Multi-parameter identification method for rotating stall of centrifugal compressor

Abstract

The invention relates to the technical field of compressor state monitoring, and particularly provides a multi-parameter identification method for rotating stall of a centrifugal compressor. The method comprises the following steps: acquiring pressure pulsation data according to a pressure sensor of the centrifugal compressor; according to a sparse dictionary corresponding to the centrifugal compressor, sparse reconstruction is conducted on the pressure pulsation data, a stall sensing signal is obtained, and the sparse dictionary is obtained through training according to the pressure pulsation data collected by a pressure sensor of the centrifugal compressor; according to the stall sensing signal, identifying a phase parameter of stall dynamic propagation through a two-channel fusion and space phase identification method; and calculating a plurality of target parameters of the centrifugal compressor according to the phase parameters. The problems that a centrifugal compressor parameter identification method in the prior art is low in accuracy and low in efficiency are solved.

Description

Technical Field

[0001] The invention relates to the technical field of compressor status monitoring, and in particular to a multi-parameter identification method for rotational stall of a centrifugal compressor. Background Art

[0002] Centrifugal compressors are core power equipment in important fields such as petrochemicals, natural gas, coal chemical industry, and aerospace. With the upgrading of related equipment, centrifugal compressors need to develop in the direction of high performance and large-scale, and their aerodynamic design requirements are also improved accordingly. At the same time, the actual working conditions of the impeller, the core component of the centrifugal compressor, have become more severe, the internal flow field pulsation under non-design conditions has intensified, the problem of impeller vibration caused by fluid excitation has become prominent, and accidents caused by impeller fatigue damage and fracture have occurred frequently.

[0003] In transient operation, centrifugal compressors may experience flow instability, leading to rotating stall, which affects the stable operation of the impeller and the entire machine. Rotating stall is a complex unsteady flow instability problem, with variable state parameters, which is difficult to accurately solve and describe. For rotating stall parameter identification, the pressure waveform analysis method is currently mainly used. However, due to the complex unsteady flow in centrifugal compressors and the harsh service environment, the pressure pulsation signal will be disturbed by noise, and there are problems such as weak rotating stall characteristics, low identification accuracy, low manual efficiency, and reliance on expert experience, which limits its operating status monitoring and diagnostic operation and maintenance.

[0004] The centrifugal compressor parameter identification method in the related art has the problems of low accuracy and low efficiency, and no effective solution has been proposed so far. Summary of the invention

[0005] The invention provides a multi-parameter identification method for rotating stall of a centrifugal compressor, which at least solves the problems of low accuracy and low efficiency in the centrifugal compressor parameter identification method in the related art.

[0006] The invention provides a multi-parameter identification method for rotational stall of a centrifugal compressor, comprising: obtaining pressure pulsation data according to a pressure sensor of the centrifugal compressor; sparsely reconstructing the pressure pulsation data according to a sparse dictionary corresponding to the centrifugal compressor to obtain a stall perception signal, wherein the sparse dictionary is trained based on the pressure pulsation data collected by the pressure sensor of the centrifugal compressor; identifying phase parameters of dynamic propagation of the stall according to the stall perception signal through dual-channel fusion and spatial phase identification methods; and calculating multiple target parameters of the centrifugal compressor according to the phase parameters.

[0007] The multi-parameter identification method for rotating stall of a centrifugal compressor provided by the embodiment of the present invention performs sparse reconstruction according to the sparse dictionary corresponding to the centrifugal compressor, and before obtaining the stall sensing signal, the method further includes: obtaining pressure pulsation signals of multiple measuring points as training data according to multiple pressure sensors, wherein the multiple pressure sensors are arranged on the leading edge channel of the diffuser of the centrifugal compressor, the interval between two adjacent pressure sensors is not less than one blade channel, and the intervals of the multiple pressure sensors are different; defining relevant parameters of the sparse representation algorithm, including the dictionary length, sparsity, and number of iterations of the initial dictionary, wherein the sparse representation algorithm is used to perform sparse dictionary learning according to the training data and train the dictionary atoms of the sparse dictionary; according to the training data and the corresponding initial dictionary, the sparse representation of the training data under the initial dictionary is found through the orthogonal matching pursuit algorithm; the initial dictionary is iteratively updated according to the sparse representation, and a trained sparse dictionary is obtained when the number of iterations reaches the defined number of iterations.

[0008] The multi-parameter identification method of rotating stall of a centrifugal compressor provided in an embodiment of the present invention finds the sparse representation of the training data under the initial dictionary through an orthogonal matching pursuit algorithm based on training data and a corresponding initial dictionary, including: segmenting different pressure pulsation signals in the training data according to the dictionary length of the initial dictionary to obtain data column vectors, and the mean and mean normalization matrix of the data column vectors; using the orthogonal matching pursuit algorithm to solve the sparse coding vector corresponding to each column vector in the mean normalization matrix; generating a sparse coefficient matrix of the pressure pulsation signal under the initial dictionary based on the sparse coding vectors of multiple column vectors in the mean normalization matrix.

