Multi-source real-time judgment method for surge boundary of fuel cell air compressor

Abstract

The invention discloses a multi-source real-time judgment method for a surge boundary of a fuel cell air compressor, and belongs to the field of fluid compression mechanical systems. According to the method, after signals collected by a sensor are filtered, the surge variance used for representing the surge fluctuation degree is introduced. And determining the moving time window length according to the sampling frequency set by the filter. And the variances of the pressure and the current signal of the air compressor are calculated in the moving time window. The temperature rise value in the time window is obtained by making the difference between the temperature at the beginning of the time window and the temperature at the end of the time window, and real-time collection and calculation of the pressure signal surge variance, the current signal surge variance and the temperature rise value of the air compressor are achieved. According to the surge recognition method based on redundant multi-source information combination, when at least two of the three parameters of the pressure surge variance, the current signal surge variance and the temperature rising value of the air compressor exceed the corresponding threshold values, it is judged that surge occurs, and the reliability and real-time performance of online recognition of the surge of the air compressor are improved.

Description

Technical Field

[0001] The invention belongs to the field of fluid compression mechanical systems and relates to a multi-source real-time judgment method for surge boundaries of a fuel cell air compressor. Background Art

[0002] The fuel cell air compressor plays a vital role in the fuel cell system. It is mainly responsible for compressing the air entering the fuel cell stack, providing sufficient pressure and flow for the cathode of the fuel cell stack, and ensuring that the chemical reaction of the fuel cell can proceed efficiently. Its performance directly affects the performance and life of the entire fuel cell system. Among them, the centrifugal air compressor has the advantages of light weight, simple structure, high pressure ratio, and easy maintenance. It is currently the most widely used form of fuel cell air compressor. However, centrifugal air compressors also have some obvious disadvantages, such as surge problems. The existence of surge lines puts higher requirements on the optimization control of air compressors.

[0003] When the air flow of a centrifugal air compressor decreases to a certain value at a certain speed, the operation becomes unstable, the airflow pulsates strongly, causing the entire air compressor to vibrate violently, and even cause damage, while emitting sharp periodic noise. This unstable working condition is called surge. The surge phenomenon is extremely harmful to the air compressor and even the entire fuel cell system. In severe cases, it may cause damage to the air compressor and even threaten the safety of personnel. Therefore, the surge boundary must be accurately defined during the performance test.

[0004] When testing the performance of a fuel cell air compressor, it is necessary to detect surge in real time according to the working conditions and define the surge boundary. However, most of the current performance tests are still performed manually, and the judgment of surge also relies on the tester's own experience, which results in large errors and uncertainties. Existing recognition algorithms also have the problem of not being able to balance speed and accuracy. When encountering sensor failures, changes in environmental factors, or other non-surge failures in the air compressor, it is very likely that the pre-designed detection algorithm will fail, misjudging the occurrence of surge and sacrificing a certain degree of accuracy. In order to improve accuracy, more complex algorithms are designed to ensure that surge identification is not inaccurate, but complex algorithms bring a greater amount of calculation, and the speed of detection cannot be guaranteed. Therefore, in order to balance speed and accuracy, it is necessary to increase the information dimension of the detection, which will bring about information redundancy, which can be used to eliminate uncertainty. Summary of the invention

[0005] In view of the above-mentioned problems, the purpose of the present invention is to provide a multi-source real-time judgment method for the surge boundary of a fuel cell air compressor, which can realize online monitoring of the surge of the fuel cell air compressor, improve the real-time and accuracy of the air compressor surge identification, and improve the accuracy of the air compressor surge boundary test.

[0006] The objective of the present invention is achieved through the following technical solutions:

[0007] The present invention discloses a multi-source real-time judgment method for surge boundaries of a fuel cell air compressor, comprising the following steps:

[0008] S1: Perform surge test on fuel cell air compressor measurement and control platform, and collect compressor inlet pressure, inlet temperature and motor current. The sensor is arranged on the double-twist flowmeter at the air compressor inlet, where the tail probe of the temperature sensor is placed in the center of the flowmeter, facing the air compressor inlet, and measuring the air compressor inlet temperature, pressure and current signals respectively. During the operation of the fuel cell system, by changing the air compressor flow and pressure ratio, gradually approaching the surge boundary of the air compressor, and dynamically collecting the air compressor inlet temperature, pressure and current signals, the acquisition of surge signals is realized.

