation of secondary sound sources. Conducting flight, wind tunnel, water tunnel, self-propelled model and other tests is the most important means to obtain the characteristics of the pulsating pressure of the turbulent boundary layer. Among them, the application of the surface array method has effectively improved the experimental capabilities in this regard. The one-dimensional line array is an important device for measuring the wave number-frequency spectrum in the flow direction. By collecting multi-channel pulsating pressure signals along the incoming flow direction, the wave number-frequency spectrum in the flow direction is obtained by using data processing methods. The core of the data processing method is the mathematical discretization algorithm of the physical definition formula. The pulsating pressure data obtained by the one-dimensional line array is processed by a computer using this algorithm to obtain the measured flow wave number-frequency spectrum. However, there are some shortcomings in the current data processing method. Some mathematical discrete methods are not sufficiently related to the physical definition, and the measured results are not wavenumber-frequency spectra; some algorithms are too old and still use the most primitive time-space Fourier transform direct algorithm (such as patent application number 202410305380.3), and cannot apply the latest cross-spectrum matrix calculation method that can be parallelized to increase the speed. Therefore, for one-dimensional linear arrays, the data processing method still needs to be improved. Summary of the invention

[0003] The purpose of the present invention is to propose a new data processing method to address the shortcomings of the current one-dimensional model surface line array data processing method for turbulent boundary layer pulsating pressure wave number-frequency spectrum measurement.

[0004] The present invention adopts the following technical solutions: A method for designing a two-dimensional array formation for measuring turbulent boundary layer pulsating pressure based on a particle swarm optimization algorithm comprises the following steps: Step 1: Use a one-dimensional linear array to carry out a pulsating pressure measurement test to obtain the time domain signal of the model surface; Install a one-dimensional line array with M channels on the surface of the test model, connect the lines, data acquisition equipment, and host computer, and conduct wind tunnel / flight / underwater tests after checking that they are correct, and collect relevant pulsating pressure signals.

[0005] Step 2: Calculate the cross-spectral matrix of array signals; Let the cross-spectral matrix be expressed as , that is dimensional matrix whose elements are cross-spectral functions ,in m , n Representative m , n The sensors are defined as follows: The time-space correlation function of the pulsating pressure at any two points is expressed as: (1) (1) Where p is the pulsating pressure, and Represented as the spatial position vector coordinates of any two points and is the coordinate difference, and Indicates two points in time. “<>” indicates ensemble average, and “*” indicates conjugate complex number. When the TBL pulse pressure is steady and isotropic, the formula is the same as and No relation, It can be written as , then the cross spectrum between a and b can be expressed as: (2) (2) In the formula, when When it is zero, it is transformed into the self-spectrum, that is, becomes ; C is the experimental coefficient, which is 1 or The wavenumber-frequency spectrum is the Fourier transform of the time-space correlation function in the time and space domains, that is: (3) here is the wave number vector, and is the wave number in the stream direction and span direction. Similarly, for . is an exponent, which can be 0, 1, 2 or 3 depending on the Fourier transform dimension. Here we have the following relationship: (4) In addition, formula (3) can also be written as: (5) The cross-spectrum matrix (CSM) contains the cross-spectrum between all sensors. The cross spectrum is: (6) In addition, if the exchange m , n Position, the following relationship exists: (7) Right now and is complex conjugate. In order to realize computer parallel computing, equation (6) is discretized, then: (8) in, N is the data block length, Represents hope, If necessary, parallel computing can be used to calculate each Improves computation efficiency.

[0006] Step 3: Using the cross-spectral matrix as input, calculate the streamwise wavenumber-frequency spectrum based on the direct discrete algorithm; For the one-dimensional wave number-frequency spectrum, that is, the streamwise wave number-frequency spectrum (3) is transformed into: (9) Equation (9) is the physical definition of the wavenumber-frequency spectrum measurement obtained by the linear array. In order to carry out the measurement, Equation (9) needs to be discretized. Based on the isotropic assumption, the discretization requires all different The sum is calculated, where is the array sensor spacing. According to the homogeneity assumption, when the distance between two sensors is the same (i.e. same), no matter which two sensors, The same, so only one of them needs to be selected. In summary, the discrete form of formula (9) can be expressed as: (10) Right now: (11) The above formula is the core formula of the direct discrete algorithm of wave number-frequency spectrum, which can be used to calculate the wave number-frequency spectrum. If necessary during the calculation process (such as a large number of sensors), a parallel calculation method can be used to obtain each value after the summation symbol, and then sum it up to improve the calculation efficiency.

[0007] Step 4: Using the cross-spectral matrix as input, calculate the streamwise wavenumber-frequency spectrum results based on the average discrete algorithm; The isotropy assumption is a theoretical result. In actual measurement, when two sensors are at the same distance, the cross-spectrum measurement results are still different. To further reduce the correlation effect, the cross-spectrum values ​​of sensors at the same distance from different groups are averaged and then brought into the connection algorithm to reduce the error effect. For example: (12) Therefore, formula (11) can be optimized as follows: (13) The above formula is the core formula of the wavenumber-frequency spectrum average discrete algorithm, which can be used to calculate the wavenumber-frequency spectrum. Similarly, this method can also use parallel computing to improve the calculation efficiency.

[0008] Step 5: Compare the results obtained by the two discrete algorithms in step 4 and step 5, and select the optimal value as the final measurement result.

[0009] Compare the results of the direct discretization algorithm with the average discretization algorithm, and select the appropriate result as the one-dimensional wave number-frequency spectrum array output. The specific selection criteria can be selected according to engineering requirements, such as comparison with the classic prediction model.

