A method for identifying dominant factors of dividing vibration zones of hydroelectric generating units by using e-type distortion rate

Abstract

This invention discloses a method for identifying the dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate. The method relates to the field of hydropower technology and comprises the following steps: acquiring multiple sets of state monitoring data from the hydropower unit, and performing preprocessing and normalization to obtain relative measurement values; dividing the state monitoring data into continuous load intervals, and accumulating the velocity gradients calculated for each load interval to obtain a velocity gradient accumulation sequence; calculating the shortest Euclidean distance from each state monitoring data point to the normalization boundary based on the relative measurement values ​​of each state monitoring data point; calculating the Euclidean distortion rate of each state monitoring data point based on the velocity gradient accumulation sequence and the shortest Euclidean distance of the same state monitoring data point; and selecting several state monitoring data points with relatively large Euclidean distortion rates to form a set of dominant factors for the vibration zone division of the hydropower unit, thereby enabling accurate division of the vibration zone. This method directly utilizes actual operating data of the unit, reducing human error and improving the efficiency and accuracy of dominant factor selection.

Description

Technical Field

[0001] This invention relates to the field of hydropower technology, and in particular to a method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate. Background Technology

[0002] With the large-scale grid connection of renewable energy sources such as wind and solar power into the power system, the volatility and randomness of grid power have significantly increased. This volatility stems from the intermittency and instability of renewable energy, posing a severe challenge to the power grid. Against this backdrop, the operating characteristics of hydropower, as a core peak-shaving and frequency-regulating power source, have undergone fundamental changes. Hydropower units traditionally focused on stable power generation, but now they must frequently adjust their output to adapt to grid demands, exhibiting characteristics such as high-frequency switching of operating conditions and frequent participation in peak shaving. These changes have led to increasingly complex and transient operating conditions for hydropower units, with units often operating under off-design conditions, thus altering their vibration characteristics. Traditional vibration zone delineation methods, based on historical operating data or limited testing, are no longer adequate for the new operating requirements. Failure to accurately determine vibration boundaries may result in units operating within vibration zones, causing equipment fatigue, damage, or even malfunctions, directly impacting equipment safety and grid stability. Therefore, the refined and dynamic delineation of hydropower unit vibration zones has become an urgent task. This classification not only ensures the safe and flexible operation of generating units in the new power system, but also effectively supports the consumption of a high proportion of renewable energy, ensuring the reliability and economy of the power grid, and is an important technical foundation for realizing the transformation of the power system.

[0003] However, existing vibration zone delineation methods have significant limitations, restricting the accuracy and practicality of the results. First, the data relied upon for delineation typically comes from traditional stability tests at a limited number of head points. These tests cover a limited range of operating conditions, making it difficult to comprehensively and accurately reflect the actual efficiency and vibration performance of hydropower units across all head and load conditions. Hydropower units operate under a wide range of head variations, with significant differences in hydraulic and vibration characteristics at different heads. Data from only a few head points cannot capture the overall behavior, potentially leading to deviations in the delineation results in practical applications. Second, detailed vibration zone testing of actual units is costly, difficult to implement, and yields low data accuracy. Due to limitations in test conditions, the acquired vibration data often contains noise and errors, resulting in vibration zone delineations that are typically based on conservative design principles, leading to broad results. While this broad delineation ensures safety, it sacrifices the unit's operational flexibility. Furthermore, the unclear boundary between the allowable stability zone and the vibration zone can confuse operators and potentially affect the unit's frequency response capability. Traditional delineation methods heavily rely on manual experience to screen and judge test data to extract key characteristic parameters. This method is highly subjective, and different experts may reach different conclusions, leading to poor consistency in the classification results. Furthermore, manual data processing is inefficient and struggles to handle the massive amounts of operational data generated by modern hydropower units, making the classification process time-consuming and labor-intensive. Therefore, the accuracy and consistency of the classification cannot be guaranteed, limiting its effectiveness in practical applications and failing to meet the demands of modern power systems for rapid and flexible unit adjustments. Summary of the Invention

[0004] In view of this, the present invention proposes a method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate, which can achieve automatic and efficient screening of the dominant factors in vibration zone division.

