Correlation Analysis Method of Pressure Pulsation in Hydropower Station Water Flow System on Low-Frequency Oscillation of Generating Unit
By setting up measuring points at key locations in the water flow system of hydropower station units, and combining signal time-frequency domain feature extraction and Gini-gamma correlation analysis, the problem of locating and quantifying low-frequency oscillations of hydropower units was solved, enabling stable operation and optimized control of the units under low-load conditions.
Patent Information
- Application Number
- CN202611158207.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-31
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies struggle to accurately pinpoint the hydraulic drive source of low-frequency oscillations in hydropower units, fail to quantify the nonlinear correlation between pressure pulsations and power oscillations, and lack systematic analysis methods, leading to unstable unit operation under low-load conditions.
Pressure measurement points were deployed at key locations in the water flow system of the hydropower station unit using multi-channel synchronous acquisition equipment. By combining signal time-frequency domain feature extraction, Hilbert transform and Gini gamma correlation analysis, a multi-dimensional quantitative evaluation index system was constructed to accurately identify the oscillation driving source and quantify its impact.
It has enabled precise positioning of low-frequency oscillations in hydropower units and quantification of nonlinear correlations, providing reliable data support, providing a basis for stable unit operation and optimized control strategies, and improving the safety and stability of unit operation.
Smart Images

Figure CN123115776A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of auxiliary technology for hydropower systems, specifically to a method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit. Background Technology
[0002] With the large-scale construction of hydropower stations and the continuous development of new power systems, hydropower units need to frequently participate in grid peak shaving and frequency regulation tasks, operating under off-design conditions such as low load and variable load for extended periods. When the units are in the low load range, the internal flow pattern of the hydropower station's water system becomes turbulent, and hydraulic instability is significantly aggravated. This can easily induce low-frequency oscillations in the active power of the units, which not only affect the quality of power generation but also threaten the operational safety of the unit's shaft system and hydraulic equipment, thus restricting the overall operational stability of the power station.
[0003] In current analysis and troubleshooting of low-frequency oscillations in generating units, traditional techniques mostly treat the hydraulic system as a black box model or a simplified linear model, focusing on electrical circuits and mechanical structures, with less in-depth analysis of the intrinsic relationship between hydraulic flow pressure pulsations and unit power oscillations. Even existing techniques that employ single time-domain or frequency-domain analysis for signal research have significant limitations: on the one hand, it is difficult to accurately characterize the complex nonlinear coupling relationship between pressure pulsations at different locations in the water flow system and low-frequency power oscillations in the unit; on the other hand, it cannot effectively capture the transient change characteristics, phase synchronization laws, and hysteresis response characteristics of the signal.
[0004] In practical engineering applications, existing analytical methods lack a systematic analytical framework that integrates time-frequency characteristic analysis, transient mode identification, and nonlinear correlation quantitative assessment. This makes it impossible to accurately pinpoint the specific hydraulic drive components inducing low-frequency power oscillations in the unit, nor can it quantify the impact and contribution of pressure pulsations at different locations on the oscillations. Furthermore, it is difficult to clarify the synchronous and delayed response patterns of pressure pulsations and power oscillations, ultimately resulting in a lack of reliable data support and theoretical basis for unit vibration damping schemes and optimized operation strategies. Therefore, there is an urgent need to develop an analytical method that can comprehensively analyze the pressure pulsation characteristics of the water system, accurately identify the hydraulic source of oscillations, and quantify the correlation between the two, ensuring the safe and stable operation of hydropower units under low-load conditions. Summary of the Invention
[0005] The technical problem to be solved by this invention is to provide a correlation analysis method for pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit. This method solves the problems of existing technologies that make it difficult to accurately locate the source of hydraulic oscillation and quantify the nonlinear correlation between pressure pulsation and power oscillation. It accurately identifies the hydraulic driving source that induces low-frequency power oscillation of the generating unit under low load conditions, and reveals the synchronous and delayed response laws of pressure pulsation and power oscillation. This provides reliable data support for the stable operation of hydropower units, optimization of control strategies, and formulation of vibration isolation schemes.
