CN-121682384-B - Method and system for evaluating efficiency of gas compressor based on fading trend prejudgment

CN121682384BCN 121682384 BCN121682384 BCN 121682384BCN-121682384-B

Abstract

The invention belongs to the technical field of compressor efficiency evaluation, and provides a method and a system for evaluating the compressor efficiency based on fading trend prejudgment, wherein the method comprises the steps of comprehensively judging whether efficiency signals vibrate at high frequency due to rotating stall groups or not through time domain analysis judgment of high-frequency periodic vibration and frequency domain judgment of characteristic peaks; if the method comprises the steps of carrying out double-frequency band division of stall group oscillation characteristic frequency bands and efficiency fading low-frequency characteristic frequency bands, collecting compressor efficiency oscillation and fading composite signals, carrying out wavelet packet decomposition according to the divided double frequency bands, removing the decomposed oscillation characteristic frequency band signals, only preserving the efficiency fading characteristic frequency band signals, reconstructing a pure efficiency fading trend curve without oscillation interference through spline interpolation and sliding average, carrying out fitting calculation analysis of linear/nonlinear fading rate based on the pure efficiency fading trend curve, and combining the morphological analysis of the pure efficiency fading trend curve to realize the judgment of the efficiency fading trend of the compressor.

Inventors

SUN BING
ZUO KEJUN
Bai Naijian
WEI LEI
LIU YONG
JIANG ZHONGFENG

Assignees

太仓点石航空动力有限公司

Dates

Publication Date: 20260505
Application Date: 20260210

Claims (8)