[0005] The technical solution of this invention is implemented as follows:

[0006] A method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate includes the following steps:

[0007] Step S1: Obtain multiple sets of status monitoring data of the hydropower unit under multiple operating conditions, and preprocess the status monitoring data;

[0008] Step S2: Normalize the preprocessed state monitoring data to obtain the relative measurement values ​​of each state monitoring data.

[0009] Step S3: Based on the unit load, divide the status monitoring data of the same group under different operating conditions into continuous load intervals.

[0010] Step S4: Calculate the velocity gradient of the relative values ​​of the measured values ​​of the condition monitoring data at both ends of each load interval, and accumulate the absolute values ​​of the velocity gradients of all load intervals to obtain the velocity gradient accumulation sequence of each condition monitoring data as the load changes.

[0011] Step S5: Based on the relative measurement values ​​of each state monitoring data, calculate the shortest Euclidean distance of each state monitoring data point from the normalization boundary;

[0012] Step S6: Based on the velocity gradient accumulation sequence and the shortest Euclidean distance of the same state monitoring data, calculate the Euclidean distortion rate of each state monitoring data;

[0013] Step S7: Select several state monitoring data with large Euclidean distortion rates to form the dominant factor set for the division of the vibration zone of the hydropower unit.

[0014] Preferably, the status monitoring data includes the swing of each guide bearing of the unit, the vibration amplitude of the unit frame, the vibration amplitude of the unit top cover, and the pressure data of the unit volute and tailrace pipe.

[0015] Preferably, the preprocessing in step S1 includes screening and noise reduction.

[0016] Preferably, the normalization calculation formula in step S2 is:

[0017]

[0018] in This corresponds to the actual monitored value of a certain condition monitoring data under a specific operating condition. These are the standard allowable values ​​for the corresponding condition monitoring data. This refers to the relative measurement value of the corresponding status monitoring data.

[0019] Preferably, the tailrace pipe pressure data in the status monitoring data is corrected using the measured head value before normalization.

[0020] Preferably, the formula for calculating the velocity gradient accumulation sequence in step S4 is:

[0021]

[0022] in Let be the cumulative velocity gradient sequence of the first i load intervals of a certain state monitoring data. These are the relative measured values ​​of the condition monitoring data at operating points n and n+1 at both ends of the same load interval. These are the unit load operating parameters at the two operating points n and n+1.

[0023] Preferably, the formula for calculating the shortest Euclidean distance is:

[0024]

[0025] in The shortest Euclidean distance is given, and 1 represents the normalized boundary value. This represents the maximum relative value among the measured values ​​of a certain condition monitoring data under the corresponding operating conditions.

[0026] Preferably, the specific steps of step S6 are as follows:

[0027] The maximum and minimum values ​​are extracted from the velocity gradient accumulation sequence, and the velocity gradient distortion rate is calculated.

[0028] The Euclidean distortion rate is obtained by dividing the velocity gradient distortion rate by the shortest Euclidean distance.

[0029] Preferably, the formula for calculating the Euclidean distortion rate is:

[0030]