[0009] The multi-parameter identification method for rotating stall of a centrifugal compressor provided by the embodiment of the present invention iteratively updates the initial dictionary according to the sparse representation, and obtains the trained sparse dictionary when the number of iterations reaches the defined number of iterations, including: calculating the residual matrix of the initial dictionary according to the sparse coefficient matrix; determining the position set of the non-zero items of the row vector of the sparse coefficient matrix according to the residual matrix; selecting the column elements of the positions corresponding to the non-zero items in the residual matrix according to the position set to form a non-zero matrix, performing singular value decomposition on the non-zero matrix to obtain a singular value matrix, a dictionary assignment matrix, and a system. The method comprises the following steps: a first step is to update the column vector of the initial dictionary according to the first column elements of the dictionary assignment matrix; a second step is to update the row vector of the sparse coefficient matrix according to the product of the non-zero singular value of the first column of the singular value matrix and the first column elements of the coefficient assignment matrix; a third step is to update the row vector of the sparse coefficient matrix by traversing and updating multiple columns of the initial dictionary and multiple rows of the sparse coefficient matrix, and increase the number of iterations by one; when the number of iterations reaches the defined number of iterations, the last updated dictionary is used as the sparse dictionary, and the corresponding sparse coefficient matrix is ​​recorded.

[0010] The multi-parameter identification method for rotating stall of a centrifugal compressor provided in an embodiment of the present invention performs sparse reconstruction on pressure pulsation data according to a sparse dictionary corresponding to the centrifugal compressor to obtain a stall perception signal, including: using the dictionary length of the sparse dictionary to segment the pressure pulsation data to obtain a corresponding mean normalization matrix; using an orthogonal matching pursuit algorithm to solve the mean normalization matrix to obtain a sparse coefficient matrix of the pressure pulsation data under the sparse dictionary; multiplying the sparse coefficient matrix with the pressure pulsation data to generate a sparsely reconstructed stall perception signal.

[0011] The invention provides a multi-parameter identification method for rotating stall of a centrifugal compressor, wherein the phase parameters include a target phase shift; according to the stall sensing signal, the phase parameters of the dynamic propagation of the stall are identified through a dual-channel fusion and a spatial phase identification method, including: calculating the mean and standard deviation of the stall sensing signal, and determining the pressure pulsation signal after de-averaging; calculating the phase shift corresponding to the dual-channel signal and the correlation coefficient corresponding to the phase shift, wherein the dual-channel signal is the pressure pulsation signal of two different pressure sensors; dynamically normalizing the correlation coefficient through the standard deviation value of the dual-channel signal to obtain a phase characterization function, wherein the phase characterization function is used to characterize the functional relationship between the phase shift and the zero-mean normalized cross-correlation value; according to the phase characterization function, by solving the local extreme value, the target phase shift of the dual-channel signal when the phase characterization function value is maximum is obtained.

[0012] The invention provides a multi-parameter identification method for rotating stall of a centrifugal compressor, wherein multiple target parameters include the number of stall clusters, the circumferential propagation velocity of the stall clusters, and the propagation direction of the stall clusters relative to the impeller; multiple target parameters of the centrifugal compressor are calculated according to the phase parameters, including: determining the relationship between the number of stall clusters, the circumferential propagation velocity and the impeller rotation frequency of the centrifugal compressor, and the non-integer order of the rotating stall; obtaining the phase characterization function, the circumferential spatial angle of two pressure sensors corresponding to the dual-channel signal, and the impeller rotation frequency, and calculating the circumferential propagation velocity of the stall cluster; obtaining the non-integer order of the rotating stall, the circumferential propagation velocity and the impeller rotation frequency, and calculating the number of stall clusters according to the relationship; determining the propagation direction of the stall cluster according to the positive or negative of the target phase shift corresponding to the maximum value of the phase characterization function corresponding to the dual-channel signal.

[0013] The invention provides a multi-parameter identification method for rotating stall of a centrifugal compressor, which determines the relationship between the number of stall clusters, the circumferential propagation velocity and the impeller rotation frequency of the centrifugal compressor, and the non-integral order of the rotating stall, including: obtaining a stall sensing signal, performing a spectrum analysis on the stall sensing signal, and obtaining a stall characteristic frequency, wherein the stall characteristic frequency is the product of the number of stall clusters and the circumferential propagation velocity of the stall cluster; normalizing the stall characteristic frequency using the impeller rotation frequency of the centrifugal compressor, and associating the stall characteristic frequency with the impeller rotation frequency to obtain a relationship, as shown in the following formula:

[0014] ;

[0015] In the formula, is the non-integral order of rotating stall, is the number of stall groups, is the impeller rotation frequency, is the circumferential propagation velocity, The parameter is based on the identification result obtained by measuring the signal in the stationary / absolute coordinate system, and defines the stall propagation direction as the same as the impeller rotation direction. Is positive.

[0016] An electronic device provided by an embodiment of the present invention includes: a processor, and a memory storing a program, wherein the program includes instructions, and when the instructions are executed by the processor, the processor executes any of the above methods.

[0017] The present invention provides a computer program product according to an embodiment of the present invention, including a computer program / instruction, wherein when the computer program / instruction is executed by a processor, any of the above methods is implemented.