1. The method for evaluating the efficiency of the gas compressor based on the prejudgment of the declining trend is characterized by comprising the following steps of: step 10, collecting real-time monitoring data of inlet and outlet pressure, flow and efficiency of the compressor in a near stall state, and comprehensively judging whether efficiency signals vibrate at high frequency due to a rotating stall group through time domain analysis judgment of high-frequency periodic vibration and frequency domain analysis judgment of characteristic peaks; Step S20, if yes, calculating stall group oscillation characteristic frequency bands through real-time rotor rotation speed and rotating stall group propagation coefficients of the compressor, and calibrating efficiency decline low-frequency characteristic frequency bands to realize double-frequency band division; s30, collecting a compressor efficiency oscillation and decay composite signal, and executing wavelet packet decomposition according to divided double frequency bands, wherein the wavelet packet decomposition comprises narrow window width decomposition of an oscillation frequency band and wide window width decomposition of a decay frequency band; S40, eliminating the decomposed oscillation characteristic frequency band signals, only reserving the efficiency degradation characteristic frequency band signals, and reconstructing a pure efficiency degradation trend curve without oscillation interference through spline interpolation and sliding average; the process of reconstructing the pure efficiency decay trend curve without oscillation interference is as follows: Performing wavelet packet inverse transformation on the wavelet packet coefficients of the reserved efficiency fading low-frequency characteristic frequency band, and restoring the wavelet packet coefficients into a time domain signal to obtain a fading signal sequence; Processing the signal gap by adopting a cubic spline interpolation algorithm, performing moving average processing on the fading signal sequence after interpolation, sequencing the fading signal sequence after smoothing according to time stamps, and constructing a pure efficiency fading trend curve with the horizontal axis as running time and the vertical axis as an efficiency value; s50, based on a pure efficiency decay trend curve, fitting calculation analysis of linear/nonlinear decay rate is carried out, and the compressor efficiency decay trend judgment is realized by combining with the morphological analysis of the pure efficiency decay trend curve; the process for judging the efficiency decline trend of the air compressor is as follows: based on a pure efficiency decline trend curve, constructing a standardized data set by using running time and efficiency values, respectively adopting a least square method to perform linear, exponential and power function model fitting, and calculating the fitting goodness of each model; Based on the fitting result, extracting a linear decay rate and converting the linear decay rate into a thousand-hour decay rate, or deriving a nonlinear model to analyze instantaneous rate change; And carrying out auxiliary judgment by combining the overall morphological characteristics of the pure efficiency decay trend curve, and comprehensively judging that the efficiency decay belongs to a specific class in linear slow decay, nonlinear acceleration decay or nonlinear stable decay through fitting goodness comparison, decay rate quantification calculation and curve morphological analysis.
2. The method for evaluating efficiency of a compressor based on declining trend prediction as set forth in claim 1, wherein the process of collecting real-time monitoring data of intake and exhaust pressure, flow and efficiency of the compressor in a near stall state comprises the steps of: Under the near stall working condition, when the anti-surge system is not triggered, triggering data acquisition, wherein the data format comprises a time stamp, air inlet pressure, air outlet pressure, air inlet flow, air inlet temperature and air outlet temperature; Outliers are removed by adopting a 3 sigma rule, the environmental random noise is filtered by a 5-order Butterworth low-pass filter, and then an efficiency real-time value is obtained by isentropic efficiency calculation.
3. The method for evaluating the efficiency of the compressor based on the fading trend prediction as set forth in claim 2, wherein the time domain analysis of the high-frequency periodic oscillation comprises the following steps: Dividing the acquired pressure, flow and efficiency data into data segments, calculating the fluctuation amplitude, fluctuation coefficient and oscillation period consistency of each data segment, and judging that the time domain analysis passes when the fluctuation amplitude and the fluctuation coefficient of the pressure, flow and efficiency signals exceed a preset threshold value and the oscillation period deviation rate is not more than 10%, otherwise, judging that any index does not meet the requirement and does not pass.
4. The method for evaluating the efficiency of a compressor based on the prejudgment of a declining trend according to claim 3, wherein the frequency domain analysis process of the characteristic peak is as follows: And carrying out fast Fourier transform on the segmented data, extracting a characteristic frequency interval related to the rotor rotating speed and the stall clique propagation coefficient, judging that the frequency domain analysis passes if a significant peak exists in the characteristic frequency interval and the peak amplitude ratio is more than or equal to 5, and the peak frequency deviation of continuous multi-segment data does not exceed 1Hz, otherwise, judging that the frequency domain analysis does not pass if any index is not satisfied, and only when the time domain and the frequency domain judgment pass, judging that the high-frequency oscillation of the efficiency signal is caused by the rotating stall clique.
5. The method for evaluating the efficiency of a compressor based on the prediction of a fading trend as set forth in claim 1, wherein the process of dividing the dual frequency bands is as follows: the method comprises the steps of obtaining the real-time rotor rotating speed of the air compressor, combining the rotating stall group propagation coefficient, calculating the theoretical oscillation frequency, setting an oscillation characteristic frequency band with the theoretical oscillation frequency as the center, covering the frequency fluctuation range with the width, and calibrating the low-frequency characteristic frequency band according to the slow-variation characteristic of efficiency decline.