[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: The correlation analysis method for pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit includes the following steps: a. Measurement point layout and synchronous signal acquisition: Pressure measurement points are set up at the spiral inlet, bladeless zone, downstream of tailrace, and under the top cover of the hydropower station unit's water flow system. Pressure pulsation signals at each measurement point are collected synchronously, along with unit active power, speed, unit frequency, and guide vane opening operation status signals. b. Time-frequency domain feature extraction: Perform time-domain statistical analysis on all acquired signals to extract signal probability statistical features; combine frequency domain analysis and filtering processing to obtain signal spectral energy characteristics and frequency distribution characteristics; c. Dynamic coupling characteristic analysis: Based on the Hilbert transform, the signal is demodulated to solve for the instantaneous amplitude, instantaneous frequency and instantaneous phase of each signal. The resonant coupling relationship and phase synchronization characteristics of the pressure pulsation signal and the unit active power signal in the low frequency range are analyzed. d. Nonlinear correlation quantitative calculation: The Gini gamma correlation analysis method is used to calculate the Gini gamma correlation coefficient between the pressure pulsation signal and the active power signal of the unit at each location, and to construct a multi-dimensional quantitative evaluation index system; e. Comprehensive determination of oscillation driving source: Combining the signal time-frequency characteristics, Hilbert transform analysis results and Gini gamma correlation coefficient, comprehensively determine the hydraulic driving source that induces low-frequency oscillation of unit power, and determine the degree of influence of pressure pulsation at different locations.
[0007] In step a above, a multi-channel synchronous acquisition device is used to achieve millisecond-level synchronous sampling of pressure pulsation signals and unit operating status signals. The sampling frequency matches the low-frequency oscillation frequency range of the unit, and the sampling duration covers the entire process of steady-state operation and load fluctuation of the unit.
[0008] In step b above, the time-domain statistical analysis includes extracting five types of probabilistic statistical features: signal mean, standard deviation, root mean square, root square amplitude, and kurtosis factor. Wherein, let the signal sequence be The total length of the data is The corresponding calculation method is as follows: Mean: ; Standard deviation: ; Mean squared error: ; Root amplitude: ; Kurtosis factor: .
[0009] In step b above, frequency domain analysis includes converting time-domain signals to frequency-domain signals based on Fourier transform, using low-pass and high-pass filters to separate the spectrum, distinguishing between low-frequency and high-frequency components, and extracting five types of frequency domain characteristic indicators: frequency centroid value, frequency variance, frequency standard deviation, root mean square of frequency amplitude, and average frequency value.
[0010] In step c above, the Hilbert transform demodulation process is specifically as follows: (1) Perform Hilbert transform on the given signal: ; In the formula, Represents the principal value integral; Indicates time; (2) Construct a complex signal; convert the signal into a signal with a 90° phase difference from the original signal. Thus, an imaginary part signal is obtained; the signal obtained through the Hilbert transform With the original signal It can form a complex signal The specific formula is as follows:
[0011] In the formula, the modulus of the complex signal i The instantaneous amplitude of the signal is represented by the equation. To calculate the instantaneous frequency of the signal, the argument of the complex signal can be calculated first, and then the derivative can be taken. The instantaneous amplitude and instantaneous phase are solved by the complex analytic signal, and the instantaneous frequency is obtained by differentiating the phase. Based on the above instantaneous parameters, the low-frequency resonant mode, dynamic coupling relationship and phase synchronization law between the pressure pulsation signal and the active power signal are analyzed.
[0012] In step d above, the Gini gamma correlation coefficient is calculated as follows: Let the two sets of time vectors be respectively , The vector length is , and They are respectively time series data x a and x b After sorting in ascending order, the corresponding t Time data and The grade number, express and The Gini-gamma correlation coefficient is obtained by combining the grade difference with the floor operation, the grade difference value, and the time series length. The final Gini-gamma correlation coefficient The solution is completed based on the grade difference. R g express a,b The Gini gamma correlation coefficient between two vectors; The Gini-gamma correlation coefficient is used to characterize the degree of nonlinear correlation between pressure pulsation and active power, and to distinguish three types of correlation: strong correlation, weak correlation, and negative correlation.