[0009] S2: Due to the influence of instrument measurement error, environmental factors, electromagnetic interference and signal transmission process during the test, high-frequency noise is generated, which affects the accuracy of signal analysis and processing. The original signals of current and inlet pressure are processed by first-order low-pass filtering, and the inlet temperature signal is not processed. The current and inlet pressure signals after first-order low-pass filtering are obtained to reduce the influence of high-frequency noise.

[0010] S2.1: According to the compressor surge signal measured in step 1, a complete surge waveform data is intercepted, and a complete surge waveform data is defined as consisting of m sampling points, which constitutes an m-dimensional signal vector

[0011]

[0012] Among them, s(0), s(1), ..., s(m-1) are the values ​​of m sampling points contained in a complete surge signal waveform.

[0013] S2.2 uses a first-order low-pass filter as shown in the following formula to perform first-order low-pass filtering on the original signals of the current and the inlet pressure to obtain the current signal and the inlet pressure signal after the first-order low-pass filtering.

[0014] y n =ax n +(1-a)y n-1

[0015] in, y n is the output value of this filtering, y n-1 is the last filtering output value, x n is the sampling value of this time, T s is the sampling period in seconds, f c is the cutoff frequency in Hertz, and the range is [0,1].

[0016] S3: Construct a surge variance to characterize the degree of surge fluctuation. Calculate the surge variance for the compressor pressure and current signals in each time window, so that the calculation of the surge variance is more localized and the accuracy of surge recognition is improved. Determine the moving time window length ΔX according to the sampling frequency set by the filter. Calculate the variance of the compressor pressure and current signals in the moving time window, and obtain the compressor pressure signal surge variance S P And the current signal surge variance S I , that is, the fluctuation characteristics of the air compressor pressure and current signals are obtained. The temperature rise value ΔT within the time window is obtained by subtracting the temperature at the beginning of the time window from the temperature at the end of the time window.

[0017] S3.1: When surge occurs, the compressor inlet pressure and current will produce large surge fluctuations. The surge variance is introduced to characterize the degree of surge fluctuation. The surge variance calculation formula is:

[0018]

[0019] Among them, c i is the sampling value at time i, m is the number of sampling points in the preset time period, is the mean value of the sampled data during this period.

[0020] Select the time window. The statistical range m of each variance is determined by the length of the time window. The window is a signal range used for surge variance calculation. That is, the surge variance is calculated for the air compressor pressure and current signals in each time window, making the calculation of surge variance more localized and improving the accuracy of surge recognition.

[0021] S3.2: Determine the moving time window length ΔX according to the sampling frequency set by the filter

[0022]

[0023] Among them, f is the signal sampling frequency, and m is the number of sampling points contained in a complete surge waveform data.

[0024] S3.3: According to 3.1, the variance of the air compressor pressure and current signals is calculated in the time window respectively, and the surge variance of the air compressor pressure signal and the surge variance of the current signal are obtained respectively, that is, the fluctuation characteristics of the air compressor pressure and current signals are obtained.

[0025] S3.4: The temperature signal will not fluctuate significantly when surge occurs, but there will be an obvious temperature rise phenomenon. Therefore, the temperature rise value within the time window is calculated, where ΔT = T2-T1, T1 represents the temperature at the beginning of the time window, and T2 represents the temperature at the end of the time window.

[0026] S4: According to step S3, the pressure surge variance and current signal surge variance of the air compressor are measured respectively when surge occurs at different speeds, and the speed-threshold relationship of the pressure signal and the current signal is obtained by polynomial fitting. At different speeds, the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained according to the speed-threshold relationship. p and the current signal variance threshold M I By setting a fixed temperature threshold M T , used to determine whether surge occurs.

[0027] S4.1: The selection of the threshold is related to the speed. The higher the speed, the higher the corresponding threshold should be. In the early stage of performance testing, the pressure variance and current variance when surge occurs at different speeds are measured respectively. There is a unique polynomial L n (x) = b0 + b1x + b2x 2 +…+b n x n , so that L n (x j )=y j (j=1,2,…,n)

[0028] Among them, x j Indicates different speeds, y j Indicates the variance value at different speeds.