6. The method for estimating efficiency of a compressor based on declining trend prediction as set forth in claim 5, wherein the performing of wavelet packet decomposition is: The method comprises the steps of determining a wavelet packet fundamental wave and a decomposition layer number, namely selecting a db4 wavelet packet as the fundamental wave, calculating the decomposition layer number according to the characteristics of dual frequency bands, executing wavelet packet decomposition on a composite signal, and adopting differential window width setting for the two frequency bands: The stall group oscillation characteristic frequency band adopts a narrow window width, the efficiency decline low-frequency characteristic frequency band adopts a wide window width, and after decomposition is completed, wavelet packet coefficients corresponding to the stall group oscillation characteristic frequency band and the efficiency decline low-frequency characteristic frequency band are respectively extracted to form two independent frequency band signals.
7. The method for evaluating the efficiency of the compressor based on the regression trend prediction according to claim 1, wherein the comprehensive judging efficiency regression belongs to the specific category of linear slow regression, nonlinear acceleration regression or nonlinear stationary regression, and the method comprises the following steps: Judging linear slow fading, namely judging that the goodness of fit of the linear fitting model is highest and not lower than 0.85, the calculated thousand-hour fading rate is lower than 0.1% and the fluctuation amplitude is limited, and meanwhile, the pure efficiency fading trend curve is in a straight line shape, and the slope deviation and fitting residual error of each period do not exceed a preset threshold value; Judging nonlinear acceleration fading, namely judging that the goodness of fit of the exponential fit model is highest and not lower than 0.85, calculating that the average fading rate of thousands of hours is not lower than 0.1 percent and the continuous incremental characteristic is presented, and simultaneously, the whole pure efficiency fading trend curve presents the concave morphological characteristic; and judging nonlinear stationary decay, namely judging that the fitting goodness of the power function fitting model is highest and not lower than 0.85, gradually reducing the calculated thousand-hour average decay rate to zero, and meanwhile, enabling the pure efficiency decay trend curve to be in an upward convex state and enabling the tail end section to be gradually gentle.
8. The compressor efficiency evaluation system based on the fading trend prejudgment is characterized by comprising the following modules: the high-frequency oscillation judging module is used for collecting real-time monitoring data of the inlet and outlet pressure, flow and efficiency of the compressor in a near stall state, and comprehensively judging whether the efficiency signal high-frequency oscillation is caused by a rotating stall group through time domain analysis judgment of high-frequency periodic oscillation and frequency domain analysis judgment of a characteristic peak; The dual-band dividing module is used for calculating stall group oscillation characteristic frequency bands through the real-time rotor rotating speed and the rotating stall group propagation coefficient of the compressor if the judgment is yes, calibrating the low-frequency characteristic frequency bands with declining efficiency, and realizing dual-band division; The dual-band decomposition module is used for collecting the compressor efficiency oscillation and decay composite signal, and executing wavelet packet decomposition according to the divided dual-band, wherein the dual-band decomposition module comprises the steps of adopting narrow window width decomposition for an oscillation frequency band and adopting wide window width decomposition for a decay frequency band; the efficiency decay reconstruction module is used for eliminating the decomposed oscillation characteristic frequency band signals, only preserving the efficiency decay characteristic frequency band signals, and reconstructing a pure efficiency decay trend curve without oscillation interference through spline interpolation and sliding average; the process of reconstructing the pure efficiency decay trend curve without oscillation interference is as follows: Performing wavelet packet inverse transformation on the wavelet packet coefficients of the reserved efficiency fading low-frequency characteristic frequency band, and restoring the wavelet packet coefficients into a time domain signal to obtain a fading signal sequence; Processing the signal gap by adopting a cubic spline interpolation algorithm, performing moving average processing on the fading signal sequence after interpolation, sequencing the fading signal sequence after smoothing according to time stamps, and constructing a pure efficiency fading trend curve with the horizontal axis as running time and the vertical axis as an efficiency value; The efficiency decay analysis module is used for carrying out fitting calculation analysis of linear/nonlinear decay rate based on a pure efficiency decay trend curve and realizing the judgment of the efficiency decay trend of the air compressor by combining with the morphological analysis of the pure efficiency decay trend curve; the process for judging the efficiency decline trend of the air compressor is as follows: based on a pure efficiency decline trend curve, constructing a standardized data set by using running time and efficiency values, respectively adopting a least square method to perform linear, exponential and power function model fitting, and calculating the fitting goodness of each model; Based on the fitting result, extracting a linear decay rate and converting the linear decay rate into a thousand-hour decay rate, or deriving a nonlinear model to analyze instantaneous rate change; And carrying out auxiliary judgment by combining the overall morphological characteristics of the pure efficiency decay trend curve, and comprehensively judging that the efficiency decay belongs to a specific class in linear slow decay, nonlinear acceleration decay or nonlinear stable decay through fitting goodness comparison, decay rate quantification calculation and curve morphological analysis.