[0013] In step e above, the criterion for determining the oscillation driving source is as follows: a Gini gamma correlation coefficient threshold is preset. If the Gini gamma correlation coefficient between the pressure pulsation at a certain measuring point and the active power of the unit is greater than the threshold, and the two have a consistent energy peak frequency and phase synchronous change trend in the low-frequency range of the Hilbert spectrum, then the pressure pulsation at that measuring point is determined to be the main hydraulic driving source that induces the low-frequency oscillation of the unit power. The influence level of each driving source is divided based on the magnitude of the correlation coefficient.
[0014] The method also includes a full load condition analysis step: collecting signal data of the unit from 0 to 600MW full load range, performing the analysis process from step a to step e in segments according to different load ranges, summarizing the correlation between water system pressure pulsation and low-frequency oscillation of the unit under different loads, and determining the unit oscillation sensitive load range and sensitive frequency band.
[0015] The method also includes outputting unit vibration isolation operation strategies and guide vane opening optimization adjustment schemes based on the full-condition analysis results, which are used to suppress low-frequency oscillations of the unit.
[0016] The system using the above-mentioned correlation analysis method of pressure pulsation in the hydropower station's water flow system to low-frequency oscillations of the generating unit includes a signal acquisition module, a time-frequency analysis module, a coupling analysis module, a correlation calculation module, and a comprehensive judgment module. The system executes the correlation analysis method.
[0017] The correlation analysis method for pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit mentioned in this invention has the following significant advantages: 1. This invention constructs a multi-dimensional signal analysis system integrating time-domain statistical analysis, frequency-domain energy analysis, and Hilbert instantaneous feature analysis, overcoming the limitations of traditional single signal analysis methods. By extracting the instantaneous amplitude, instantaneous phase, and instantaneous frequency of the signal through Hilbert transform, it can accurately capture the transient changes in pressure pulsation and active power, effectively identify low-frequency resonant modes and dynamic coupling relationships that are difficult to detect using traditional methods, and comprehensively reconstruct the entire signal evolution process.
[0018] 2. This invention introduces the Gini gamma correlation coefficient as a quantitative evaluation index. Compared with the traditional linear correlation analysis method, it can effectively characterize the degree of nonlinear correlation between pressure pulsation and unit power fluctuation, realize the quantitative distinction of the influence intensity of pressure pulsation at different measuring points, and greatly improve the accuracy and reliability of oscillation driving source identification.
[0019] 3. This invention conducts simultaneous analysis on multiple key hydraulic locations, including the volute inlet, bladeless zone, downstream tailrace, and under the top cover, which can accurately locate the specific hydraulic components that induce low-frequency power oscillations in the unit and clarify the mechanism of pressure pulsation in each component.
[0020] 4. This invention can complete the full-process analysis of the unit's 0~600MW full load range, sort out the unit's oscillation characteristics under different loads, accurately locate the unit's oscillation-sensitive load range and sensitive frequency band, and formulate scientific vibration-avoidance operation strategies and guide vane adjustment optimization schemes to effectively suppress the unit's low-frequency oscillations.