[0031] in The Euclidean distortion rate of the condition monitoring data. The maximum and minimum values ​​of the velocity gradient accumulation sequence for the corresponding state monitoring data are respectively calculated. This represents the shortest Euclidean distance to the corresponding status monitoring data.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] This invention discloses a method for identifying the dominant factor in the vibration zone division of a hydropower unit using Euclidean distortion rate. The method involves collecting multiple sets of condition monitoring data from the hydropower unit under various operating conditions, then normalizing and preprocessing the data. The condition monitoring data includes multi-source monitoring data, allowing for the acquisition of the same condition monitoring data under continuous switching of multiple operating conditions. After preprocessing the condition monitoring data, normalization is performed to obtain relative measurement values, eliminating differences between data from different measuring points. Then, the same set of condition monitoring data is divided into continuous load intervals, and the velocity gradient of each load interval is calculated. These gradients are then accumulated to obtain a velocity gradient accumulation sequence. Simultaneously, based on… The shortest Euclidean distance is calculated based on the relative values ​​of the measurements from each condition monitoring data point. The reciprocal of the shortest Euclidean distance is used as a weighting coefficient to characterize the proximity of the relative values ​​of the condition monitoring data to the boundary. A Euclidean distortion rate is constructed based on the shortest Euclidean distance and the maximum and minimum values ​​in the velocity gradient accumulation sequence to assess the severity of changes in the condition monitoring data. Ultimately, the Euclidean distortion rate can be used to automatically and objectively identify the dominant factors that are most sensitive to changes in operating conditions and have the most significant impact on safe operation from numerous condition monitoring data points. This reduces human error, improves the reliability and consistency of the classification results, and provides a theoretical basis for the optimized operation and condition monitoring of hydropower units. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart illustrating the calculation process of a method for identifying dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate, as described in this invention.

[0036] Figure 2 This is a line graph showing the variation of guide bearing vibration runout with load in the original data of a power station, provided as an embodiment of the method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate according to the present invention.

[0037] Figure 3 This is an embodiment of the method for identifying the dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate according to the present invention, which provides a velocity gradient accumulation piecewise linear graph of a guide bearing of a power station as a function of load.

[0038] Figure 4 This is a comparison chart of the final Euclidean distortion rate calculation values ​​of a power station guide bearing, provided as an embodiment of the method for identifying the dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate according to an embodiment of the present invention. Detailed Implementation

[0039] To better understand the technical content of this invention, a specific embodiment is provided below, and the invention will be further described in conjunction with the accompanying drawings.

[0040] See Figure 1 The present invention provides a method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate, comprising the following steps:

[0041] Step S1: Obtain multiple sets of status monitoring data of the hydropower unit under multiple operating conditions, and preprocess the status monitoring data;

[0042] Step S2: Normalize the preprocessed state monitoring data to obtain the relative measurement values ​​of each state monitoring data.

[0043] Step S3: Based on the unit load, divide the status monitoring data of the same group under different operating conditions into continuous load intervals.

[0044] Step S4: Calculate the velocity gradient of the relative values ​​of the measured values ​​of the condition monitoring data at both ends of each load interval, and accumulate the absolute values ​​of the velocity gradients of all load intervals to obtain the velocity gradient accumulation sequence of each condition monitoring data as the load changes.

[0045] Step S5: Based on the relative measurement values ​​of each state monitoring data, calculate the shortest Euclidean distance of each state monitoring data point from the normalization boundary;

[0046] Step S6: Based on the velocity gradient accumulation sequence and the shortest Euclidean distance of the same state monitoring data, calculate the Euclidean distortion rate of each state monitoring data;

[0047] Step S7: Select several state monitoring data with large Euclidean distortion rates to form the dominant factor set for the division of the vibration zone of the hydropower unit.

[0048] This invention discloses a method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate. Unlike traditional identification methods, this method does not rely on manual observation and data screening. Instead, it directly calculates and processes the monitoring data generated by the unit under actual operating conditions to automatically and quickly screen the dominant factors for vibration zone division. The core of this invention is to construct a comprehensive index called Euclidean distortion rate. This index can simultaneously assess the degree of change in condition monitoring data and its proximity to the safety boundary, thereby automatically and objectively identifying the dominant factors that are most sensitive to changes in operating conditions and have the most significant impact on safe operation from numerous monitoring data.