Description

Method and system for evaluating efficiency of gas compressor based on fading trend prejudgment Technical Field The invention belongs to the technical field of compressor efficiency evaluation, and particularly relates to a method and a system for evaluating compressor efficiency based on fading trend prediction. Background In the field of compressor performance monitoring and efficiency evaluation, accurate judgment of efficiency decline trend in a near stall state is important to equipment safe and stable operation and maintenance decision-making. When the compressor enters a near-stall working condition, the attack angle of airflow in a flow channel is abnormal, a low-speed airflow cluster which is transmitted in the circumferential direction, namely a rotating stall cluster, is easy to form, the stall cluster can induce the air inlet and outlet pressure and flow signals to generate high-frequency periodic oscillation, and the oscillation frequency is directly related to the transmission speed of the stall cluster. Since the efficiency calculation depends on the temperature-pressure flow parameters, the oscillation signals are synchronously added into the efficiency calculation value, so that the efficiency signals are severely fluctuated. The core contradiction of the problem is that high-frequency large-amplitude transient efficiency oscillation caused by rotating stall and low-frequency small-amplitude slow variation trend inherent to efficiency decline are mutually overlapped in a signal dimension, and the traditional evaluation method is difficult to realize effective separation of the high-frequency large-amplitude transient efficiency oscillation and the low-frequency small-amplitude slow variation trend. Although the time average method can weaken oscillation interference, the real efficiency decay value can be directly smoothed, and the instantaneous value method cannot avoid the oscillation interference, so that the slow decay trend is difficult to identify from the clutter signal. The periodic oscillation signal completely covers the trend change, so that the efficiency decline trend extraction distortion is caused, and the evaluation accuracy is affected. Therefore, the invention provides a method and a system for evaluating the efficiency of a gas compressor based on the prediction of a declining trend. Disclosure of Invention In order to overcome the deficiencies of the prior art, at least one technical problem presented in the background art is solved. The technical scheme adopted by the invention for solving the technical problems is that the method for evaluating the efficiency of the air compressor based on the prejudgment of the declining trend comprises the following steps: the high-frequency oscillation judging module is used for collecting real-time monitoring data of the inlet and outlet pressure, flow and efficiency of the compressor in a near stall state, and comprehensively judging whether the efficiency signal high-frequency oscillation is caused by a rotating stall group through time domain analysis judgment of high-frequency periodic oscillation and frequency domain analysis judgment of a characteristic peak; The dual-band dividing module is used for calculating stall group oscillation characteristic frequency bands through the real-time rotor rotating speed and the rotating stall group propagation coefficient of the compressor if the judgment is yes, calibrating the low-frequency characteristic frequency bands with declining efficiency, and realizing dual-band division; The dual-band decomposition module is used for collecting the compressor efficiency oscillation and decay composite signal, and executing wavelet packet decomposition according to the divided dual-band, wherein the dual-band decomposition module comprises the steps of adopting narrow window width decomposition for an oscillation frequency band and adopting wide window width decomposition for a decay frequency band; the efficiency decay reconstruction module is used for eliminating the decomposed oscillation characteristic frequency band signals, only preserving the efficiency decay characteristic frequency band signals, and reconstructing a pure efficiency decay trend curve without oscillation interference through spline interpolation and sliding average; And the efficiency decay analysis module is used for carrying out fitting calculation analysis of linear/nonlinear decay rate based on a pure efficiency decay trend curve and realizing the judgment of the efficiency decay trend of the air compressor by combining with the morphological analysis of the pure efficiency decay trend curve. As a further scheme of the invention, the process of collecting the real-time monitoring data of the inlet and outlet pressure, flow and efficiency of the compressor in the near stall state is as follows: Under the near stall working condition, when the anti-surge system is not triggered, triggering data acquisition, wherein th