[0021] 5. The technical solution of this invention has strong engineering practicality and is suitable for various actual operating conditions of large hydropower units, such as low load and variable load. It can provide important technical support for the stable operation of large hydropower station units and the safe regulation of new power systems, and has broad application value. Attached Figure Description
[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the overall process of the correlation analysis method of pressure pulsation in the water flow system of a hydropower station on low-frequency oscillation of the generating unit according to the present invention; Figure 2 This is a time-domain response diagram of the active power of the 27 units in this embodiment of the invention within a specified time period, showing the periodic fluctuation characteristics of the active power. Figure 3 This is a time-frequency evolution characteristic diagram of unit pressure pulsation and active power in an embodiment of the present invention; Figure 4 This is a Hilbert analysis-raw signal diagram from an embodiment of the present invention; Figure 5 This is a Hilbert analysis-analysis signal diagram from an embodiment of the present invention; Figure 6 This is a Hilbert analysis-instantaneous amplitude (envelope) diagram in an embodiment of the present invention; Figure 7 This is a Hilbert analysis-instantaneous phase diagram in an embodiment of the present invention; Figure 8 This is a Hilbert analysis-instantaneous frequency diagram in an embodiment of the present invention; Figure 9This is a frequency domain spectral array diagram characterizing the low-frequency energy distribution pattern in an embodiment of the present invention; Figure 10 This is a Gini gamma correlation heatmap used in an embodiment of the present invention to quantify the coupling strength between various state variables; Figure 11 This is a bar chart that visually demonstrates the correlation ranking of the influence of each monitoring parameter on power oscillation in this embodiment of the invention. Detailed Implementation
[0023] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0024] The correlation analysis method for pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit includes the following steps: a. Measurement point layout and synchronous signal acquisition: Pressure measurement points are set up at the spiral inlet, bladeless zone, downstream of tailrace, and under the top cover of the hydropower station unit's water flow system. Pressure pulsation signals at each measurement point are collected synchronously, along with unit active power, speed, unit frequency, and guide vane opening operation status signals. b. Time-frequency domain feature extraction: Perform time-domain statistical analysis on all acquired signals to extract signal probability statistical features; combine frequency domain analysis and filtering processing to obtain signal spectral energy characteristics and frequency distribution characteristics; c. Dynamic coupling characteristic analysis: Based on the Hilbert transform, the signal is demodulated to solve for the instantaneous amplitude, instantaneous frequency and instantaneous phase of each signal. The resonant coupling relationship and phase synchronization characteristics of the pressure pulsation signal and the unit active power signal in the low frequency range are analyzed. d. Nonlinear correlation quantitative calculation: The Gini gamma correlation analysis method is used to calculate the Gini gamma correlation coefficient between the pressure pulsation signal and the active power signal of the unit at each location, and to construct a multi-dimensional quantitative evaluation index system; e. Comprehensive determination of oscillation driving source: Combining the signal time-frequency characteristics, Hilbert transform analysis results and Gini gamma correlation coefficient, comprehensively determine the hydraulic driving source that induces low-frequency oscillation of unit power, and determine the degree of influence of pressure pulsation at different locations.
[0025] In step a above, a multi-channel synchronous acquisition device is used to achieve millisecond-level synchronous sampling of pressure pulsation signals and unit operating status signals. The sampling frequency matches the low-frequency oscillation frequency range of the unit, and the sampling duration covers the entire process of steady-state operation and load fluctuation of the unit.
[0026] In step b above, the time-domain statistical analysis includes extracting five types of probabilistic statistical features: signal mean, standard deviation, root mean square, root square amplitude, and kurtosis factor. Wherein, let the signal sequence be The total length of the data is The corresponding calculation method is as follows: Mean: ; Standard deviation: ; Mean squared error: ; Root amplitude: ; Kurtosis factor: .
[0027] In step b above, frequency domain analysis includes converting the time-domain signal to the frequency-domain signal based on Fourier transform, using low-pass and high-pass filters to separate the spectrum, distinguishing between low-frequency and high-frequency components; and extracting five types of frequency domain feature indicators: frequency centroid value, frequency variance, frequency standard deviation, root mean square frequency amplitude, and average frequency value. Among these, low-frequency components reflect the overall signal structure and large-scale fluctuation information, while high-frequency components highlight the signal's detailed features and local fluctuation information; let... This represents the total number of spectral lines. For the first The frequency values corresponding to each spectral line are used to characterize the energy distribution and dispersion of the signal spectrum based on various frequency domain characteristic indicators.
[0028] In step c above, the Hilbert transform demodulation process is specifically as follows: (1) Perform Hilbert transform on the given signal: ; In the formula, Represents the principal value integral; Indicates time; (2) Construct a complex signal; convert the signal into a signal with a 90° phase difference from the original signal. Thus, an imaginary part signal is obtained; the signal obtained through the Hilbert transform With the original signal It can form a complex signal The specific formula is as follows: ; In the formula, the modulus of the complex signal iThe instantaneous amplitude of the signal is represented by the equation. To calculate the instantaneous frequency of the signal, the argument of the complex signal can be calculated first, and then the derivative can be taken. The instantaneous amplitude and instantaneous phase are solved by the complex analytic signal, and the instantaneous frequency is obtained by differentiating the phase. Based on the above instantaneous parameters, the low-frequency resonant mode, dynamic coupling relationship and phase synchronization law between the pressure pulsation signal and the active power signal are analyzed.