[0029] The speed-threshold relationship of the pressure signal and the current signal is obtained, and the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained at different speeds. p and the current signal variance threshold M I .

[0030] S4.2: When surge occurs, the temperature change is negligible due to the system vibration. By setting a fixed temperature threshold M T , used to determine whether surge occurs.

[0031] S5: collect the surge signal described in step S1 in real time through the sensor, determine the length of the time window using step 2, and obtain the air compressor pressure surge variance, current signal surge variance and temperature rise value within the time window according to the calculation in step S3; compare the air compressor pressure surge variance within the time window calculated in step S3 with the pressure signal variance threshold M p Compare the current signal surge variance in the time window calculated in step S3 with the current signal variance threshold M I Compare the temperature rise value in the time window calculated in step S3 with the temperature threshold M TCompare; when at least two of the three parameters of air compressor pressure surge variance, current signal surge variance and temperature rise value exceed the corresponding threshold value, it is judged that surge occurs, otherwise, surge does not occur, that is, multi-source real-time judgment of fuel cell air compressor surge boundary is realized.

[0032] Beneficial effects:

[0033] 1. The present invention discloses a multi-source real-time judgment method for surge boundary of fuel cell air compressor. The collected air compressor inlet temperature, pressure and current signals are analyzed in the time domain. It is found that when surge occurs at various speeds of the electric air compressor, the pressure and current will produce periodic pulsating signals, and the temperature will rise at the same time. It is concluded that the signal characteristics of pressure and current during surge are low-frequency and large-amplitude pulsating signals. After filtering the signals collected by the sensor, the surge variance is introduced to characterize the degree of surge fluctuation. The surge variance is calculated for the air compressor pressure and current signals in each time window, so that the calculation of the surge variance is more localized, and the recognition accuracy of surge is improved.

[0034] 2. The present invention discloses a multi-source real-time judgment method for the surge boundary of a fuel cell air compressor. Based on the beneficial effect 1, the moving time window length ΔX is determined according to the sampling frequency set by the filter. The air compressor pressure and current signal variances are calculated in the moving time window, respectively, and the air compressor pressure signal surge variance and the current signal surge variance are obtained, respectively, that is, the fluctuation characteristics of the air compressor pressure and current signals are obtained. By subtracting the temperature at the beginning of the time window from the temperature at the end of the time window, the temperature rise value in the time window is obtained, which can realize the real-time collection and solution of the air compressor pressure signal surge variance, the current signal surge variance and the temperature rise value in the time window.

[0035] 3. A multi-source real-time judgment method for surge boundary of a fuel cell air compressor disclosed in the present invention, based on the beneficial effect 2, obtains the speed-threshold relationship of the pressure signal and the current signal through polynomial fitting, and obtains the corresponding pressure signal variance threshold M for judging whether surge occurs according to the speed-threshold relationship at different speeds. p and the current signal variance threshold M I , and set a fixed temperature threshold M T . By using the redundant information generated by different sensors measuring the same feature, a surge identification method combining redundant multi-source information is constructed. When at least two of the three parameters of air compressor pressure surge variance, current signal surge variance and temperature rise value exceed the corresponding threshold, surge is judged to occur, thereby improving the reliability and real-time performance of online identification of air compressor surge. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1The present invention discloses a method for real-time multi-source judgment of surge boundaries of a fuel cell air compressor.

[0037] Figure 2 It is the air compressor inlet pressure signal at different speeds.

[0038] Figure 3 It is the air compressor inlet current signal at different speeds.

[0039] Figure 4 It is the air compressor inlet temperature signal at different speeds.

[0040] Figure 5 is the pressure signal and pressure signal variance when surge occurs.

[0041] Figure 6 is the current signal and current signal variance when surge occurs.

[0042] Figure 7 The sensor signals and alarm information at 50,000 rpm. DETAILED DESCRIPTION

[0043] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0044] Embodiment 1:

[0045] like Figure 1 As shown, this embodiment discloses a multi-source real-time judgment method for surge boundary of a fuel cell air compressor, and the specific implementation steps are as follows:

[0046] S1: Perform surge test on fuel cell air compressor measurement and control platform, and collect compressor inlet pressure, inlet temperature and motor current. The sensor is arranged on the double-twist flowmeter at the air compressor inlet, where the tail probe of the temperature sensor is placed in the center of the flowmeter, facing the air compressor inlet, and measuring the air compressor inlet temperature, pressure and current signals respectively. During the operation of the fuel cell system, by changing the air compressor flow and pressure ratio, gradually approaching the surge boundary of the air compressor, and dynamically collecting the air compressor inlet temperature, pressure and current signals, the sampling frequency is f = 100Hz, to achieve the acquisition of surge signals.