[0049] Specifically, the Euclidean distortion rate is calculated based on the time series data of the monitoring data, determining its deviation from the baseline state, and quantifying the risk by combining it with a safety threshold. This data-driven approach can accurately reflect the vibration characteristics of the unit under different operating conditions, identify key monitoring data that lead to increased vibration, and provide a crucial basis for subsequent precise vibration zone division, making the division results more consistent with actual operating conditions. At the same time, the method of this invention has a fast processing speed and can adapt to the needs of real-time or near-real-time data analysis, effectively improving the safety and economy of hydropower stations in flexible operation in new power systems. By identifying automated dominant factors, it reduces human error, improves the reliability and consistency of the division results, and provides a powerful tool for the optimized operation and condition monitoring of hydropower units, helping to promote the more efficient integration of hydropower stations in high-proportion renewable energy grids.

[0050] Before classifying the dominant factors, the subject to be monitored is first determined, using the same component of the same hydropower unit as the monitoring object. This object is then continuously operated under multiple different operating conditions. Multiple sets of status monitoring data are collected throughout the entire operating cycle. This status monitoring data includes data from multiple sources, thus ultimately providing time-series variation data for each dimension of status monitoring data under various operating condition switching. The status monitoring data is then preprocessed. Due to the different dimensions of the multi-source data and the significant differences in amplitude variations at different locations, normalization is required to include all status monitoring data within a unified framework for comparison. This yields the relative measurement values ​​of the status monitoring data. After obtaining the relative measurement values, velocity gradients and the shortest Euclidean distance can be calculated based on these values. The shortest Euclidean distance can be calculated based on the difference between the normalized boundary and the maximum value of the relative measurement. When calculating the velocity gradient, a basis is needed... The condition monitoring data of the same group is divided into multiple continuous load intervals based on the unit load. The two ends of each load interval are adjacent operating points. The velocity gradient of each load interval can be calculated based on the relative values ​​of the condition monitoring data measured at the two ends of each load interval. Then, the absolute values ​​of the velocity gradients of all load intervals are accumulated to obtain the velocity gradient accumulation sequence of the corresponding condition monitoring data as the load changes. Finally, the Euclidean distortion rate can be calculated based on the velocity gradient accumulation sequence and the shortest Euclidean distance. After obtaining the Euclidean distortion rate of all condition monitoring data, a comprehensive comparison is used to select 5-6 condition monitoring data with the largest Euclidean distortion rate as the dominant factor set for output. This can be used to divide the vibration zone, so as to effectively avoid dangerous operating conditions during scheduling and operation, help extend the life of key unit components, and release the optimization scheduling space that is restricted by conservative division, thereby improving the overall safety and economy of power plant operation.

[0051] This invention achieves automatic screening and comparison of dominant factors in multi-source condition monitoring data through a unified calculation model, significantly reducing the workload, time cost, and potential subjective errors of manual analysis, and improving the efficiency and accuracy of screening dominant factors of hydropower unit vibration. This method directly relies on actual unit operation or test data, thus naturally reflecting the individual characteristics of different units, and can adapt to changes in unit status or operating environment by updating data, providing an effective technical basis for implementing personalized and dynamic vibration zone division and management.

[0052] Preferably, the status monitoring data includes the swing of each guide bearing of the unit, the vibration amplitude of the unit frame, the vibration amplitude of the unit top cover, and the pressure data of the unit volute and tailrace pipe.

[0053] The guide bearing is a key component supporting the main shaft of the hydro-generator unit. Its runout refers to the radial runout of the main shaft around the center during operation, reflecting the concentricity of the main shaft. Excessive runout indicates severe main shaft runout, which can easily cause bearing wear and increased unit vibration. It is a core indicator for judging the mechanical stability of the unit. Excessive frame vibration amplitude indicates that the vibration of the core components of the unit has not been effectively attenuated, which can easily lead to loose frame bolts and foundation cracking. It can also indirectly reflect abnormal operation of the guide bearing and runner. The vibration amplitude of the top cover mainly reflects the vibration caused by the hydraulic excitation of the turbine. It is a key state monitoring data linking hydraulic factors and mechanical vibration, and is also an important basis for judging the hydraulic stability of the unit. The turbine casing and draft tube are flow-through components of the turbine. They mainly monitor the pressure pulsation value during operation, characterizing the hydraulic operating state of the unit. They are core indicators for judging hydraulic vibration.