[0018] The invention provides a multi-parameter identification method for rotating stall of a centrifugal compressor. The sparse dictionary for the application scenario of the centrifugal compressor and the corresponding sensor is pre-trained, so that the sparse dictionary has better processing effect and accuracy for the pressure pulsation data in the scenario of the centrifugal compressor and the sensor. The pressure pulsation data acquired by the pressure sensor is sparsely reconstructed using the trained sparse dictionary, which can reduce the influence of noise on the perception signal and obtain a stall perception signal with higher accuracy. The phase parameters of the dynamic propagation of the stall are identified based on the stall perception signal, and multiple target parameters are identified based on the phase parameters, so that the parametric characterization results of the stall state in the centrifugal compressor can be obtained, and online monitoring can be realized. This solves the problems of low accuracy and low efficiency in the centrifugal compressor parameter identification method in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] 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 use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other embodiments can be obtained based on these drawings without creative work.

[0020] Figure 1 It is a flow chart of the steps of a multi-parameter identification method for rotating stall of a centrifugal compressor in an embodiment of the present invention.

[0021] Figure 2 It is a schematic diagram of the installation position of the pressure sensor in the embodiment of the invention.

[0022] Figure 3 It is a schematic diagram of the installation positions of the P1-P5 pressure pulsation sensors used for testing in the embodiments of the present invention.

[0023] Figure 4 It is a schematic diagram of the original signal of pressure pulsation obtained by the P3 pressure pulsation sensor in the embodiment of the invention.

[0024] Figure 5 yes Figure 4 Schematic diagram of the envelope spectrum corresponding to the original signal shown.

[0025] Figure 6 It is a schematic diagram of the envelope spectrum of the stall sensing signal corresponding to the P3 pressure pulsation sensor in the embodiment of the invention.

[0026] Figure 7 It is the phase characterization function in the embodiment of the invention Schematic diagram of the experimental results of the propagation direction of the stall group.

[0027] Figure 8 The phase characterization function in the embodiment of the invention is Schematic diagram of the test results of the maximum and average values.

[0028] Fig. 9 It is a schematic diagram of the test results of the number of stall groups in the embodiment of the invention.

[0029] Fig.10 It is a schematic diagram of the structure of an electronic device in an embodiment of the invention. DETAILED DESCRIPTION

[0030] The embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.

[0031] The pressure waveform analysis method is currently mainly used to identify the rotating stall parameters of centrifugal compressors. However, due to the complex unsteady flow in centrifugal compressors and the harsh service environment, the pressure pulsation signal will be disturbed by noise, and there are problems such as weak rotating stall characteristics, low identification accuracy, low manual efficiency and reliance on expert experience, which limits its operating status monitoring and diagnostic operation and maintenance.

[0032] To this end, the present invention provides an embodiment of a multi-parameter identification method for rotating stall of a centrifugal compressor.

[0033] Please refer to Figure 1 As shown, the above method includes:

[0034] Step S101, obtaining pressure pulsation data according to the pressure sensor of the centrifugal compressor.

[0035] Step S102, sparsely reconstructing the pressure pulsation data according to the sparse dictionary corresponding to the centrifugal compressor to obtain a stall sensing signal, wherein the sparse dictionary is trained based on the pressure pulsation data collected by the pressure sensor of the centrifugal compressor.

[0036] Step S103: According to the stall sensing signal, the phase parameter of the stall dynamic propagation is identified by using a dual-channel fusion and spatial phase identification method.

[0037] Step S104: Calculate multiple target parameters of the centrifugal compressor according to the phase parameters.

[0038] It is understandable that the pressure pulsation data refers to a set of measured values ​​of periodic or non-periodic fluctuations of pressure over time in the internal flow channel of the centrifugal compressor, such as unsteady air flow pulsation data.

[0039] Sparse dictionary is the core content of sparse representation (K-Singular Value Decomposition, K-SVD for short). This algorithm is mainly used in many fields such as signal processing, image denoising, image classification, etc. It is an improvement on the traditional singular value decomposition (SVD) and is used to find the sparsest representation of the signal given a dictionary.

[0040] The target parameters include but are not limited to the circumferential propagation speed of the stall cluster, the number of stall clusters, and the propagation direction of the stall cluster. The stall cluster refers to the area of ​​local airflow stagnation or reverse flow in the flow field inside the impeller. Specifically, when a centrifugal compressor has a rotational stall, the airflow in the impeller no longer flows evenly and stably, but forms relatively stable low-speed or reverse flow areas in certain areas, which are called stall clusters.

[0041] The above method provided by the embodiment of the invention pre-trains a sparse dictionary in the application scenario of a centrifugal compressor and a corresponding sensor, so that the sparse dictionary has better processing effect and accuracy for the pressure pulsation data in the scenario of the centrifugal compressor and the sensor, and uses the trained sparse dictionary to sparsely reconstruct the pressure pulsation data obtained by the pressure sensor, which can reduce the influence of noise on the perception signal and obtain a stall perception signal with higher accuracy; based on the stall perception signal, the phase parameters of the stall dynamic propagation are identified, and based on the phase parameters, multiple target parameters are identified, so that the parametric characterization result of the stall state in the centrifugal compressor can be obtained, and online monitoring can be realized.