[0010] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: It is directly obtained from the physical definition of the wave number-frequency spectrum of the turbulent boundary layer pulsating pressure flow, and is directly related to the physical definition; the cross-spectrum matrix method is used to greatly improve the calculation and analysis capabilities. In addition, if necessary, parallel computing can be used in the cross-spectrum matrix calculation, direct discrete algorithm, and average discrete algorithm to greatly improve the calculation speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The present invention will now be described by way of example with reference to the accompanying drawings, in which: Figure 1 is a schematic diagram of a line array. Figure 2 Line array data processing method, Figure 3 The one-dimensional wavenumber-frequency spectrum measurement results of the direct discrete algorithm, Figure 4 The one-dimensional wavenumber-frequency spectrum measurement results of the average discrete algorithm, Figure 5 The direct discrete algorithm one-dimensional wave number-frequency spectrum measurement results are compared with the Chase I model. Figure 6 The one-dimensional wavenumber-frequency spectrum measurement results of the average discrete algorithm are compared with the Chase I model; Among them, 1 represents the total number of sensors M , 2 represents the sensor spacing . DETAILED DESCRIPTION

[0012] All features disclosed in this specification, or steps in all methods or processes disclosed, except mutually exclusive features and / or steps, can be combined in any manner.

[0013] Any feature disclosed in this specification (including any additional claims, abstract and drawings), unless otherwise stated, may be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

[0014] The specific implementation mode of the present invention is described below with reference to the accompanying drawings, and the specific implementation steps are shown in FIG1 .

[0015] Step 1: Use a one-dimensional linear array to carry out a pulsating pressure measurement test to obtain the time domain signal of the model surface.

[0016] As shown in Figure 2, this is a number of M= 32 line array, the sensor spacing The pulsating pressure data was collected through wind tunnel test, with sampling frequency of 51.2kHz, wind speed of 30.7m / s and sampling time of 30s.

[0017] Step 2: Calculate the cross-spectral matrix of array signals.

[0018] For 32-channel data, the cross-spectrum matrix is ​​calculated. Each element in the matrix is ​​the cross-spectrum between two sensors. During the calculation process, the block length is set to 8192, the overlap rate is 50%, and the Hanning window is used as the window function to calculate and obtain the cross-spectrum matrix Note that the parallel computing method is used here to improve the computing efficiency, that is, the different elements in the matrix are calculated in parallel and then assigned to the matrix after completion.

[0019] Step 3: Using the cross-spectral matrix as input, calculate the streamwise wavenumber-frequency spectrum based on the direct discrete algorithm.

[0020] Substitute the data into the above formula (11) to obtain the calculation result of the direct discrete algorithm, where M= 32. Note that the summation here uses parallel computing to improve the computational efficiency. Here C is Figure 3 is the calculation result of the direct discrete algorithm.

[0021] Step 4: Using the cross-spectral matrix as input, calculate the streamwise wavenumber-frequency spectrum results based on the average discrete algorithm.

[0022] Substitute the data into the above formula (13) to obtain the average discrete algorithm calculation result. Parallel computing is also used here to improve the calculation efficiency. Figure 4 This is the calculation result of the direct discrete algorithm.

[0023] Step 5: Compare the results obtained by the two discrete algorithms in step 4 and step 5, and select the optimal value as the final measurement result.

[0024] To compare the calculation results of the two algorithms, we compared them with the classic Chase I model. The wavenumber-frequency spectrum below is compared with the Chase I model. Figure 5 , Figure 6 The one-dimensional wavenumber-frequency spectrum measurement results obtained by the direct discrete algorithm and the average discrete algorithm are compared with the Chase I model, where the black realization is the prediction result of the Chase I model and the scattered points are the experimental results. As can be seen from the figure, the average discrete method result has a smaller fluctuation, which is better than the direct discrete algorithm result, so the average discrete method result is selected as the measurement output.

[0025] The above specific implementation methods are also applicable to flight, wind tunnel, water tunnel, self-propelled model and other tests.

[0026] The present invention is not limited to the above-mentioned specific embodiments, but extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed.

Claims

1. A linear array data processing method for surface pulse pressure wave number-frequency spectrum measurement, characterized in that The following steps are involved: S1: Use a one-dimensional linear array to carry out a pulsating pressure measurement test to obtain the time domain signal of the model surface; S2: Calculate the cross-spectral matrix of array signals; S3: Taking the cross-spectral matrix as input, the flow wavenumber-frequency spectrum is calculated based on the direct discrete algorithm and the result is: ， S4: Taking the cross-spectral matrix as input, the streamwise wavenumber-frequency spectrum is calculated based on the average discrete algorithm and the result is: ， S5: Compare the results obtained by the two discrete algorithms S3 and S4, and select the optimal value as the final measurement result according to the degree of agreement between the results of the direct discrete algorithm and the average discrete algorithm and the corresponding prediction model; in: is the flow wave number, is the angular frequency, is the experimental coefficient, is the array sensor spacing; is the number of sensors; Represents the corresponding Sensors; Refers to the autopower spectrum of the first sensor; Representative No. 1, The cross power spectrum between sensors; Representative , the cross power spectrum between 1 sensors, For the The autospectrum of each sensor, and For the and The cross spectrum between the sensors is the complex conjugate of the two.

2. A linear array data processing method for surface pulse pressure wavenumber-frequency spectrum measurement according to claim 1, characterized in that: In the calculation process of S3 and S4, the parallel calculation method is first used to obtain each value after the summation symbol, and then the sum is calculated.

Images