[0054] Preferably, the preprocessing in step S1 includes screening and noise reduction.

[0055] By filtering and noise reduction, invalid and interfering data can be eliminated, improving the accuracy and effectiveness of condition monitoring data. This provides a reliable data foundation for subsequent normalization, Euclidean distortion rate calculation, and dominant factor identification, ensuring the accuracy of the vibration zone division results for hydropower units.

[0056] Preferably, the normalization calculation formula in step S2 is:

[0057]

[0058] in This corresponds to the actual monitored value of a certain condition monitoring data under a specific operating condition. These are the standard allowable values ​​for the corresponding condition monitoring data. This refers to the relative measurement value of the corresponding status monitoring data.

[0059] Because the dimensions of multi-source condition monitoring data are different and the amplitude changes in different parts vary greatly, normalization processing is required. Normalization can be achieved by dividing the actual monitored value of the collected condition monitoring data by the standard allowable value. The standard allowable value is the allowable value obtained by referring to the standard GB / T 32584-2016.

[0060] Preferably, the tailrace pipe pressure data in the status monitoring data is corrected using the measured head value before normalization.

[0061] For condition monitoring data such as pressure pulsation in the tailrace pipe and cone pipe, which are relatively minor in unit vibration analysis, the measured head value of the measuring point is used as a standardization benchmark during data processing. The corresponding condition monitoring data is then corrected and normalized to eliminate the influence of head changes on the absolute value of the condition monitoring data. This allows for fair comparison with condition monitoring data from other measuring points on a consistent basis, and objectively assesses the relative degree of influence of different condition monitoring data on the unit's condition.

[0062] Preferably, the formula for calculating the velocity gradient accumulation sequence in step S4 is:

[0063]

[0064] in Let be the cumulative velocity gradient sequence of the first i load intervals of a certain state monitoring data. These are the relative measured values ​​of the condition monitoring data at operating points n and n+1 at both ends of the same load interval. These are the unit load operating parameters at the two operating points n and n+1.

[0065] After obtaining the relative measurement values ​​of the same unit and the same component under different operating conditions, since the relative measurement values ​​of different components vary with the operating conditions, a velocity gradient is introduced to accurately reflect the changes in the relative measurement values ​​of different components as the operating conditions change. The condition monitoring data is divided into multiple continuous load intervals based on the unit load. The slope of the relative measurement values ​​of the same unit and the same component under adjacent operating conditions within the same load interval is used to reflect the changes. Simultaneously, to represent the overall change of the curve, the velocity gradient uses the absolute values ​​of the calculated slopes and accumulates them to characterize the changes in the same unit and the same component under different operating conditions. In this method, using absolute values ​​effectively ensures the monotonicity of the velocity gradient, eliminates the fluctuation effects of positive and negative changes, and amplifies the rate of data change, thus more effectively reflecting the fluctuations in the relative measurement values ​​of the same unit and the same component under adjacent operating conditions.

[0066] Preferably, the formula for calculating the shortest Euclidean distance is:

[0067]

[0068] in The shortest Euclidean distance is given, and 1 represents the normalized boundary value. This represents the maximum relative value among the measured values ​​of a certain condition monitoring data under the corresponding operating conditions.

[0069] When screening the dominant factors of hydropower unit vibration, in addition to considering the stability of the relative values ​​of the condition monitoring data as the operation changes, it is also necessary to consider the distance of the relative values ​​from the boundary. Therefore, this invention uses the reciprocal of the shortest Euclidean distance as a weighting coefficient to characterize the closeness of the relative values ​​of the condition monitoring data to the boundary "1". The smaller the shortest Euclidean distance, the closer the relative values ​​of the condition monitoring data are to the boundary, and the larger the weighting coefficient obtained. At the same time, considering that when the relative values ​​of the condition monitoring data reach or exceed the boundary, the shortest Euclidean distance is 0 or negative, losing its comparative significance, therefore, when... At that time, it was stipulated This will increase the weight value of the status monitoring data.