[0042] Preferably, before step S101, the sparse dictionary needs to be trained, specifically:

[0043] Step S001, obtaining pressure pulsation signals of multiple measuring points as training data according to multiple pressure sensors, wherein the multiple pressure sensors are arranged on the leading edge channel of the diffuser of the centrifugal compressor, the interval between two adjacent pressure sensors is not less than one blade channel, and the intervals between the multiple pressure sensors are different.

[0044] Step S002, defining relevant parameters of the sparse representation algorithm, including the dictionary length, sparsity, and number of iterations of the initial dictionary, wherein the sparse representation algorithm is used to perform sparse dictionary learning based on training data and train dictionary atoms of the sparse dictionary.

[0045] Step S003, based on the training data and the corresponding initial dictionary, find the sparse representation of the training data under the initial dictionary through the orthogonal matching pursuit algorithm.

[0046] Step S004, iteratively updating the initial dictionary according to the sparse representation, and obtaining a trained sparse dictionary when the number of iterations reaches a defined number of iterations.

[0047] It is understandable that the centrifugal compressor is mainly composed of the air inlet chamber, impeller, diffuser, bend, return flow device, volute, etc. Among them, the impeller is one of the most important parts of the centrifugal compressor. Its function is to do work on the gas to increase the speed and pressure of the gas. When the gas flows out of the impeller at high speed, it enters the diffuser.

[0048] The main function of the diffuser is to effectively convert the kinetic energy of the high-speed airflow at the impeller outlet into pressure energy. Specifically, the gas velocity at the impeller outlet is very high and has a large kinetic energy. The diffuser reduces the gas velocity. According to the law of conservation of energy, this part of the kinetic energy will be converted into pressure energy, thereby increasing the gas pressure.

[0049] For example, please refer to Figure 2 As shown, a pressure sensor is arranged on the leading edge channel of the diffuser of the centrifugal compressor. The pressure sensor is locked on the cover plate of the leading edge channel of the diffuser and is flush with the inner wall surface of the cover plate.

[0050] Preferably, the pressure sensor is a pressure pulsation sensor, which has a wide frequency response range and can accurately capture rapid changes in pressure in a short period of time. The number of the pressure sensors is greater than or equal to 2.

[0051] Specifically, for a given rotating stall signal and dictionary, the sparse representation algorithm first finds the sparse representation of each signal under the dictionary through the orthogonal matching pursuit (OMP) algorithm, so that each signal can be approximated by as few atoms as possible in the dictionary. After obtaining the sparse representation, the sparse representation algorithm optimizes the dictionary quality by iteratively updating the atoms of the dictionary. Specifically, the other atoms in the dictionary are fixed one by one, only the currently considered atom is updated, and the optimal update direction is found by minimizing the reconstruction error. During the update process, singular value decomposition is used to find the best approximate representation to update the current atom of the dictionary.

[0052] The objective optimization function of the above sparse representation algorithm can be expressed as:

[0053] ;

[0054] In the formula, represents the training sample set of the above training data, represents the initial dictionary, represents a sparse coefficient matrix, represents the estimation error calculated according to the F norm, represents the i-th column of a sparse sparse matrix, Indicates sparsity.

[0055] Preferably, in step S003, based on the training data and the corresponding initial dictionary, an orthogonal matching pursuit algorithm is used to find a sparse representation of the training data under the initial dictionary, including:

[0056] Step S0031, according to the dictionary length of the initial dictionary, the different pressure pulsation signals in the training data are segmented to obtain data column vectors, as well as the mean of the data column vectors and the mean normalization matrix. Specifically:

[0057] ;

[0058] In the formula, represents the mean of the column vector of the i-th column data, Indicates the dictionary length of the initial dictionary, represents the jth element in the i-th column data column vector, Represents the mean normalization matrix of the i-th column data vector, Represents the column vector of the i-th column data.

[0059] Among them, from left to right, with a step size of 1 and a window size of The different pressure pulsation signals in the training data are segmented in a ×1 manner to obtain data column vectors.

[0060] Step S0032, using the orthogonal matching pursuit algorithm to solve the sparse coding vector corresponding to each column vector in the mean normalization matrix.

[0061] For example, the orthogonal matching pursuit algorithm is used to solve the i-th column data vector The corresponding sparse coding vector .

[0062] Step S0033, generating a sparse coefficient matrix of the pressure pulsation signal under the initial dictionary according to the sparse coding vectors of multiple column vectors in the mean normalization matrix .

[0063] Preferably, in step S004, the initial dictionary is iteratively updated according to the sparse representation, and when the number of iterations reaches a defined number of iterations, a trained sparse dictionary is obtained, including:

[0064] Step S0041, according to the sparse coefficient matrix , calculate the residual matrix of the initial dictionary , specifically:

[0065] ;

[0066] In the formula, and Represent the jth column vector and the jth column vector in the initial dictionary D respectively. Column vector, Represents a sparse coefficient matrix No. A row vector.

[0067] Step S0042, determining the position set of the non-zero items of the row vector of the sparse coefficient matrix according to the residual matrix, specifically:

[0068] According to the residual matrix For the first The column vector is updated and the sparse coefficient matrix is ​​determined The location set of the non-zero entries of the row vector .