[0029] In step d above, the Gini gamma correlation coefficient is calculated as follows: Let the two sets of time vectors be respectively , The vector length is , and They are respectively time series data x a and x b After sorting in ascending order, the corresponding t Time data and The grade number, express and The Gini-gamma correlation coefficient is obtained by combining the grade difference with the floor operation, the grade difference value, and the time series length. The final Gini-gamma correlation coefficient The solution is completed based on the grade difference. R g express a,b The Gini gamma correlation coefficient between two vectors; The Gini-gamma correlation coefficient is used to characterize the degree of nonlinear correlation between pressure pulsation and active power, and to distinguish three types of correlation: strong correlation, weak correlation, and negative correlation.
[0030] Gini-gamma correlation coefficient The calculation formula is: ; This means rounding a number down to the nearest integer.
[0031] In step e above, the criterion for determining the oscillation driving source is as follows: a Gini gamma correlation coefficient threshold is preset. If the Gini gamma correlation coefficient between the pressure pulsation at a certain measuring point and the active power of the unit is greater than the threshold, and the two have a consistent energy peak frequency and phase synchronous change trend in the low-frequency range of the Hilbert spectrum, then the pressure pulsation at that measuring point is determined to be the main hydraulic driving source that induces the low-frequency oscillation of the unit power. The influence level of each driving source is divided based on the magnitude of the correlation coefficient.
[0032] The method also includes a full load condition analysis step: collecting signal data of the unit from 0 to 600MW full load range, performing the analysis process from step a to step e in segments according to different load ranges, summarizing the correlation between water system pressure pulsation and low-frequency oscillation of the unit under different loads, and determining the unit oscillation sensitive load range and sensitive frequency band.
[0033] The method also includes outputting unit vibration isolation operation strategies and guide vane opening optimization adjustment schemes based on the full-condition analysis results, which are used to suppress low-frequency oscillations of the unit.
[0034] The system using the above-mentioned correlation analysis method of pressure pulsation in the hydropower station's water flow system to low-frequency oscillations of the generating unit includes a signal acquisition module, a time-frequency analysis module, a coupling analysis module, a correlation calculation module, and a comprehensive judgment module. The system executes the correlation analysis method.
[0035] Example 1: This embodiment takes the No. 27 turbine generator unit of a hydropower station as the analysis object. The unit's load regulation range is 0~600MW. The unit has long suffered from low-frequency oscillation of active power under low load conditions. The method of this invention is used to conduct a correlation analysis between pressure pulsation and low-frequency oscillation of the unit.
[0036] I. Measurement Point Layout and Signal Acquisition Perform the measurement point setup and signal acquisition work according to step a of this invention: Pressure sensors were installed at four key locations in the unit's water flow system: WKJ at the volute inlet, WYQ in the bladeless zone, WSS downstream of the tailrace, and DGX under the top cover. Unit operating parameter acquisition devices were also deployed simultaneously. Multi-channel synchronous acquisition equipment was used to achieve millisecond-level synchronous sampling of pressure pulsation signals, active power, speed, unit frequency, and guide vane opening signals.
[0037] The load of the test unit was gradually increased from 0MW to 600MW, fully covering the full load conditions. Analytical data was selected from measured data on October 21, 2025, with a focus on in-depth analysis of the 211.545MW low-load condition. Figure 2 The figure shows the time-domain response of Unit 27, where the active power exhibits more than 10 periodic fluctuations during the period from 15:51:35 to 15:51:55. Figure 3 – Figure 11 This is a comprehensive analysis diagram of the low-frequency oscillation of the power of Unit 27 in this embodiment of the invention; Figure 3 This is a time-frequency evolution characteristic diagram of unit pressure pulsation and active power in an embodiment of the present invention; Figures 4-8 Hilbert analysis plot to reveal the synchronous relationship between instantaneous amplitude and phase of a signal: Figure 4 This is a Hilbert analysis-raw signal diagram from an embodiment of the present invention; Figure 5This is a Hilbert analysis-analysis signal diagram from an embodiment of the present invention; Figure 6 This is a Hilbert analysis-instantaneous amplitude (envelope) diagram in an embodiment of the present invention; Figure 7 This is a Hilbert analysis-instantaneous phase diagram in an embodiment of the present invention; Figure 8 This is a Hilbert analysis-instantaneous frequency diagram in an embodiment of the present invention; Figure 9 This is a frequency domain spectral array diagram characterizing the low-frequency energy distribution pattern in an embodiment of the present invention; Figure 10 This is a Gini gamma correlation heatmap used in an embodiment of the present invention to quantify the coupling strength between various state variables; Figure 11 This is a bar chart that visually demonstrates the correlation ranking of the influence of each monitoring parameter on power oscillation in this embodiment of the invention.