[0047] S2: Due to the influence of instrument measurement error, environmental factors, electromagnetic interference and signal transmission process during the test, high-frequency noise is generated, which affects the accuracy of signal analysis and processing. The original signals of inlet pressure and current are processed by first-order low-pass filtering, and the inlet temperature signal is not processed. The current and inlet pressure signals after first-order low-pass filtering are obtained. The inlet pressure signal is as follows: Figure 2 As shown, the import current signal is Figure 3 As shown, the inlet temperature signal is Figure 4 shown.

[0048] S2.1: According to the compressor surge signal measured in step 1, a complete surge waveform data is intercepted, and a complete surge waveform data is defined as consisting of 50 sampling points, which constitutes a 50-dimensional signal vector

[0049]

[0050] Among them, s(0), s(1), ..., s(49) are the values ​​of 50 sampling points contained in a complete surge signal waveform.

[0051] S2.2 uses a first-order low-pass filter as shown in the following formula to perform first-order low-pass filtering on the original signals of the inlet temperature and pressure to obtain the inlet temperature signal and the inlet pressure signal after the first-order low-pass filtering.

[0052] y n =ax n +(1-a)y n-1

[0053] in, y n is the output value of this filtering, y n-1 is the last filtering output value, x n is the sampling value of this time, T s is the sampling period in seconds, f c is the cutoff frequency in Hertz.

[0054] T s =0.01s, f c =275Hz, after calculation, a=0.054.

[0055] S3: Construct a surge variance to characterize the degree of surge fluctuation. Calculate the surge variance for the compressor pressure and current signals in each time window, so that the calculation of the surge variance is more localized and the accuracy of surge recognition is improved. Determine the moving time window length ΔX according to the sampling frequency set by the filter. Calculate the variance of the compressor pressure and current signals in the moving time window, and obtain the compressor pressure signal surge variance S P And the current signal surge variance S I , that is, the fluctuation characteristics of the air compressor pressure and current signals are obtained. The temperature rise value ΔT within the time window is obtained by subtracting the temperature at the beginning of the time window from the temperature at the end of the time window.

[0056] S3.1: When surge occurs, the compressor inlet pressure and current will produce large surge fluctuations. The surge variance is introduced to characterize the degree of surge fluctuation. The surge variance calculation formula is:

[0057]

[0058] Among them, c i is the sampling value at time i, m is the number of sampling points in the preset time period, which is 50, is the mean value of the sampled data during this period.

[0059] Select a suitable time window. The statistical range m of each variance is determined by the length of the time window. The window is a signal range used for surge variance calculation. That is, the surge variance is calculated for the air compressor pressure and current signals in each time window, making the calculation of surge variance more localized and improving the accuracy of surge recognition.

[0060] S3.2: Determine the moving time window length ΔX according to the sampling frequency set by the filter

[0061]

[0062] Wherein, f=100 Hz, m=50, and the calculated time window length ranges from 1.5 to 2 seconds. In this experiment, the time window length ΔX=1.5 s.

[0063] S3.3: According to 3.1, the variance of the air compressor pressure and current signals is calculated in the time window, and the air compressor pressure surge variance and the current signal surge variance are obtained, respectively, as follows: Figure 5 , 6 As shown, the fluctuation characteristics of the air compressor pressure and current signals are obtained.

[0064] S3.4: The temperature signal will not fluctuate significantly when surge occurs, but there will be an obvious temperature rise phenomenon. Therefore, the temperature rise value within the time window is calculated, where ΔT = T2-T1, T1 represents the temperature at the beginning of the time window, and T2 represents the temperature at the end of the time window.

[0065] S4: According to step S3, the pressure surge variance and current signal surge variance of the air compressor are measured respectively when surge occurs at different speeds, and the speed-threshold relationship of the pressure signal and the current signal is obtained by polynomial fitting. At different speeds, the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained according to the speed-threshold relationship. p and the current signal variance threshold M I By setting a fixed temperature threshold M T , used to determine whether surge occurs.