[0069] Step S0043, based on the location set , select the residual matrix The column elements corresponding to the non-zero items in the form a non-zero matrix , for non-zero matrices Perform singular value decomposition SVD to obtain the singular value matrix , dictionary assignment matrix , coefficient assignment matrix , specifically:

[0070] ;

[0071] Step S0044, according to the first column element of the dictionary assignment matrix, the column vector of the initial dictionary is assigned and updated, specifically:

[0072] Assign matrix according to dictionary The first column element of the initial dictionary The atoms in the column vector of are updated column by column.

[0073] Step S0045, according to the singular value matrix The non-zero singular value Δ(1,1) of the first column of The first element in the The product of is used to update the row vector of the sparse coefficient matrix. Specifically:

[0074] ;

[0075] Step S0046, by traversing and updating multiple columns of the initial dictionary and multiple rows of the sparse coefficient matrix, an updated dictionary and sparse coefficient matrix are obtained, and the number of iterations is increased by one.

[0076] Step S0047, when the number of iterations reaches the defined number of iterations, the dictionary obtained by the latest update is used as a sparse dictionary, and the corresponding sparse coefficient matrix is ​​recorded.

[0077] It is understandable that the specific value of the defined number of iterations can be determined by those skilled in the art based on empirical values ​​and equipment operating parameter requirements.

[0078] It is understandable that, in actual application, the pressure pulsation data obtained in step S101 does not need to be processed according to the training step, which refers to the step of processing the pressure pulsation signal as training data in the above steps S001 to S004.

[0079] Preferably, in step S102, according to the sparse dictionary corresponding to the centrifugal compressor, the pressure pulsation data is sparsely reconstructed to obtain the stall sensing signal, including:

[0080] Step S1021 , using the dictionary length of the sparse dictionary, segment the pressure pulsation data to obtain a corresponding mean normalization matrix.

[0081] Step S1022: using an orthogonal matching pursuit algorithm to solve the mean normalization matrix, and obtaining a sparse coefficient matrix of the pressure pulsation data under a sparse dictionary.

[0082] Step S1023, multiplying the sparse coefficient matrix by the pressure pulsation data to generate a sparsely reconstructed stall perception signal.

[0083] Preferably, in step S103, the phase parameter includes a target phase shift.

[0084] According to the stall sensing signal, the phase parameters of the stall dynamic propagation are identified through dual-channel fusion and spatial phase identification methods, including:

[0085] Step S1031, calculating the mean value of the stall sensing signal Z and standard deviation , determine the pressure pulsation signal after removing the mean , specifically:

[0086] ;

[0087] Step S1032, calculating the phase shift n and the correlation coefficient Γ corresponding to the dual-channel signal, wherein the dual-channel signal is the pressure pulsation signal of two different pressure sensors. and , specifically:

[0088] ;

[0089] In the formula, ; I represents the number of sampling points within one rotation period of the impeller, specifically:

[0090] ;

[0091] In the formula, represents the signal sampling frequency, Indicates the impeller rotation frequency.

[0092] Here, the phase shift n can be understood as phase delay.

[0093] Step S1033, dynamically normalize the correlation coefficient by the standard deviation value of the dual-channel signal to obtain a phase characterization function, wherein the phase characterization function is used to characterize the functional relationship between the phase shift and the zero-mean normalized cross-correlation value, specifically:

[0094] ;

[0095] In the formula, To characterize the phase shift n and stall sensing signal and The phase characterization function of the zero-mean normalized cross-correlation (ZNCC); and Represents stall sensing signal and The corresponding standard deviation.

[0096] Step S1034, according to the phase characterization function, by solving the local extreme value, the target phase shift of the dual-channel signal is obtained when the phase characterization function value is the maximum.

[0097] It can be understood that since the stall perception signal is related to time, the phase characterization function associates the stall perception signal with the phase shift. After determining the target phase shift based on the phase characterization function, the corresponding instantaneous phase information can be determined based on the target phase shift.

[0098] Preferably, in step S104, the multiple target parameters include the number of stall clusters, the circumferential propagation speed of the stall clusters, and the propagation direction of the stall clusters relative to the impeller.

[0099] Based on the phase parameters, multiple target parameters of the centrifugal compressor are calculated, including:

[0100] Step S1041, determining the relationship between the number of stall clusters, the circumferential propagation velocity and the impeller rotation frequency of the centrifugal compressor, and the non-integral order of the rotating stall.

[0101] Step S1042, obtaining the phase characterization function, the circumferential spatial angle of the two pressure sensors corresponding to the dual-channel signals, and the impeller rotation frequency, and calculating the circumferential propagation speed of the stall group.

[0102] Step S1043, obtaining the non-integral order of rotating stall, the circumferential propagation velocity and the impeller rotation frequency, and calculating the number of stall clusters according to the relationship.

[0103] Step S1044, determining the propagation direction of the stall group according to the positive or negative sign of the target phase shift corresponding to the maximum value of the phase characterization function corresponding to the dual-channel signal.