[0038] II. Temporal Feature Extraction Following step b, conduct time-domain statistical analysis to extract five time-domain features of each pressure measurement point and active power signal: mean, standard deviation, root mean square, kurtosis factor, and root mean square amplitude. The feature data are shown in Table 1.
[0039] Table 1. Time-frequency domain characteristic values of unit pressure pulsation and active power
[0040] III. Frequency Domain Feature Extraction and Spectrum Analysis The time-domain signal is converted into a frequency-domain signal based on Fourier transform, and spectrum separation is completed by combining low-pass and high-pass filters. Five frequency domain features are extracted: frequency centroid, frequency variance, frequency standard deviation, root mean square of frequency amplitude, and average frequency value. The data are shown in Table 1. At the same time, the first three characteristic frequencies of each signal are extracted, as shown in Table 2.
[0041] Table 2 Characteristic frequency values of unit pressure pulsation and active power
[0042] Frequency domain analysis reveals that the first three dominant frequencies of the pressure pulsation in the bladeless zone (1.13Hz, 1.33Hz, and 2.27Hz) are completely consistent with the dominant frequency of the active power, indicating a strong correlation between pressure pulsation and power oscillation in the bladeless zone. The characteristic frequencies at the volute inlet and under the top cover partially overlap with the active power, while the characteristic frequencies downstream of the tailrace show significant differences. Combined with the frequency domain spectral array, it is evident that each pressure pulsation signal exhibits energy concentration in the low-frequency range of 0.5–2Hz, coupling with the low-frequency oscillation band of the unit's power.
[0043] IV. Analysis of Coupling Characteristics of Hilbert Transform Perform Hilbert transform on all signals according to step c, construct complex analytic signals, solve for instantaneous amplitude, instantaneous phase, and instantaneous frequency, and plot the Hilbert analysis spectrum.
[0044] Analysis results show that: the unit's active power exhibits a significant energy peak in the 1-2Hz range, which is the core oscillation frequency; the energy peak and instantaneous phase change trend of pressure pulsations at the volute inlet and bladeless zone are highly synchronized with the active power; the energy peak of pressure pulsations downstream of the tailrace is weaker in the same frequency band; the pressure pulsations under the top cover have a dispersed spectrum, exhibiting multi-frequency resonance. These results demonstrate that the pressure pulsations in the water flow system and the unit's active power exhibit significant resonant coupling and phase synchronization characteristics in the low-frequency range.
[0045] The frequency domain spectral array shows a clear concentration of low-frequency energy in the pressure pulsations, concentrated between approximately 0.5 and 2 Hz. This indicates that the pressure pulsations have high energy in the low-frequency range and may be coupled with system power oscillations. The WKJ pressure pulsation spectrum is generally reddish, especially with sustained high energy in the 1–3 Hz range, indicating a sustained positive correlation between the pressure pulsation energy and active power changes. The WYQ pressure pulsation spectrum is smoother, with stronger energy in the low-frequency portion (<1 Hz), suggesting a long-term, gradual relationship between the pressure pulsations and active power oscillations, representing a response to overall system low-frequency fluctuations.
[0046] V. Quantitative Calculation of Gini Gamma Correlation Calculate the Gini gamma correlation coefficient between pressure pulsation, guide vane opening and active power of the unit at each measuring point according to step d, and conduct analysis in conjunction with correlation heatmap and correlation ranking bar chart.