[0066] S4.1: The selection of the threshold is related to the speed. The higher the speed, the higher the corresponding threshold should be. In the early stage of performance testing, the pressure variance and current variance when surge occurs at different speeds are measured respectively. There is a unique polynomial L n(x) = b0 + b1x + b2x 2 +…+b n x n , so that L n (x j )=y j (j=1,2,…,n)

[0067] Among them, x j Indicates different speeds, y j Indicates the variance value at different speeds.

[0068] The function result after fitting the pressure link is:

[0069] L(x)=0.000563x 4 +0.00132x 3 +0.0024356x 2 +0.0030659x+0.002186

[0070] The function result after fitting the pressure link is:

[0071] L(x)=5×10 -7 x-0.001

[0072] The speed-threshold relationship of the pressure signal and the current signal is obtained, and the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained at different speeds. p and the current signal variance threshold M I .

[0073] S4.2: When surge occurs, the temperature change is negligible due to the system vibration. By setting a fixed temperature threshold M T , after multiple comparisons, the threshold M T =0.06.

[0074] S5: collect the surge signal described in step S1 in real time through the sensor, determine the length of the time window using step 2, and obtain the air compressor pressure surge variance, current signal surge variance and temperature rise value within the time window according to the calculation in step S3; compare the air compressor pressure surge variance within the time window calculated in step S3 with the pressure signal variance threshold M p Compare the current signal surge variance in the time window calculated in step S3 with the current signal variance threshold M I Compare the temperature rise value in the time window calculated in step S3 with the temperature threshold M TWhen at least two of the three parameters, namely, the air compressor pressure surge variance, the current signal surge variance, and the temperature rise value, exceed the corresponding threshold, it is determined that surge occurs; otherwise, surge does not occur, thus realizing multi-source real-time judgment of the fuel cell air compressor surge boundary. Figure 7 As shown in the figure, the current link has an alarm signal at 4.8s. At this time, the air compressor is in the normal adjustment process. Since the pressure and temperature links have not issued an alarm at this time, the redundant combination module has not issued an alarm signal. As the valve opening gradually closes, the pressure, current and temperature links issue alarms in turn. The redundant combination module gives an alarm signal in time after surge occurs.

[0075] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-source real-time judgment method for surge boundary of a fuel cell air compressor, characterized in that: The following steps are included: S1: A surge test is conducted on the fuel cell air compressor measurement and control platform, and the compressor inlet pressure, inlet temperature and motor current are collected; the sensor is arranged on the double-negative wire flowmeter at the air compressor inlet, where the tail probe of the temperature sensor is placed in the center of the flowmeter, facing the air compressor inlet, and the air compressor inlet temperature, pressure and current signals are measured respectively. During the operation of the fuel cell system, the air compressor flow rate and pressure ratio are changed to gradually approach the surge boundary of the air compressor, and the air compressor inlet temperature, pressure and current signals are dynamically collected to realize the collection of surge signals; S2: Due to the influence of instrument measurement error, environmental factors, electromagnetic interference and signal transmission process during the test, high-frequency noise is generated, which affects the accuracy of signal analysis and processing; the original signals of current and inlet pressure are processed by first-order low-pass filtering, and the inlet temperature signal is not processed to obtain the current and inlet pressure signals after first-order low-pass filtering to reduce the influence of high-frequency noise; S3: Construct a surge variance for characterizing the degree of surge fluctuation, calculate the surge variance for the air compressor pressure and current signals in each time window, so that the calculation of the surge variance is more localized and the recognition accuracy of surge is improved; determine the moving time window length ΔX according to the sampling frequency set by the filter; calculate the variance of the air compressor pressure and current signals in the moving time window, and obtain the air compressor pressure signal surge variance S respectively. P And the current signal surge variance S I , that is, to obtain the fluctuation characteristics of the air compressor pressure and current signals; The temperature rise value ΔT within the time window is obtained by subtracting the temperature at the beginning of the time window from the temperature at the end of the time window; S4: According to step S3, the pressure surge variance and current signal surge variance of the air compressor are measured respectively when surge occurs at different speeds, and the speed-threshold relationship of the pressure signal and the current signal is obtained by polynomial fitting. At different speeds, the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained according to the speed-threshold relationship. p and the current signal variance threshold M I ; By setting a fixed temperature threshold M T , used to determine whether surge occurs; S5: collect the surge signal described in step S1 in real time through the sensor, determine the length of the time window using step 2, and obtain the air compressor pressure surge variance, current signal surge variance and temperature rise value within the time window according to the calculation in step S3; compare the air compressor pressure surge variance within the time window calculated in step S3 with the pressure signal variance threshold M p Compare the current signal surge variance in the time window calculated in step S3 with the current signal variance threshold M I Compare the temperature rise value in the time window calculated in step S3 with the temperature threshold M T Compare; when at least two of the three parameters of air compressor pressure surge variance, current signal surge variance and temperature rise value exceed the corresponding threshold value, it is judged that surge occurs, otherwise, surge does not occur, that is, multi-source real-time judgment of fuel cell air compressor surge boundary is realized.