[0104] Preferably, in step S1041, determining the relationship between the number of stall clusters, the circumferential propagation velocity, the impeller rotation frequency of the centrifugal compressor, and the non-integral order of the rotating stall includes:

[0105] A stall sensing signal is acquired, and a spectrum analysis is performed on the stall sensing signal to obtain a stall characteristic frequency, wherein the stall characteristic frequency is the product of the number of stall clusters and the circumferential propagation speed of the stall clusters.

[0106] The stall characteristic frequency is normalized by the impeller rotation frequency of the centrifugal compressor, and the stall characteristic frequency is associated with the impeller rotation frequency to obtain the relationship as follows:

[0107] ;

[0108] In the formula, is the non-integral order of rotating stall, is the number of stall groups, is the impeller rotation frequency, is the circumferential propagation velocity, The parameter is based on the identification result obtained by measuring the signal in the stationary / absolute coordinate system, and defines the stall propagation direction as the same as the impeller rotation direction. Is positive.

[0109] Furthermore, based on the phase characterization function Determine the phase shift n, the circumferential propagation velocity Can be matched with the impeller rotation frequency The specific relationship is:

[0110] ;

[0111] In the formula, Indicates pressure pulsation signal and The circumferential spatial angle between the corresponding two pressure sensors.

[0112] It can be understood that based on the above circumferential propagation speed The impeller rotation frequency The relationship between Identify and establish the phase shift n and the circumferential space angle The quantitative relationship between them.

[0113] Furthermore, based on the circumferential propagation velocity and rotating stall non-integral order , the number of stall groups can be identified :

[0114] ;

[0115] In addition, the numbers of the two adjacent pressure sensors are defined as p and q, respectively, wherein the pressure sensor located before the other pressure sensor in the impeller rotation direction is numbered p. Based on the physical law of dynamic propagation, the propagation direction of the stall group relative to the impeller is for:

[0116] ;

[0117] In the formula, sign (.) represents the sign function, Indicates taking the maximum value, Indicates the stall sensor signal corresponding to the pressure sensor numbered p, Indicates the stall sensor signal corresponding to the pressure sensor numbered q.

[0118] In summary, the above method provided by the embodiment of the present invention can identify the circumferential propagation speed of the stall group. 、Number of stall groups , the propagation direction of the stall group .

[0119] Exemplarily, the present invention also provides test data of the above method to verify its technical effect, which is as follows:

[0120] On the centrifugal compressor test bench, taking a vaned diffuser with 20 blades as an example, the pressure pulsation on the side wall of the diffuser channel cover was measured with a sampling frequency of f s Set to 10kHz.

[0121] The installation position of the pressure sensor distributed along the leading edge is as follows: Figure 3 As shown, P1-P5 correspond to the pressure pulsation sensors numbered 1-5, each pressure pulsation sensor is installed in the middle of the leading edge channel, and the blade channel interval angle δθ=360 o / 20=18 o .

[0122] Take the P1 pressure pulsation sensor as the reference position (0 o ), P2-P5 pressure pulsation sensors are arranged in a counterclockwise direction, and their circumferential positions are: 72 o , 144 o , 180 o , 288 oThe circumferential spatial angles between adjacent pressure pulsation sensors are 4δθ, 4δθ, 2δθ, and 6δθ, respectively, and the circumferential spatial angle between the P1 pressure pulsation sensor and the P5 pressure pulsation sensor is 360°-288°=72°.

[0123] Based on the pressure pulsation sensor mentioned above, the pressure pulsation data is obtained. Taking the amplitude of the pressure pulsation obtained by the P3 pressure pulsation sensor as an example, please refer to Figure 4 As shown; perform spectrum analysis and obtain the corresponding envelope spectrum as shown Figure 5 shown.

[0124] Based on the above step S102, the amplitude of the pressure pulsation obtained by the P3 pressure pulsation sensor is sparsely reconstructed to obtain a corresponding stall sensing signal. The envelope spectrum of the stall sensing signal is as follows: Figure 6 shown.

[0125] contrast Figure 5 and Figure 6 It can be found that the amplitude of the pressure pulsation after sparse reconstruction has weakened the influence of noise to a certain extent, and the amplitude of the corresponding frequency under the rotating stall characteristic disturbance has been enhanced. Figure 5 and Figure 6 The red dots in the figure indicate that the amplitude of the corresponding frequency changes from Figure 5 About 70% of Figure 6 Among them, there are more than 100, and the growth rate is obvious.

[0126] Taking the stall sensing signals Z3 and Z4 of the P3 and P4 pressure pulsation sensors as examples, the corresponding stall propagation speed can be calculated by the above method. , Phase characterization function under the number of rotations and the propagation direction of the stall group, as shown in Figure 7 As shown, the phase characterization function The maximum and average values ​​of Figure 8 As shown in the figure, and the number of stall groups, as shown in the figure Fig. 9 shown.

[0127] It is understandable that the stall propagation speed , and the impeller rotation frequency Multiplying them together gives the circumferential propagation velocity of the stall group. .

[0128] In summary, the above method provided by the embodiment of the present invention can accurately identify the circumferential propagation speed of the stall group, the number of stall groups, and the propagation direction of the stall group.