[0047] Experimental data show that the Gini gamma correlation coefficient between pressure pulsation in the bladeless zone, guide vane opening, and active power is greater than 0.6, indicating a strong positive correlation; the pressure pulsation at the volute inlet and under the top cover shows a moderate correlation; and the absolute value of the Gini gamma correlation coefficient between pressure pulsation downstream of the tailrace, unit frequency, and active power is less than 0.2, indicating an extremely weak correlation.
[0048] VI. Comprehensive Determination of Oscillation Driving Source Based on the judgment criteria in step e, and combining time-frequency characteristics, Hilbert spectrum, and Gini gamma correlation coefficient, a comprehensive judgment is made: The pressure pulsation in the bladeless zone WYQ not only perfectly matches the main frequency and active power, but also has a highly synchronized low-frequency phase, and the Gini gamma correlation coefficient exceeds the set threshold. Therefore, the pressure pulsation in the bladeless zone is determined to be the main hydraulic driving source that induces the low-frequency power oscillation under low load conditions of the unit. The pressure pulsation at the volute inlet and under the top cover is a secondary influencing factor. The pressure pulsation downstream of the tailrace has basically no impact on this low-frequency oscillation.
[0049] VII. Full-condition analysis and operation strategy formulation A full-load segment analysis of the unit from 0 to 600MW was conducted to summarize the evolution patterns of pressure pulsation and low-frequency oscillations under different loads. The low-load range of the unit was identified as the oscillation-sensitive range, with 1-2Hz as the core sensitive frequency band. Based on the analysis results, a vibration-damping operation plan and a guide vane opening optimization adjustment strategy were formulated to avoid the sensitive load range during actual operation. Fine-tuning of the guide vane opening effectively suppressed the low-frequency power oscillations of the unit.
[0050] This embodiment uses the method of the present invention to accurately locate the hydraulic drive source of low-frequency power oscillation in Unit 27 of a hydropower station under low-load conditions, quantifies the influence of pressure pulsation in various parts, and clarifies the coupling law between pressure pulsation and power oscillation. After optimizing the unit operation mode based on the analysis results, the low-frequency oscillation phenomenon of the unit is significantly suppressed, and the unit's operational stability and power quality are significantly improved, verifying the scientific nature, accuracy, and engineering applicability of the method of the present invention.
[0051] The above description is merely a preferred 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 within the scope of protection of the present invention. Structures, connections, and method steps not described in detail in the present invention can be understood as conventional structures, connections, and method steps in the prior art.
Claims
1. A method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit, characterized in that, Includes the following steps: a. Measurement point layout and synchronous signal acquisition: Pressure measurement points are set up at the spiral inlet, bladeless zone, downstream of tailrace, and under the top cover of the hydropower station unit's water flow system. Pressure pulsation signals at each measurement point are collected synchronously, along with unit active power, speed, unit frequency, and guide vane opening operation status signals. b. Time-frequency domain feature extraction of signals: Perform time-domain statistical analysis on all acquired signals to extract the probabilistic statistical features of the signals; By combining frequency domain analysis and filtering, the signal's spectral energy characteristics and frequency distribution features can be obtained. c. Dynamic coupling characteristic analysis: Based on the Hilbert transform, the signal is demodulated to solve for the instantaneous amplitude, instantaneous frequency and instantaneous phase of each signal. The resonant coupling relationship and phase synchronization characteristics of the pressure pulsation signal and the unit active power signal in the low frequency range are analyzed. d. Nonlinear correlation quantitative calculation: The Gini gamma correlation analysis method is used to calculate the Gini gamma correlation coefficient between the pressure pulsation signal and the active power signal of the unit at each location, and to construct a multi-dimensional quantitative evaluation index system; e. Comprehensive determination of oscillation driving source: Combining the signal time-frequency characteristics, Hilbert transform analysis results and Gini gamma correlation coefficient, comprehensively determine the hydraulic driving source that induces low-frequency oscillation of unit power, and determine the degree of influence of pressure pulsation at different locations.
2. The correlation analysis method for pressure pulsation in the water flow system of a hydropower station on low-frequency oscillations of the generating unit according to claim 1, characterized in that, In step a, a multi-channel synchronous acquisition device is used to achieve millisecond-level synchronous sampling of pressure pulsation signals and unit operating status signals. The sampling frequency matches the low-frequency oscillation frequency range of the unit, and the sampling duration covers the entire process of steady-state operation and load fluctuation of the unit.