2. A multi-source real-time judgment method for surge boundary of a fuel cell air compressor according to claim 1, characterized in that: The implementation method of step S2 is: S2.1: According to the compressor surge signal measured in step 1, a complete surge waveform data is intercepted, and a complete surge waveform data is defined as consisting of m sampling points, which constitutes an m-dimensional signal vector Among them, s(0), s(1), ..., s(m-1) are the values ​​of m sampling points contained in a complete surge signal waveform; S2.2 uses a first-order low-pass filter as shown in the following formula to perform first-order low-pass filtering on the original signals of the current and the inlet pressure to obtain the current signal and the inlet pressure signal after the first-order low-pass filtering; y n =ax n +(1-a)y n-1 in, y n is the output value of this filtering, y n-1 is the last filtering output value, x n is the sampling value of this time, T s is the sampling period in seconds, f c is the cutoff frequency in Hertz, and the range is [0,1].

3. A multi-source real-time judgment method for surge boundary of a fuel cell air compressor as claimed in claim 2, characterized in that: The implementation method of step S3 is: S3.1: When surge occurs, the compressor inlet pressure and current will produce large surge fluctuations. The surge variance is introduced to characterize the degree of surge fluctuation. The surge variance calculation formula is: Among them, c i is the sampling value at time i, m is the number of sampling points in the preset time period, is the mean value of the sampled data during this period; Select the time window. The statistical range m of each variance is determined by the length of the time window. The window is a signal range used for surge variance calculation. That is, the surge variance is calculated for the air compressor pressure and current signals in each time window, making the calculation of surge variance more localized and improving the accuracy of surge recognition. S3.2: Determine the moving time window length ΔX according to the sampling frequency set by the filter Among them, f is the signal sampling frequency, and m is the number of sampling points contained in a complete surge waveform data; S3.3: According to 3.1, the variance of the air compressor pressure and current signals is calculated in the time window, and the surge variance of the air compressor pressure signal and the surge variance of the current signal are obtained, that is, the fluctuation characteristics of the air compressor pressure and current signals are obtained; S3.4: The temperature signal will not fluctuate significantly when surge occurs, but there will be an obvious temperature rise phenomenon. Therefore, the temperature rise value within the time window is calculated, where ΔT = T2-T1, T1 represents the temperature at the beginning of the time window, and T2 represents the temperature at the end of the time window.

4. A multi-source real-time judgment method for surge boundary of a fuel cell air compressor as claimed in claim 3, characterized in that: The implementation method of step S4 is: S4.1: The selection of the threshold is related to the speed. The higher the speed, the higher the corresponding threshold should be. In the early stage of performance testing, the pressure variance and current variance when surge occurs at different speeds are measured respectively. There is a unique polynomial L n (x) = b0 + b1x + b2x 2 +…+b n x n , so that L n (x j )=y j (j=1,2,…,n) Among them, x j Indicates different speeds, y j Indicates the variance value at different speeds; The speed-threshold relationship of the pressure signal and the current signal is obtained, and the corresponding pressure signal variance threshold M for judging whether surge occurs is obtained at different speeds. p and the current signal variance threshold M I ; S4.2: When surge occurs, the temperature change is negligible due to the influence of system vibration. By setting a fixed temperature threshold M T , used to determine whether surge occurs.

Images