[0129] The present invention provides a computer program product according to an embodiment of the present invention, including a computer program / instruction, wherein when the computer program / instruction is executed by a processor, any of the above methods is implemented.

[0130] The embodiments of the present invention also provide a non-transitory machine-readable medium storing a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to execute the method of the embodiments of the present invention.

[0131] The embodiments of the present invention further provide a computer program product, including a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to execute the method of the embodiments of the present invention.

[0132] The embodiment of the invention also provides an electronic device, comprising: at least one processor; and a memory connected to the at least one processor in communication. The memory stores a computer program that can be executed by the at least one processor, and the computer program is used to enable the electronic device to perform the method of the embodiment of the invention when executed by the at least one processor.

[0133] refer to Fig.10 , a block diagram of an electronic device that can be used as a server or client of an embodiment of the invention will now be described, which is an example of hardware devices that can be applied to various aspects of the invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the invention described and / or required herein.

[0134] like Fig.10 As shown, the electronic device includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1002 or a computer program loaded from a storage unit 1008 to a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the electronic device can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0135] Multiple components in the electronic device are connected to the I / O interface 1005, including: an input unit 1006, an output unit 1007, a storage unit 1008, and a communication unit 1009. The input unit 1006 can be any type of device that can input information to the electronic device, and the input unit 1006 can receive input digital or character information, and generate key signal input related to user settings and / or function control of the electronic device. The output unit 1007 can be any type of device that can present information, and can include but is not limited to a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. The storage unit 1008 can include but is not limited to a disk, an optical disk. The communication unit 1009 allows the electronic device to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks, and can include but is not limited to a modem, a network card, an infrared communication device, and / or a wireless communication transceiver, such as a Bluetooth device, a WiFi device, a WiMax device, a cellular communication device, and / or the like.

[0136] The computing unit 1001 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a CPU, a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing units, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 1001 performs the various methods and processes described above. For example, in some embodiments, the method embodiments created by the present invention may be implemented as a computer program, which is tangibly contained in a machine-readable medium, such as a storage unit 1008. In some embodiments, part or all of the computer program may be loaded and / or installed on an electronic device via ROM 1002 and / or a communication unit 1009. In some embodiments, the computing unit 1001 may be configured to perform the above method in any other appropriate manner (e.g., by means of firmware).

[0137] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the computer program, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The computer program can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0138] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable signal medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, or infrared systems, devices, or equipment, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0139] It should be noted that the term "including" and its variations used in the embodiments of the present invention are open inclusions, that is, "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments". The modifications of "one" and "multiple" mentioned in the embodiments of the present invention are illustrative and not restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0140] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0141] The various steps described in the method implementation methods provided by the embodiments of the present invention can be performed in different orders and / or in parallel. In addition, the method implementation methods may include additional steps and / or omit the steps shown. The scope of protection of the present invention is not limited in this respect.

[0142] The term "embodiment" in this specification refers to specific features, structures or characteristics described in conjunction with the embodiment that can be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily mean the same embodiment, nor does it mean that it is mutually exclusive with other embodiments and is independent or optional. The various embodiments in this specification are described in a related manner, and the same and similar parts between the various embodiments refer to each other. In particular, for the device, equipment, and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts refer to the partial description of the method embodiment.

[0143] The above-described embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of protection. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the attached claims.

Claims

1. A multi-parameter identification method for rotating stall of a centrifugal compressor, characterized in that: include: Obtain pressure pulsation data based on the pressure sensor of the centrifugal compressor; According to the sparse dictionary corresponding to the centrifugal compressor, the pressure pulsation data is sparsely reconstructed to obtain a stall sensing signal, wherein the sparse dictionary is trained based on the pressure pulsation data collected by the pressure sensor of the centrifugal compressor; According to the stall sensing signal, the phase parameter of the stall dynamic propagation is identified by dual-channel fusion and spatial phase identification method; A plurality of target parameters of the centrifugal compressor are calculated according to the phase parameter.

2. The method according to claim 1, characterized in that Before performing sparse reconstruction according to the sparse dictionary corresponding to the centrifugal compressor to obtain the stall sensing signal, the method further includes: According to the plurality of pressure sensors, pressure pulsation signals of a plurality of measuring points are obtained as training data, wherein the plurality of pressure sensors are arranged on the leading edge channel of the diffuser of the centrifugal compressor, the interval between two adjacent pressure sensors is not less than one blade channel, and the intervals between the plurality of pressure sensors are different; Defining relevant parameters of a sparse representation algorithm, including a dictionary length, sparsity, and number of iterations of an initial dictionary, wherein the sparse representation algorithm is used to perform sparse dictionary learning based on the training data and train dictionary atoms of the sparse dictionary; According to the training data and the corresponding initial dictionary, a sparse representation of the training data under the initial dictionary is found by an orthogonal matching pursuit algorithm; The initial dictionary is iteratively updated according to the sparse representation, and when the number of iterations reaches a defined number of iterations, the trained sparse dictionary is obtained.