[0070] Preferably, the specific steps of step S6 are as follows:

[0071] The maximum and minimum values ​​are extracted from the velocity gradient accumulation sequence, and the velocity gradient distortion rate is calculated.

[0072] The Euclidean distortion rate is calculated by dividing the velocity gradient distortion rate by the shortest Euclidean distance. The formula for calculating the Euclidean distortion rate is as follows:

[0073]

[0074] in The Euclidean distortion rate of the condition monitoring data. These correspond to the maximum and minimum values ​​of the velocity gradient accumulation sequence of the state monitoring data, respectively. This represents the shortest Euclidean distance to the corresponding status monitoring data.

[0075] Taking into account both the stability of the relative values ​​of measured parameters under varying operating conditions and the distance of the relative values ​​of measured parameters from the boundary, this invention uses the Euclidean distortion rate as the final parameter to screen the dominant factors of vibration in hydropower units. For the same unit and the same component, the Euclidean distortion rate is calculated by dividing the velocity gradient distortion rate by the corresponding shortest Euclidean distance. The velocity gradient distortion rate is calculated as the difference between the maximum and minimum velocity gradient values ​​for the same unit and the same component. The Euclidean distortion rate is used to reflect the trade-off between the gradient of vibration data with load conditions and the allowable value from the standard. The larger the value, the greater the gradient of data change and the closer it is to the allowable value from the standard. Therefore, a larger value of Euclidean distortion rate should be selected when choosing the dominant factor. Finally, 5-6 condition monitoring data are selected to form a set of dominant factors, and the vibration zone is divided based on this.

[0076] The present invention will be further described in detail below through an embodiment:

[0077] Reference Figure 2-4 As shown, the analysis is based on actual operating data from a power plant. Figure 2 A line graph showing the relationship between the runout of the guide bearing in the power station unit and load variation is presented. The graph visually demonstrates the significant change in runout within a specific load range, allowing for rapid identification of the potential correlation between the measuring point parameter and vibration characteristics. However, this method, based on subjective visual judgment, lacks quantitative evidence and suffers from insufficient repeatability and objectivity. When faced with large datasets and an increasing number of monitoring points, the efficiency of manual interpretation drops sharply, consistency is difficult to guarantee, and errors are easily introduced, affecting the accuracy and reliability of the final vibration zone division results. To address this issue, the method of identifying the dominant factors for vibration zone division of hydropower units based on Euclidean distortion rate, as described in this invention, is used to process the X-axis runout of the upper guide bearing in this power station. The specific steps for screening the dominant factors are as follows:

[0078] Step 1: Collect 20 sets of guide bearing vibration amplitude data of the unit within a specific load range at intervals of 30MW. After processing, divide the data into 19 segments according to the load.

[0079] Step 2: Normalize the collected data. For example, according to GB / T 32584-2016, the maximum allowable value of the X-direction runout of the upper guide bearing is 250µm. Therefore, divide the collected X-direction runout of the upper guide bearing by 250 to obtain the normalized data.

[0080] Step 3: Calculate the velocity gradient value of the bearing vibration amplitude at adjacent working points within each segment. That is, with the working condition as the abscissa and the vibration amplitude as the ordinate, calculate the velocity gradient value at each point, and accumulate all the velocity gradient values ​​within each segment to obtain 19 cumulative gradient values ​​H1 to H19;

[0081] Step 4: Calculate the shortest Euclidean distance of the normalized data. For example, if the normalized value of the X-direction runout of a certain upper guide bearing is 0.464, then its shortest Euclidean distance is... ;

[0082] Step 5: Calculate the Euclidean distortion rate, which is the product of the difference between the maximum and minimum values ​​in the cumulative velocity gradient sequence and the Euclidean shortest distance.

[0083] The vibration amplitude data, normalized data, velocity gradient, and gradient accumulation value used for calculating the X-axis runout of the above-mentioned guide bearing are shown in Table 1:

[0084] Table 1. Parameter calculation table using the X-direction runout of the above guide bearing as an example.