3. The correlation analysis method for pressure pulsation in the water flow system of a hydropower station on low-frequency oscillations of the generating unit according to claim 1, characterized in that, In step b, the time-domain statistical analysis includes extracting five types of probabilistic statistical features: signal mean, standard deviation, root mean square, root square amplitude, and kurtosis factor. Wherein, let the signal sequence be The total length of the data is The corresponding calculation method is as follows: Mean: ; Standard deviation: ; Mean squared error: ; Root amplitude: ; Kurtosis factor: 。 4. The correlation analysis method for pressure pulsation in the water flow system of a hydropower station on low-frequency oscillations of the generating unit according to claim 1, characterized in that, In step b, the frequency domain analysis includes converting the time domain signal to the frequency domain signal based on Fourier transform, using low-pass and high-pass filters to separate the spectrum, and extracting five types of frequency domain feature indicators: frequency centroid value, frequency variance, frequency standard deviation, root mean square frequency amplitude, and average frequency value.
5. The method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit according to claim 1, characterized in that, In step c, the Hilbert transform demodulation process specifically involves: (1) Perform Hilbert transform on the given signal: ; In the formula, Represents the principal value integral; Indicates time; (2) Construct a complex signal; convert the signal into a signal with a 90° phase difference from the original signal. Thus, an imaginary part signal is obtained; the signal obtained through the Hilbert transform With the original signal It can form a complex signal The specific formula is as follows: ; In the formula, the modulus of the complex signal i It represents the instantaneous amplitude of the signal; to calculate the instantaneous frequency of the signal, you can first calculate the argument of the complex signal and then take the derivative.
6. The method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit according to claim 1, characterized in that, In step d, the Gini-gamma correlation coefficient is calculated as follows: Let the two sets of time vectors be respectively , The vector length is , and They are respectively time series data x a and x b After sorting in ascending order, the corresponding t Time data and The grade number, express and The difference in grades between them, and the final Gini gamma correlation coefficient The solution is obtained based on the grade difference. R g express a,b The Gini gamma correlation coefficient between two vectors; The Gini-gamma correlation coefficient is used to characterize the degree of nonlinear correlation between pressure pulsation and active power, and to distinguish three types of correlation: strong correlation, weak correlation, and negative correlation.
7. The correlation analysis method for pressure pulsation in the water flow system of a hydropower station on low-frequency oscillations of the generating unit according to claim 1, characterized in that, In step e, the criterion for determining the oscillation driving source is as follows: a Gini gamma correlation coefficient threshold is preset. If the Gini gamma correlation coefficient between the pressure pulsation at a certain measuring point and the active power of the unit is greater than the threshold, and the two have consistent energy peak frequency and phase synchronous change trend in the low-frequency range of the Hilbert spectrum, then the pressure pulsation at that measuring point is determined to be the main hydraulic driving source that induces low-frequency oscillation of the unit power. The influence level of each driving source is divided based on the magnitude of the correlation coefficient.
8. The method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit according to claim 1, characterized in that, The method also includes a full load condition analysis step: collecting signal data of the unit from 0 to 600MW full load range, performing the analysis process from step a to step e in segments according to different load ranges, summarizing the correlation between water system pressure pulsation and low-frequency oscillation of the unit under different loads, and determining the unit oscillation sensitive load range and sensitive frequency band.
9. The method for correlation analysis of pressure pulsation in the water flow system of a hydropower station and low-frequency oscillation of the generating unit according to claim 8, characterized in that, The method also includes outputting unit vibration isolation operation strategies and guide vane opening optimization adjustment schemes based on the full-condition analysis results, which are used to suppress low-frequency oscillations of the unit.
10. A system using the correlation analysis method for pressure pulsation in the water flow system of a hydropower station and low-frequency oscillations of the generating unit as described in any one of claims 1-9, characterized in that, It includes a signal acquisition module, a time-frequency analysis module, a coupling analysis module, a correlation calculation module, and a comprehensive judgment module. The system executes the correlation analysis method.