3. The method according to claim 2, characterized in that According to the training data and the corresponding initial dictionary, a sparse representation of the training data under the initial dictionary is found by an orthogonal matching pursuit algorithm, including: According to the dictionary length of the initial dictionary, different pressure pulsation signals in the training data are segmented to obtain data column vectors, and means and mean normalization matrices of the data column vectors; Using an orthogonal matching pursuit algorithm, solving the sparse coding vector corresponding to each column vector in the mean normalization matrix; A sparse coefficient matrix of the pressure pulsation signal under the initial dictionary is generated according to the sparse coding vectors of the plurality of column vectors in the mean normalization matrix.

4. The method according to claim 3, characterized in that Iteratively updating the initial dictionary according to the sparse representation, and obtaining the trained sparse dictionary when the number of iterations reaches a defined number of iterations, comprises: Calculating a residual matrix of the initial dictionary according to the sparse coefficient matrix; Determine a position set of non-zero items of a row vector of the sparse coefficient matrix according to the residual matrix; According to the position set, column elements of positions corresponding to non-zero items in the residual matrix are selected to form a non-zero matrix, and singular value decomposition is performed on the non-zero matrix to obtain a singular value matrix, a dictionary assignment matrix, and a coefficient assignment matrix; According to the first column elements of the dictionary assignment matrix, the column vector of the initial dictionary is assigned and updated; Assign and update the row vector of the sparse coefficient matrix according to the product of the non-zero singular value of the first column of the singular value matrix and the elements of the first column in the coefficient assignment matrix; By traversing and updating multiple columns of the initial dictionary and multiple rows of the sparse coefficient matrix, an updated dictionary and sparse coefficient matrix are obtained, and the number of iterations is increased by one; When the number of iterations reaches the defined number of iterations, the dictionary obtained by the last update is used as the sparse dictionary, and the corresponding sparse coefficient matrix is ​​recorded.

5. The method according to claim 1, characterized in that According to the sparse dictionary corresponding to the centrifugal compressor, the pressure pulsation data is sparsely reconstructed to obtain a stall sensing signal, including: Using the dictionary length of the sparse dictionary, segmenting the pressure pulsation data to obtain a corresponding mean normalization matrix; An orthogonal matching pursuit algorithm is used to solve the mean normalization matrix to obtain a sparse coefficient matrix of the pressure pulsation data under the sparse dictionary; The sparse coefficient matrix is ​​multiplied by the pressure pulsation data to generate the stall sensing signal after sparse reconstruction.

6. The method according to claim 1, characterized in that The phase parameters include a target phase shift; According to the stall sensing signal, the phase parameters of the stall dynamic propagation are identified through dual-channel fusion and spatial phase identification methods, including: Calculating the mean and standard deviation of the stall sensing signal to determine a pressure pulsation signal after removing the mean; Calculating a phase shift corresponding to a dual-channel signal and a correlation coefficient corresponding to the phase shift, wherein the dual-channel signal is the pressure pulsation signal of two different pressure sensors; Dynamically normalizing the correlation coefficient by using the standard deviation value of the dual-channel signal to obtain a phase characterization function, wherein the phase characterization function is used to characterize the functional relationship between the phase shift and the zero-mean normalized cross-correlation value; According to the phase characterization function, by solving the local extreme value, the target phase shift of the dual-channel signal when the phase characterization function value is maximum is obtained.

7. The method according to claim 6, characterized in that The multiple target parameters include the number of stall clusters, the circumferential propagation speed of the stall clusters, and the propagation direction of the stall clusters relative to the impeller; Calculating a plurality of target parameters of the centrifugal compressor according to the phase parameter includes: Determine the relationship between the number of stall clusters, the circumferential propagation velocity, the impeller rotation frequency of the centrifugal compressor, and the non-integral order of the rotating stall; Obtaining the phase characterization function, the circumferential spatial angle of the two pressure sensors corresponding to the dual-channel signals, and the impeller rotation frequency, and calculating the circumferential propagation speed of the stall group; Obtaining the non-integral order of the rotating stall, the circumferential propagation velocity and the impeller rotation frequency, and calculating the number of the stall clusters according to the relationship; The propagation direction of the stall group is determined according to the positive or negative value of the target phase shift corresponding to the maximum value of the phase characterization function corresponding to the dual-channel signal.

8. The method according to claim 7, characterized in that Determine the number of stall clusters, the relationship between the circumferential propagation velocity and the impeller rotation frequency of the centrifugal compressor and the non-integral order of the rotating stall, including: Acquire a stall sensing signal, perform spectrum analysis on the stall sensing signal, and obtain a stall characteristic frequency, wherein the stall characteristic frequency is the product of the number of stall clusters and the circumferential propagation speed of the stall clusters; The stall characteristic frequency is normalized by using the impeller rotation frequency of the centrifugal compressor, and the stall characteristic frequency is associated with the impeller rotation frequency to obtain the relationship, which is as follows: ； In the formula, is the rotating stall non-integral order, is the number of stall groups, is the impeller rotation frequency, is the circumferential propagation velocity, subscript The parameter is based on the identification result obtained by measuring the signal in the stationary / absolute coordinate system, and defines the stall propagation direction as the same as the impeller rotation direction. Is positive.

9. An electronic device, comprising: A processor and a memory storing a program, wherein the program comprises instructions, which, when executed by the processor, cause the processor to perform the method according to any one of claims 1 to 8.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images