[0085]

[0086] Similarly, calculate the velocity gradients and cumulative gradient values ​​of the X-axis and Y-axis runouts of the lower guide bearing and the water guide bearing, and plot them as shown below. Figure 3 The line graph shown;

[0087] Step 6: Compare the Euclidean distortion distances of different data groups, referring to... Figure 4 As shown, the larger the value, the greater the vibration amplitude in the overall parameters and the closer it is to the allowable value of the national standard. The data shows that the gradient accumulation value sequence of the water guide bearing data is the largest, indicating that the water guide bearing parameters have the most significant relationship with the unit vibration on the guide bearing and are the dominant parameters for dividing this vibration zone.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate, characterized in that, Includes the following steps: Step S1: Obtain multiple sets of status monitoring data of the hydropower unit under multiple operating conditions, and preprocess the status monitoring data; Step S2: Normalize the preprocessed state monitoring data to obtain the relative measurement values ​​of each state monitoring data. Step S3: Based on the unit load, divide the status monitoring data of the same group under different operating conditions into continuous load intervals. Step S4: Calculate the velocity gradient of the relative values ​​of the measured values ​​of the condition monitoring data at both ends of each load interval, and accumulate the absolute values ​​of the velocity gradients of all load intervals to obtain the velocity gradient accumulation sequence of each condition monitoring data as the load changes. Step S5: Based on the relative measurement values ​​of each state monitoring data, calculate the shortest Euclidean distance of each state monitoring data point from the normalization boundary; Step S6: Based on the velocity gradient accumulation sequence and the shortest Euclidean distance of the same state monitoring data, calculate the Euclidean distortion rate of each state monitoring data; Step S7: Select several state monitoring data with large Euclidean distortion rates to form the dominant factor set for the division of the vibration zone of the hydropower unit. The formula for calculating the velocity gradient accumulation sequence in step S4 is as follows: in Let be the cumulative velocity gradient sequence of the first i load intervals of a certain state monitoring data. These are the relative measured values ​​of the condition monitoring data at operating points n and n+1 at both ends of the same load interval. The load condition parameters of the unit at the two operating points n and n+1 are given. The formula for calculating the shortest Euclidean distance is: in The shortest Euclidean distance is given, and 1 represents the normalized boundary value. This refers to the maximum value among the relative values ​​of the measured data under a certain operating condition; The specific steps of step S6 are as follows: The maximum and minimum values ​​are extracted from the velocity gradient accumulation sequence, and the velocity gradient distortion rate is calculated. The product of the velocity gradient distortion rate and the shortest Euclidean distance is taken as the Euclidean distortion rate; The formula for calculating the Euclidean distortion rate is as follows: in The Euclidean distortion rate of the condition monitoring data. These represent the maximum and minimum values ​​of the velocity gradient accumulation sequence corresponding to the state monitoring data. This represents the shortest Euclidean distance to the corresponding status monitoring data.

2. The method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate according to claim 1, characterized in that, The condition monitoring data includes the swing of each guide bearing of the unit, the vibration amplitude of the unit frame, the vibration amplitude of the unit top cover, and the pressure data of the unit volute and tailrace pipe.

3. The method for identifying the dominant factors in the vibration zone division of hydropower units using Euclidean distortion rate according to claim 1, characterized in that, The preprocessing in step S1 includes screening and noise reduction.

4. The method for identifying the dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate according to claim 2, characterized in that, The normalization calculation formula in step S2 is as follows: in This corresponds to the actual monitored value of a certain condition monitoring data under a specific operating condition. These are the standard allowable values ​​for the corresponding condition monitoring data. This refers to the relative measurement value of the corresponding status monitoring data.

5. The method for identifying the dominant factors in the vibration zone division of a hydropower unit using Euclidean distortion rate according to claim 4, characterized in that, Before normalization, the tailrace pressure data in the condition monitoring data is corrected using the measured head value.

Images